avoid symbolic link to files in packaging

Hatch cannot handle those correctly and will add a symbolic link to the
source package.

.flake8

@@ -6,3 +6,6 @@ extend-ignore =
     E711
 per-file-ignores =
     __init__.py: F401
+    test/python/utils/test_json_writer.py: E131
+    **/conftest.py: E402
+    test/bdd/*: F821

.github/actions/setup-postgresql/action.yml

@@ -11,10 +11,8 @@ runs:
     steps:
         - name: Remove existing PostgreSQL
           run: |
+              sudo /usr/share/postgresql-common/pgdg/apt.postgresql.org.sh -y
               sudo apt-get purge -yq postgresql*
-              sudo apt install curl ca-certificates gnupg
-              curl https://www.postgresql.org/media/keys/ACCC4CF8.asc | gpg --dearmor | sudo tee /etc/apt/trusted.gpg.d/apt.postgresql.org.gpg >/dev/null
-              sudo sh -c 'echo "deb https://apt.postgresql.org/pub/repos/apt $(lsb_release -cs)-pgdg main" > /etc/apt/sources.list.d/pgdg.list'
               sudo apt-get update -qq
 
           shell: bash

.github/workflows/ci-tests.yml

@@ -37,18 +37,20 @@ jobs:
         needs: create-archive
         strategy:
             matrix:
-                flavour: ["ubuntu-20", "ubuntu-24"]
+                flavour: ["ubuntu-22", "ubuntu-24"]
                 include:
-                    - flavour: ubuntu-20
-                      ubuntu: 20
+                    - flavour: ubuntu-22
+                      ubuntu: 22
                       postgresql: 12
                       lua: '5.1'
                       dependencies: pip
+                      python: '3.9'
                     - flavour: ubuntu-24
                       ubuntu: 24
                       postgresql: 17
                       lua: '5.3'
                       dependencies: apt
+                      python: 'builtin'
 
         runs-on: ubuntu-${{ matrix.ubuntu }}.04
 
@@ -68,26 +70,40 @@ jobs:
               with:
                   dependencies: ${{ matrix.dependencies }}
 
+            - uses: actions/cache@v4
+              with:
+                  path: |
+                     /usr/local/bin/osm2pgsql
+                  key: osm2pgsql-bin-22-1
+              if: matrix.ubuntu == '22'
+
+            - name: Set up Python
+              uses: actions/setup-python@v5
+              with:
+                  python-version: ${{ matrix.python }}
+              if: matrix.python != 'builtin'
+
             - name: Compile osm2pgsql
               run: |
-                  sudo apt-get install -y -qq libboost-system-dev libboost-filesystem-dev libexpat1-dev zlib1g-dev libbz2-dev libpq-dev libproj-dev libicu-dev liblua${LUA_VERSION}-dev lua-dkjson nlohmann-json3-dev 
-                  mkdir osm2pgsql-build
-                  cd osm2pgsql-build
-                  git clone https://github.com/osm2pgsql-dev/osm2pgsql
-                  mkdir build
-                  cd build
-                  cmake ../osm2pgsql
-                  make
-                  sudo make install
-                  cd ../..
-                  rm -rf osm2pgsql-build
-              if: matrix.ubuntu == '20'
+                  if [ ! -f /usr/local/bin/osm2pgsql ]; then
+                      sudo apt-get install -y -qq libboost-system-dev libboost-filesystem-dev libexpat1-dev zlib1g-dev libbz2-dev libpq-dev libproj-dev libicu-dev liblua${LUA_VERSION}-dev lua-dkjson nlohmann-json3-dev 
+                      mkdir osm2pgsql-build
+                      cd osm2pgsql-build
+                      git clone https://github.com/osm2pgsql-dev/osm2pgsql
+                      mkdir build
+                      cd build
+                      cmake ../osm2pgsql
+                      make
+                      sudo make install
+                      cd ../..
+                      rm -rf osm2pgsql-build
+                  else
+                      sudo apt-get install -y -qq libexpat1 liblua${LUA_VERSION}
+                  fi
+              if: matrix.ubuntu == '22'
               env:
                   LUA_VERSION: ${{ matrix.lua }}
 
-            - name: Install test prerequisites
-              run: ./venv/bin/pip install behave==1.2.6
-
             - name: Install test prerequisites (apt)
               run: sudo apt-get install -y -qq python3-pytest python3-pytest-asyncio uvicorn python3-falcon python3-aiosqlite python3-pyosmium
               if: matrix.dependencies == 'apt'
@@ -96,11 +112,14 @@ jobs:
               run: ./venv/bin/pip install pytest-asyncio falcon starlette asgi_lifespan aiosqlite osmium uvicorn
               if: matrix.dependencies == 'pip'
 
+            - name: Install test prerequisites
+              run: ./venv/bin/pip install pytest-bdd
+
             - name: Install latest flake8
               run: ./venv/bin/pip install -U flake8
 
             - name: Python linting
-              run: ../venv/bin/python -m flake8 src
+              run: ../venv/bin/python -m flake8 src test/python test/bdd
               working-directory: Nominatim
 
             - name: Install mypy and typechecking info
@@ -108,7 +127,7 @@ jobs:
               if: matrix.dependencies == 'pip'
 
             - name: Python static typechecking
-              run: ../venv/bin/python -m mypy --strict --python-version 3.8 src
+              run: ../venv/bin/python -m mypy --strict --python-version 3.9 src
               working-directory: Nominatim
               if: matrix.dependencies == 'pip'
 
@@ -118,8 +137,8 @@ jobs:
 
             - name: BDD tests
               run: |
-                  ../../../venv/bin/python -m behave -DREMOVE_TEMPLATE=1 --format=progress3
-              working-directory: Nominatim/test/bdd
+                  ../venv/bin/python -m pytest test/bdd --nominatim-purge
+              working-directory: Nominatim
 
     install:
         runs-on: ubuntu-latest

CONTRIBUTING.md

@@ -113,3 +113,5 @@ Checklist for releases:
   * run `nominatim --version` to confirm correct version
 * [ ] tag new release and add a release on github.com
 * [ ] build pip packages and upload to pypi
+  * `make build`
+  * `twine upload dist/*`

ChangeLog

@@ -1,3 +1,22 @@
+5.1.0
+ * replace datrie with simple internal trie implementation
+ * add pattern-based postcode parser for queries,
+   postcodes no longer need to be present in OSM to be found
+ * take variants into account when computing token similarity
+ * add extratags output to geocodejson format
+ * fix default layer setting used for structured queries
+ * update abbreviation lists for Russian and English
+   (thanks @shoorick, @IvanShift, @mhsrn21)
+ * fix variant generation for Norwegian
+ * fix normalization around space-like characters
+ * improve postcode search and handling of postcodes in queries
+ * reorganise internal query structure and get rid of slow enums
+ * enable code linting for tests
+ * various code moderinsations in test code (thanks @eumiro)
+ * remove setting osm2pgsql location via config.lib_dir
+ * make SQL functions parallel save as far as possible (thanks @otbutz)
+ * various fixes and improvements to documentation (thanks @TuringVerified)
+
 5.0.0
  * increase required versions for PostgreSQL (12+), PostGIS (3.0+)
  * remove installation via cmake and debundle osm2pgsql

Makefile

@@ -18,16 +18,16 @@ build-api:
 tests: mypy lint pytest bdd
 
 mypy:
-	mypy --strict --python-version 3.8 src
+	mypy --strict --python-version 3.9 src
 
 pytest:
 	pytest test/python
 
 lint:
-	flake8 src
+	flake8 src test/python test/bdd
 
 bdd:
-	cd test/bdd; behave -DREMOVE_TEMPLATE=1
+	pytest test/bdd --nominatim-purge
 
 # Documentation
 

README.md

@@ -27,18 +27,25 @@ can be found at nominatim.org as well.
 
 A quick summary of the necessary steps:
 
-1. Create a Python virtualenv and install the packages:
+
+1. Clone this git repository and download the country grid
+
+        git clone https://github.com/osm-search/Nominatim.git
+        wget -O Nominatim/data/country_osm_grid.sql.gz https://nominatim.org/data/country_grid.sql.gz
+
+2. Create a Python virtualenv and install the packages:
 
         python3 -m venv nominatim-venv
         ./nominatim-venv/bin/pip install packaging/nominatim-{api,db}
 
-2. Create a project directory, get OSM data and import:
+3. Create a project directory, get OSM data and import:
 
         mkdir nominatim-project
         cd nominatim-project
-        ../nominatim-venv/bin/nominatim import --osm-file <your planet file>
+        ../nominatim-venv/bin/nominatim import --osm-file <your planet file> 2>&1 | tee setup.log
 
-3. Start the webserver:
+
+4. Start the webserver:
 
         ./nominatim-venv/bin/pip install uvicorn falcon
         ../nominatim-venv/bin/nominatim serve

SECURITY.md

@@ -9,7 +9,8 @@ versions.
 
 | Version | End of support for security updates |
 | ------- | ----------------------------------- |
-| 5.0.x   | 2027-02-06
+| 5.1.x   | 2027-04-01                          |
+| 5.0.x   | 2027-02-06                          |
 | 4.5.x   | 2026-09-12                          |
 | 4.4.x   | 2026-03-07                          |
 | 4.3.x   | 2025-09-07                          |

docs/admin/Installation.md

@@ -27,7 +27,7 @@ For running Nominatim:
   * [PostgreSQL](https://www.postgresql.org) (12+ will work, 13+ strongly recommended)
   * [PostGIS](https://postgis.net) (3.0+ will work, 3.2+ strongly recommended)
   * [osm2pgsql](https://osm2pgsql.org) (1.8+)
-  * [Python 3](https://www.python.org/) (3.7+)
+  * [Python 3](https://www.python.org/) (3.9+)
 
 Furthermore the following Python libraries are required:
 
@@ -37,7 +37,6 @@ Furthermore the following Python libraries are required:
   * [Jinja2](https://palletsprojects.com/p/jinja/)
   * [PyICU](https://pypi.org/project/PyICU/)
   * [PyYaml](https://pyyaml.org/) (5.1+)
-  * [datrie](https://github.com/pytries/datrie)
 
 These will be installed automatically when using pip installation.
 

docs/admin/Setup-Nominatim-UI.md

@@ -36,11 +36,11 @@ The website is now available at `http://localhost:8765`.
 ## Forwarding searches to nominatim-ui
 
 Nominatim used to provide the search interface directly by itself when
-`format=html` was requested. For all endpoints except for `/reverse` and
-`/lookup` this even used to be the default.
+`format=html` was requested. For the `/search` endpoint this even used
+to be the default.
 
 The following section describes how to set up Apache or nginx, so that your
-users are forwarded to nominatim-ui when they go to URL that formerly presented
+users are forwarded to nominatim-ui when they go to a URL that formerly presented
 the UI.
 
 ### Setting up forwarding in Nginx
@@ -73,41 +73,28 @@ map $args $format {
 
 # Determine from the URI and the format parameter above if forwarding is needed.
 map $uri/$format $forward_to_ui {
-    default               1;   # The default is to forward.
-    ~^/ui                 0;   # If the URI point to the UI already, we are done.
-    ~/other$              0;   # An explicit non-html format parameter. No forwarding.
-    ~/reverse.*/default   0;   # Reverse and lookup assume xml format when
-    ~/lookup.*/default    0;   #   no format parameter is given. No forwarding.
+    default               0;  # no forwarding by default
+    ~/search.*/default    1;  # Use this line only, if search should go to UI by default.
+    ~/reverse.*/html      1;  # Forward API calls that UI supports, when
+    ~/status.*/html       1;  #   format=html is explicitly requested.
+    ~/search.*/html       1;
+    ~/details.*/html      1;
 }
 ```
 
 The `$forward_to_ui` parameter can now be used to conditionally forward the
 calls:
 
-```
-# When no endpoint is given, default to search.
-# Need to add a rewrite so that the rewrite rules below catch it correctly.
-rewrite ^/$ /search;
-
-location @php {
-    # fastcgi stuff..
+``` nginx
+location / {
     if ($forward_to_ui) {
-        rewrite ^(/[^/]*) https://yourserver.com/ui$1.html redirect;
+        rewrite ^(/[^/.]*) https://$http_host/ui$1.html redirect;
     }
-}
 
-location ~ [^/]\.php(/|$) {
-    # fastcgi stuff..
-    if ($forward_to_ui) {
-        rewrite (.*).php https://yourserver.com/ui$1.html redirect;
-    }
+    # proxy_pass commands
 }
 ```
 
-!!! warning
-    Be aware that the rewrite commands are slightly different for URIs with and
-    without the .php suffix.
-
 Reload nginx and the UI should be available.
 
 ### Setting up forwarding in Apache
@@ -159,18 +146,16 @@ directory like this:
   RewriteBase "/nominatim/"
 
   # If no endpoint is given, then use search.
-  RewriteRule ^(/|$)   "search.php"
+  RewriteRule ^(/|$)   "search"
 
   # If format-html is explicitly requested, forward to the UI.
   RewriteCond %{QUERY_STRING} "format=html"
-  RewriteRule ^([^/]+)(.php)? ui/$1.html [R,END]
+  RewriteRule ^([^/.]+) ui/$1.html [R,END]
 
-  # If no format parameter is there then forward anything
-  # but /reverse and /lookup to the UI.
+  # Optionally: if no format parameter is there then forward /search.
   RewriteCond %{QUERY_STRING} "!format="
-  RewriteCond %{REQUEST_URI}  "!/lookup"
-  RewriteCond %{REQUEST_URI}  "!/reverse"
-  RewriteRule ^([^/]+)(.php)? ui/$1.html [R,END]
+  RewriteCond %{REQUEST_URI}  "/search"
+  RewriteRule ^([^/.]+) ui/$1.html [R,END]
 </Directory>
 ```
 

docs/api/Output.md

@@ -106,8 +106,11 @@ The following feature attributes are implemented:
  * `name` - localised name of the place
  * `housenumber`, `street`, `locality`, `district`, `postcode`, `city`,
    `county`, `state`, `country` -
-   provided when it can be determined from the address
+   provided when it can be determined from the address (only with `addressdetails=1`)
  * `admin` - list of localised names of administrative boundaries (only with `addressdetails=1`)
+ * `extra` - dictionary with additional useful tags like `website` or `maxspeed`
+   (only with `extratags=1`)
+
 
 Use `polygon_geojson` to output the full geometry of the object instead
 of the centroid.

docs/api/Search.md

@@ -212,7 +212,7 @@ other layers.
 The featureType allows to have a more fine-grained selection for places
 from the address layer. Results can be restricted to places that make up
 the 'state', 'country' or 'city' part of an address. A featureType of
-settlement selects any human inhabited feature from 'state' down to
+`settlement` selects any human inhabited feature from 'state' down to
 'neighbourhood'.
 
 When featureType is set, then results are automatically restricted

docs/customize/Import-Styles.md

@@ -36,18 +36,27 @@ local flex = require('flex-base')
 ### Using preset configurations
 
 If you want to start with one of the existing presets, then you can import
-its settings using the `import_topic()` function:
+its settings using the `load_topic()` function:
 
-```
+``` lua
 local flex = require('flex-base')
 
-flex.import_topic('streets')
+flex.load_topic('streets')
 ```
 
-The `import_topic` function takes an optional second configuration
+The `load_topic` function takes an optional second configuration
 parameter. The available options are explained in the
 [themepark section](#using-osm2pgsql-themepark).
 
+Available topics are: `admin`, `street`, `address`, `full`. These topic
+correspond to the [import styles](../admin/Import.md#filtering-imported-data)
+you can choose during import. To start with the 'extratags' style, use the
+`full` topic with the appropriate config parameter:
+
+``` lua
+flex.load_topic('full', {with_extratags = true})
+```
+
 !!! note
     You can also directly import the preset style files, e.g.
     `local flex = require('import-street')`. It is not possible to
@@ -116,8 +125,10 @@ value without key, then this is used as default for values that are not listed.
 
 `set_main_tags()` will completely replace the current main tag configuration
 with the new configuration. `modify_main_tags()` will merge the new
-configuration with the existing one. Otherwise, the two functions do exactly
-the same.
+configuration with the existing one. Merging is done at value level.
+For example, when the current setting is `highway = {'always', primary = 'named'}`,
+then `set_main_tags{highway = 'delete'}` will result in a rule
+`highway = {'delete', primary = 'named'}`.
 
 !!! example
     ``` lua
@@ -134,9 +145,9 @@ the same.
     when it has a value of `administrative`. Objects with `highway` tags are
     always included with two exceptions: the troll tag `highway=no` is
     deleted on the spot. And when the value is `street_lamp` then the object
-    must have a name, too. Finally, if a `landuse` tag is present then
-    it will be used independently of the concrete value when neither boundary
-    nor highway tags were found and the object is named.
+    must also have a name, to be included. Finally, if a `landuse` tag is
+    present then it will be used independently of the concrete value when
+    neither boundary nor highway tags were found and the object is named.
 
 ##### Presets
 
@@ -556,16 +567,6 @@ the Nominatim topic.
     ```
     Discarding country-level boundaries when running under themepark.
 
-## osm2pgsql gazetteer output
-
-Nominatim still allows you to configure the gazetteer output to remain
-backwards compatible with older imports. It will be automatically used
-when the style file name ends in `.style`. For documentation of the
-old import style, please refer to the documentation of older releases
-of Nominatim. Do not use the gazetteer output for new imports. There is no
-guarantee that new versions of Nominatim are fully compatible with the
-gazetteer output.
-
 ## Changing the style of existing databases
 
 There is usually no issue changing the style of a database that is already

docs/customize/Settings.md

@@ -602,25 +602,44 @@ results gathered so far.
 Note that under high load you may observe that users receive different results
 than usual without seeing an error. This may cause some confusion.
 
-### Logging Settings
-
-#### NOMINATIM_LOG_DB
+#### NOMINATIM_OUTPUT_NAMES
 
 | Summary            |                                                     |
 | --------------     | --------------------------------------------------- |
-| **Description:**   | Log requests into the database |
-| **Format:**        | boolean |
-| **Default:**       | no |
-| **After Changes:** | run `nominatim refresh --website` |
+| **Description:**   | Specifies order of name tags |
+| **Format:**        | string: comma-separated list of tag names |
+| **Default:**       | name:XX,name,brand,official_name:XX,short_name:XX,official_name,short_name,ref |
 
-Enable logging requests into a database table with this setting. The logs
-can be found in the table `new_query_log`.
+Specifies the order in which different name tags are used.
+The values in this list determine the preferred order of name variants,
+including language-specific names (in OSM: the name tag with and without any language suffix).
 
-When using this logging method, it is advisable to set up a job that
-regularly clears out old logging information. Nominatim will not do that
-on its own.
+Comma-separated list, where :XX stands for language suffix
+(e.g. name:en) and no :XX stands for general tags (e.g. name).
 
-Can be used as the same time as NOMINATIM_LOG_FILE.
+See also [NOMINATIM_DEFAULT_LANGUAGE](#nominatim_default_language).
+
+!!! note
+    If NOMINATIM_OUTPUT_NAMES = `name:XX,name,short_name:XX,short_name` the search follows
+
+        ```
+        'name', 'short_name'
+        ```
+
+    if we have no preferred language order for showing search results.
+
+    For languages ['en', 'es'] the search follows
+
+        ```
+        'name:en', 'name:es',
+        'name',
+        'short_name:en', 'short_name:es',
+        'short_name'
+        ```
+
+    For those familiar with the internal implementation, the `_place_*` expansion is added, but to simplify, it is not included in this example.
+
+### Logging Settings
 
 #### NOMINATIM_LOG_FILE
 
@@ -645,8 +664,6 @@ given in seconds and includes the entire time the query was queued and executed
 in the frontend.
 type contains the name of the endpoint used.
 
-Can be used as the same time as NOMINATIM_LOG_DB.
-
 #### NOMINATIM_DEBUG_SQL
 
 | Summary            |                                                     |

docs/customize/Tokenizers.md

@@ -67,7 +67,13 @@ Here is an example configuration file:
 
 ``` yaml
 query-preprocessing:
-    - normalize
+    - step: split_japanese_phrases
+    - step: regex_replace
+      replacements:
+        - pattern: https?://[^\s]* # Filter URLs starting with http or https
+          replace: ''
+    - step: normalize
+
 normalization:
     - ":: lower ()"
     - "ß > 'ss'" # German szet is unambiguously equal to double ss
@@ -88,8 +94,8 @@ token-analysis:
             replacements: ['ä', 'ae']
 ```
 
-The configuration file contains four sections:
-`normalization`, `transliteration`, `sanitizers` and `token-analysis`.
+The configuration file contains five sections:
+`query-preprocessing`, `normalization`, `transliteration`, `sanitizers` and `token-analysis`.
 
 #### Query preprocessing
 
@@ -106,6 +112,19 @@ The following is a list of preprocessors that are shipped with Nominatim.
         heading_level: 6
         docstring_section_style: spacy
 
+##### regex-replace
+
+::: nominatim_api.query_preprocessing.regex_replace
+    options:
+        members: False
+        heading_level: 6
+        docstring_section_style: spacy
+    description: 
+        This option runs any given regex pattern on the input and replaces values accordingly
+    replacements:
+        - pattern: regex pattern
+          replace: string to replace with
+
 
 #### Normalization and Transliteration
 

docs/develop/Database-Layout.md

@@ -3,8 +3,7 @@
 ### Import tables
 
 OSM data is initially imported using [osm2pgsql](https://osm2pgsql.org).
-Nominatim uses its own data output style 'gazetteer', which differs from the
-output style created for map rendering.
+Nominatim uses a custom flex style to create the initial import tables.
 
 The import process creates the following tables:
 
@@ -14,7 +13,7 @@ The `planet_osm_*` tables are the usual backing tables for OSM data. Note
 that Nominatim uses them to look up special relations and to find nodes on
 ways.
 
-The gazetteer style produces a single table `place` as output with the following
+The osm2pgsql import produces a single table `place` as output with the following
 columns:
 
  * `osm_type` - kind of OSM object (**N** - node, **W** - way, **R** - relation)

docs/develop/Development-Environment.md

@@ -25,15 +25,15 @@ following packages should get you started:
 
 ## Prerequisites for testing and documentation
 
-The Nominatim test suite consists of behavioural tests (using behave) and
+The Nominatim test suite consists of behavioural tests (using pytest-bdd) and
 unit tests (using pytest). It has the following additional requirements:
 
-* [behave test framework](https://behave.readthedocs.io) >= 1.2.6
 * [flake8](https://flake8.pycqa.org/en/stable/) (CI always runs the latest version from pip)
 * [mypy](http://mypy-lang.org/) (plus typing information for external libs)
 * [Python Typing Extensions](https://github.com/python/typing_extensions) (for Python < 3.9)
 * [pytest](https://pytest.org)
 * [pytest-asyncio](https://pytest-asyncio.readthedocs.io)
+* [pytest-bdd](https://pytest-bdd.readthedocs.io)
 
 For testing the Python search frontend, you need to install extra dependencies
 depending on your choice of webserver framework:
@@ -48,9 +48,6 @@ The documentation is built with mkdocs:
 * [mkdocs-material](https://squidfunk.github.io/mkdocs-material/)
 * [mkdocs-gen-files](https://oprypin.github.io/mkdocs-gen-files/)
 
-Please be aware that tests always run against the globally installed
-osm2pgsql, so you need to have this set up. If you want to test against
-the vendored version of osm2pgsql, you need to set the PATH accordingly.
 
 ### Installing prerequisites on Ubuntu/Debian
 
@@ -69,9 +66,10 @@ To set up the virtual environment with all necessary packages run:
 ```sh
 virtualenv ~/nominatim-dev-venv
 ~/nominatim-dev-venv/bin/pip install\
-    psutil psycopg[binary] PyICU SQLAlchemy \
-    python-dotenv jinja2 pyYAML datrie behave \
-    mkdocs mkdocstrings mkdocs-gen-files pytest pytest-asyncio flake8 \
+    psutil 'psycopg[binary]' PyICU SQLAlchemy \
+    python-dotenv jinja2 pyYAML \
+    mkdocs 'mkdocstrings[python]' mkdocs-gen-files \
+    pytest pytest-asyncio pytest-bdd flake8 \
     types-jinja2 types-markupsafe types-psutil types-psycopg2 \
     types-pygments types-pyyaml types-requests types-ujson \
     types-urllib3 typing-extensions unicorn falcon starlette \
@@ -94,7 +92,7 @@ but executes against the code in the source tree. For example:
 ```
 me@machine:~$ cd Nominatim
 me@machine:~Nominatim$ ./nominatim-cli.py --version
-Nominatim version 4.4.99-1
+Nominatim version 5.1.0-0
 ```
 
 Make sure you have activated the virtual environment holding all

docs/develop/ICU-Tokenizer-Modules.md

@@ -60,13 +60,19 @@ The order of phrases matters to Nominatim when doing further processing.
 Thus, while you may split or join phrases, you should not reorder them
 unless you really know what you are doing.
 
-Phrase types (`nominatim_api.search.PhraseType`) can further help narrowing
-down how the tokens in the phrase are interpreted. The following phrase types
-are known:
+Phrase types can further help narrowing down how the tokens in the phrase
+are interpreted. The following phrase types are known:
 
-::: nominatim_api.search.PhraseType
-    options:
-        heading_level: 6
+| Name           | Description |
+|----------------|-------------|
+| PHRASE_ANY     | No specific designation (i.e. source is free-form query) |
+| PHRASE_AMENITY | Contains name or type of a POI |
+| PHRASE_STREET  | Contains a street name optionally with a housenumber |
+| PHRASE_CITY    | Contains the postal city |
+| PHRASE_COUNTY  | Contains the equivalent of a county |
+| PHRASE_STATE   | Contains a state or province |
+| PHRASE_POSTCODE| Contains a postal code |
+| PHRASE_COUNTRY | Contains the country name or code |
 
 
 ## Custom sanitizer modules

docs/develop/Testing.md

@@ -43,53 +43,53 @@ The name of the pytest binary depends on your installation.
 ## BDD Functional Tests (`test/bdd`)
 
 Functional tests are written as BDD instructions. For more information on
-the philosophy of BDD testing, see the
-[Behave manual](http://pythonhosted.org/behave/philosophy.html).
-
-The following explanation assume that the reader is familiar with the BDD
-notations of features, scenarios and steps.
-
-All possible steps can be found in the `steps` directory and should ideally
-be documented.
+the philosophy of BDD testing, read the Wikipedia article on
+[Behaviour-driven development](https://en.wikipedia.org/wiki/Behavior-driven_development).
 
 ### General Usage
 
 To run the functional tests, do
 
-    cd test/bdd
-    behave
+    pytest test/bdd
 
-The tests can be configured with a set of environment variables (`behave -D key=val`):
+The BDD tests create databases for the tests. You can set name of the databases
+through configuration variables in your `pytest.ini`:
 
- * `TEMPLATE_DB` - name of template database used as a skeleton for
-                   the test databases (db tests)
- * `TEST_DB` - name of test database (db tests)
- * `API_TEST_DB` - name of the database containing the API test data (api tests)
- * `API_TEST_FILE` - OSM file to be imported into the API test database (api tests)
- * `API_ENGINE` - webframe to use for running search queries, same values as
-                  `nominatim serve --engine` parameter
- * `DB_HOST` - (optional) hostname of database host
- * `DB_PORT` - (optional) port of database on host
- * `DB_USER` - (optional) username of database login
- * `DB_PASS` - (optional) password for database login
- * `REMOVE_TEMPLATE` - if true, the template and API database will not be reused
-                       during the next run. Reusing the base templates speeds
-                       up tests considerably but might lead to outdated errors
-                       for some changes in the database layout.
- * `KEEP_TEST_DB` - if true, the test database will not be dropped after a test
-                    is finished. Should only be used if one single scenario is
-                    run, otherwise the result is undefined.
+ * `nominatim_test_db` defines the name of the temporary database created for
+    a single test (default: `test_nominatim`)
+ * `nominatim_api_test_db` defines the name of the database containing
+    the API test data, see also below (default: `test_api_nominatim`)
+ * `nominatim_template_db` defines the name of the template database used
+    for creating the temporary test databases. It contains some static setup
+    which usually doesn't change between imports of OSM data
+    (default: `test_template_nominatim`)
+
+To change other connection parameters for the PostgreSQL database, use
+the [libpq enivronment variables](https://www.postgresql.org/docs/current/libpq-envars.html).
+Never set a password through these variables. Use a
+[password file](https://www.postgresql.org/docs/current/libpq-pgpass.html) instead.
+
+The API test database and the template database are only created once and then
+left untouched. This is usually what you want because it speeds up subsequent
+runs of BDD tests. If you do change code that has an influence on the content
+of these databases, you can run pytest with the `--nominatim-purge` parameter
+and the databases will be dropped and recreated from scratch.
+
+When running the BDD tests with make (using `make tests` or `make bdd`), then
+the databases will always be purged.
+
+The temporary test database is usually dropped directly after the test, so
+it does not take up unnecessary space. If you want to keep the database around,
+for example while debugging a specific BDD test, use the parameter
+`--nominatim-keep-db`.
 
-Logging can be defined through command line parameters of behave itself. Check
-out `behave --help` for details. Also have a look at the 'work-in-progress'
-feature of behave which comes in handy when writing new tests.
 
 ### API Tests (`test/bdd/api`)
 
 These tests are meant to test the different API endpoints and their parameters.
 They require to import several datasets into a test database. This is normally
 done automatically during setup of the test. The API test database is then
-kept around and reused in subsequent runs of behave. Use `behave -DREMOVE_TEMPLATE`
+kept around and reused in subsequent runs of behave. Use `--nominatim-purge`
 to force a reimport of the database.
 
 The official test dataset is saved in the file `test/testdb/apidb-test-data.pbf`
@@ -109,12 +109,12 @@ test the correctness of osm2pgsql. Each test will write some data into the `plac
 table (and optionally the `planet_osm_*` tables if required) and then run
 Nominatim's processing functions on that.
 
-These tests need to create their own test databases. By default they will be
-called `test_template_nominatim` and `test_nominatim`. Names can be changed with
-the environment variables `TEMPLATE_DB` and `TEST_DB`. The user running the tests
-needs superuser rights for postgres.
+These tests use the template database and create temporary test databases for
+each test.
 
 ### Import Tests (`test/bdd/osm2pgsql`)
 
-These tests check that data is imported correctly into the place table. They
-use the same template database as the DB Creation tests, so the same remarks apply.
+These tests check that data is imported correctly into the place table.
+
+These tests also use the template database and create temporary test databases
+for each test.

docs/develop/overview.md

@@ -9,7 +9,7 @@ the address computation and the search frontend.
 The __data import__ stage reads the raw OSM data and extracts all information
 that is useful for geocoding. This part is done by osm2pgsql, the same tool
 that can also be used to import a rendering database. It uses the special
-gazetteer output plugin in `osm2pgsql/src/output-gazetter.[ch]pp`. The result of
+flex output style defined in the directory `/lib-lua`. The result of
 the import can be found in the database table `place`.
 
 The __address computation__ or __indexing__ stage takes the data from `place`

lib-lua/themes/nominatim/presets.lua

@@ -187,7 +187,7 @@ module.MAIN_TAGS_POIS = function (group)
                passing_place = group,
                street_lamp = 'named',
                traffic_signals = 'named'},
-    historic = {'always',
+    historic = {'fallback',
                 yes = group,
                 no = group},
     information = {include_when_tag_present('tourism', 'information'),
@@ -196,6 +196,7 @@ module.MAIN_TAGS_POIS = function (group)
                    trail_blaze = 'never'},
     junction = {'fallback',
                 no = group},
+    landuse = {cemetery = 'always'},
     leisure = {'always',
                nature_reserve = 'fallback',
                swimming_pool = 'named',
@@ -229,6 +230,7 @@ module.MAIN_TAGS_POIS = function (group)
     shop = {'always',
             no = group},
     tourism = {'always',
+               attraction = 'fallback',
                no = group,
                yes = group,
                information = exclude_when_key_present('information')},
@@ -330,7 +332,7 @@ module.NAME_TAGS.core = {main = {'name', 'name:*',
                         }
 module.NAME_TAGS.address = {house = {'addr:housename'}}
 module.NAME_TAGS.poi = group_merge({main = {'brand'},
-                                    extra = {'iata', 'icao'}},
+                                    extra = {'iata', 'icao', 'faa'}},
                                    module.NAME_TAGS.core)
 
 -- Address tagging

lib-sql/functions.sql

@@ -8,7 +8,6 @@
 {% include('functions/utils.sql') %}
 {% include('functions/ranking.sql') %}
 {% include('functions/importance.sql') %}
-{% include('functions/address_lookup.sql') %}
 {% include('functions/interpolation.sql') %}
 
 {% if 'place' in db.tables %}

lib-sql/functions/address_lookup.sql

@@ -1,334 +0,0 @@
--- SPDX-License-Identifier: GPL-2.0-only
---
--- This file is part of Nominatim. (https://nominatim.org)
---
--- Copyright (C) 2022 by the Nominatim developer community.
--- For a full list of authors see the git log.
-
--- Functions for returning address information for a place.
-
-DROP TYPE IF EXISTS addressline CASCADE;
-CREATE TYPE addressline as (
-  place_id BIGINT,
-  osm_type CHAR(1),
-  osm_id BIGINT,
-  name HSTORE,
-  class TEXT,
-  type TEXT,
-  place_type TEXT,
-  admin_level INTEGER,
-  fromarea BOOLEAN,
-  isaddress BOOLEAN,
-  rank_address INTEGER,
-  distance FLOAT
-);
-
-
-CREATE OR REPLACE FUNCTION get_name_by_language(name hstore, languagepref TEXT[])
-  RETURNS TEXT
-  AS $$
-DECLARE
-  result TEXT;
-BEGIN
-  IF name is null THEN
-    RETURN null;
-  END IF;
-
-  FOR j IN 1..array_upper(languagepref,1) LOOP
-    IF name ? languagepref[j] THEN
-      result := trim(name->languagepref[j]);
-      IF result != '' THEN
-        return result;
-      END IF;
-    END IF;
-  END LOOP;
-
-  -- as a fallback - take the last element since it is the default name
-  RETURN trim((avals(name))[array_length(avals(name), 1)]);
-END;
-$$
-LANGUAGE plpgsql IMMUTABLE;
-
-
---housenumber only needed for tiger data
-CREATE OR REPLACE FUNCTION get_address_by_language(for_place_id BIGINT,
-                                                   housenumber INTEGER,
-                                                   languagepref TEXT[])
-  RETURNS TEXT
-  AS $$
-DECLARE
-  result TEXT[];
-  currresult TEXT;
-  prevresult TEXT;
-  location RECORD;
-BEGIN
-
-  result := '{}';
-  prevresult := '';
-
-  FOR location IN
-    SELECT name,
-       CASE WHEN place_id = for_place_id THEN 99 ELSE rank_address END as rank_address
-    FROM get_addressdata(for_place_id, housenumber)
-    WHERE isaddress order by rank_address desc
-  LOOP
-    currresult := trim(get_name_by_language(location.name, languagepref));
-    IF currresult != prevresult AND currresult IS NOT NULL
-       AND result[(100 - location.rank_address)] IS NULL
-    THEN
-      result[(100 - location.rank_address)] := currresult;
-      prevresult := currresult;
-    END IF;
-  END LOOP;
-
-  RETURN array_to_string(result,', ');
-END;
-$$
-LANGUAGE plpgsql STABLE;
-
-DROP TYPE IF EXISTS addressdata_place;
-CREATE TYPE addressdata_place AS (
-  place_id BIGINT,
-  country_code VARCHAR(2),
-  housenumber TEXT,
-  postcode TEXT,
-  class TEXT,
-  type TEXT,
-  name HSTORE,
-  address HSTORE,
-  centroid GEOMETRY
-);
-
--- Compute the list of address parts for the given place.
---
--- If in_housenumber is greator or equal 0, look for an interpolation.
-CREATE OR REPLACE FUNCTION get_addressdata(in_place_id BIGINT, in_housenumber INTEGER)
-  RETURNS setof addressline
-  AS $$
-DECLARE
-  place addressdata_place;
-  location RECORD;
-  country RECORD;
-  current_rank_address INTEGER;
-  location_isaddress BOOLEAN;
-BEGIN
-  -- The place in question might not have a direct entry in place_addressline.
-  -- Look for the parent of such places then and save it in place.
-
-  -- first query osmline (interpolation lines)
-  IF in_housenumber >= 0 THEN
-    SELECT parent_place_id as place_id, country_code,
-           in_housenumber as housenumber, postcode,
-           'place' as class, 'house' as type,
-           null as name, null as address,
-           ST_Centroid(linegeo) as centroid
-      INTO place
-      FROM location_property_osmline
-      WHERE place_id = in_place_id
-            AND in_housenumber between startnumber and endnumber;
-  END IF;
-
-  --then query tiger data
-  {% if config.get_bool('USE_US_TIGER_DATA') %}
-  IF place IS NULL AND in_housenumber >= 0 THEN
-    SELECT parent_place_id as place_id, 'us' as country_code,
-           in_housenumber as housenumber, postcode,
-           'place' as class, 'house' as type,
-           null as name, null as address,
-           ST_Centroid(linegeo) as centroid
-      INTO place
-      FROM location_property_tiger
-      WHERE place_id = in_place_id
-            AND in_housenumber between startnumber and endnumber;
-  END IF;
-  {% endif %}
-
-  -- postcode table
-  IF place IS NULL THEN
-    SELECT parent_place_id as place_id, country_code,
-           null::text as housenumber, postcode,
-           'place' as class, 'postcode' as type,
-           null as name, null as address,
-           null as centroid
-      INTO place
-      FROM location_postcode
-      WHERE place_id = in_place_id;
-  END IF;
-
-  -- POI objects in the placex table
-  IF place IS NULL THEN
-    SELECT parent_place_id as place_id, country_code,
-           coalesce(address->'housenumber',
-                    address->'streetnumber',
-                    address->'conscriptionnumber')::text as housenumber,
-           postcode,
-           class, type,
-           name, address,
-           centroid
-      INTO place
-      FROM placex
-      WHERE place_id = in_place_id and rank_search > 27;
-  END IF;
-
-  -- If place is still NULL at this point then the object has its own
-  -- entry in place_address line. However, still check if there is not linked
-  -- place we should be using instead.
-  IF place IS NULL THEN
-    select coalesce(linked_place_id, place_id) as place_id,  country_code,
-           null::text as housenumber, postcode,
-           class, type,
-           null as name, address,
-           null as centroid
-      INTO place
-      FROM placex where place_id = in_place_id;
-  END IF;
-
---RAISE WARNING '% % % %',searchcountrycode, searchhousenumber, searchpostcode;
-
-  -- --- Return the record for the base entry.
-
-  current_rank_address := 1000;
-  FOR location IN
-    SELECT placex.place_id, osm_type, osm_id, name,
-           coalesce(extratags->'linked_place', extratags->'place') as place_type,
-           class, type, admin_level,
-           CASE WHEN rank_address = 0 THEN 100
-                WHEN rank_address = 11 THEN 5
-                ELSE rank_address END as rank_address,
-           country_code
-      FROM placex
-      WHERE place_id = place.place_id
-  LOOP
---RAISE WARNING '%',location;
-    -- mix in default names for countries
-    IF location.rank_address = 4 and place.country_code is not NULL THEN
-      FOR country IN
-        SELECT coalesce(name, ''::hstore) as name FROM country_name
-          WHERE country_code = place.country_code LIMIT 1
-      LOOP
-        place.name := country.name || place.name;
-      END LOOP;
-    END IF;
-
-    IF location.rank_address < 4 THEN
-      -- no country locations for ranks higher than country
-      place.country_code := NULL::varchar(2);
-    ELSEIF place.country_code IS NULL AND location.country_code IS NOT NULL THEN
-      place.country_code := location.country_code;
-    END IF;
-
-    RETURN NEXT ROW(location.place_id, location.osm_type, location.osm_id,
-                    location.name, location.class, location.type,
-                    location.place_type,
-                    location.admin_level, true,
-                    location.type not in ('postcode', 'postal_code'),
-                    location.rank_address, 0)::addressline;
-
-    current_rank_address := location.rank_address;
-  END LOOP;
-
-  -- --- Return records for address parts.
-
-  FOR location IN
-    SELECT placex.place_id, osm_type, osm_id, name, class, type,
-           coalesce(extratags->'linked_place', extratags->'place') as place_type,
-           admin_level, fromarea, isaddress,
-           CASE WHEN rank_address = 11 THEN 5 ELSE rank_address END as rank_address,
-           distance, country_code, postcode
-      FROM place_addressline join placex on (address_place_id = placex.place_id)
-      WHERE place_addressline.place_id IN (place.place_id, in_place_id)
-            AND linked_place_id is null
-            AND (placex.country_code IS NULL OR place.country_code IS NULL
-                 OR placex.country_code = place.country_code)
-      ORDER BY rank_address desc,
-               (place_addressline.place_id = in_place_id) desc,
-               (CASE WHEN coalesce((avals(name) && avals(place.address)), False) THEN 2
-                     WHEN isaddress THEN 0
-                     WHEN fromarea
-                          and place.centroid is not null
-                          and ST_Contains(geometry, place.centroid) THEN 1
-                     ELSE -1 END) desc,
-               fromarea desc, distance asc, rank_search desc
-  LOOP
-    -- RAISE WARNING '%',location;
-    location_isaddress := location.rank_address != current_rank_address;
-
-    IF place.country_code IS NULL AND location.country_code IS NOT NULL THEN
-      place.country_code := location.country_code;
-    END IF;
-    IF location.type in ('postcode', 'postal_code')
-       AND place.postcode is not null
-    THEN
-      -- If the place had a postcode assigned, take this one only
-      -- into consideration when it is an area and the place does not have
-      -- a postcode itself.
-      IF location.fromarea AND location_isaddress
-         AND (place.address is null or not place.address ? 'postcode')
-      THEN
-        place.postcode := null; -- remove the less exact postcode
-      ELSE
-        location_isaddress := false;
-      END IF;
-    END IF;
-    RETURN NEXT ROW(location.place_id, location.osm_type, location.osm_id,
-                    location.name, location.class, location.type,
-                    location.place_type,
-                    location.admin_level, location.fromarea,
-                    location_isaddress,
-                    location.rank_address,
-                    location.distance)::addressline;
-
-    current_rank_address := location.rank_address;
-  END LOOP;
-
-  -- If no country was included yet, add the name information from country_name.
-  IF current_rank_address > 4 THEN
-    FOR location IN
-      SELECT name || coalesce(derived_name, ''::hstore) as name FROM country_name
-        WHERE country_code = place.country_code LIMIT 1
-    LOOP
---RAISE WARNING '% % %',current_rank_address,searchcountrycode,countryname;
-      RETURN NEXT ROW(null, null, null, location.name, 'place', 'country', NULL,
-                      null, true, true, 4, 0)::addressline;
-    END LOOP;
-  END IF;
-
-  -- Finally add some artificial rows.
-  IF place.country_code IS NOT NULL THEN
-    location := ROW(null, null, null, hstore('ref', place.country_code),
-                    'place', 'country_code', null, null, true, false, 4, 0)::addressline;
-    RETURN NEXT location;
-  END IF;
-
-  IF place.name IS NOT NULL THEN
-    location := ROW(in_place_id, null, null, place.name, place.class,
-                    place.type, null, null, true, true, 29, 0)::addressline;
-    RETURN NEXT location;
-  END IF;
-
-  IF place.housenumber IS NOT NULL THEN
-    location := ROW(null, null, null, hstore('ref', place.housenumber),
-                    'place', 'house_number', null, null, true, true, 28, 0)::addressline;
-    RETURN NEXT location;
-  END IF;
-
-  IF place.address is not null and place.address ? '_unlisted_place' THEN
-    RETURN NEXT ROW(null, null, null, hstore('name', place.address->'_unlisted_place'),
-                    'place', 'locality', null, null, true, true, 25, 0)::addressline;
-  END IF;
-
-  IF place.postcode is not null THEN
-    location := ROW(null, null, null, hstore('ref', place.postcode), 'place',
-                    'postcode', null, null, false, true, 5, 0)::addressline;
-    RETURN NEXT location;
-  ELSEIF place.address is not null and place.address ? 'postcode'
-         and not place.address->'postcode' SIMILAR TO '%(,|;)%' THEN
-    location := ROW(null, null, null, hstore('ref', place.address->'postcode'), 'place',
-                    'postcode', null, null, false, true, 5, 0)::addressline;
-    RETURN NEXT location;
-  END IF;
-
-  RETURN;
-END;
-$$
-LANGUAGE plpgsql STABLE;

lib-sql/functions/importance.sql

@@ -65,7 +65,7 @@ BEGIN
   RETURN NULL;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 {% else %}
 
@@ -78,7 +78,7 @@ SELECT convert_from(CAST(E'\\x' || array_to_string(ARRAY(
     FROM regexp_matches($1, '%[0-9a-f][0-9a-f]|.', 'gi') AS r(m)
 ), '') AS bytea), 'UTF8');
 $$ 
-LANGUAGE SQL IMMUTABLE STRICT;
+LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION catch_decode_url_part(p varchar)
@@ -91,7 +91,7 @@ EXCEPTION
   WHEN others THEN return null;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE STRICT;
+LANGUAGE plpgsql IMMUTABLE STRICT PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION get_wikipedia_match(extratags HSTORE, country_code varchar(2))
@@ -139,7 +139,7 @@ BEGIN
   RETURN NULL;
 END;
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 {% endif %}
 
@@ -203,5 +203,5 @@ BEGIN
   RETURN result;
 END;
 $$
-LANGUAGE plpgsql;
+LANGUAGE plpgsql PARALLEL SAFE;
 

lib-sql/functions/interpolation.sql

@@ -34,7 +34,7 @@ BEGIN
   RETURN in_address;
 END;
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 
 
@@ -70,7 +70,7 @@ BEGIN
   RETURN parent_place_id;
 END;
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION reinsert_interpolation(way_id BIGINT, addr HSTORE,
@@ -309,7 +309,7 @@ BEGIN
         IF NEW.startnumber IS NULL THEN
             NEW.startnumber := startnumber;
             NEW.endnumber := endnumber;
-            NEW.linegeo := sectiongeo;
+            NEW.linegeo := ST_ReducePrecision(sectiongeo, 0.0000001);
             NEW.postcode := postcode;
         ELSE
           INSERT INTO location_property_osmline
@@ -317,7 +317,8 @@ BEGIN
                   startnumber, endnumber, step,
                   address, postcode, country_code,
                   geometry_sector, indexed_status)
-          VALUES (sectiongeo, NEW.partition, NEW.osm_id, NEW.parent_place_id,
+          VALUES (ST_ReducePrecision(sectiongeo, 0.0000001),
+                  NEW.partition, NEW.osm_id, NEW.parent_place_id,
                   startnumber, endnumber, NEW.step,
                   NEW.address, postcode,
                   NEW.country_code, NEW.geometry_sector, 0);

lib-sql/functions/partition-functions.sql

@@ -58,7 +58,7 @@ BEGIN
   RAISE EXCEPTION 'Unknown partition %', in_partition;
 END
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION get_address_place(in_partition SMALLINT, feature GEOMETRY,
@@ -87,7 +87,7 @@ BEGIN
   RAISE EXCEPTION 'Unknown partition %', in_partition;
 END;
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 
 create or replace function deleteLocationArea(in_partition INTEGER, in_place_id BIGINT, in_rank_search INTEGER) RETURNS BOOLEAN AS $$
@@ -172,7 +172,7 @@ BEGIN
   RAISE EXCEPTION 'Unknown partition %', in_partition;
 END
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 CREATE OR REPLACE FUNCTION getNearestNamedPlacePlaceId(in_partition INTEGER,
                                                        point GEOMETRY,
@@ -202,7 +202,7 @@ BEGIN
   RAISE EXCEPTION 'Unknown partition %', in_partition;
 END
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 create or replace function insertSearchName(
   in_partition INTEGER, in_place_id BIGINT, in_name_vector INTEGER[],
@@ -310,7 +310,7 @@ BEGIN
   RAISE EXCEPTION 'Unknown partition %', in_partition;
 END
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 CREATE OR REPLACE FUNCTION getNearestParallelRoadFeature(in_partition INTEGER,
                                                          line GEOMETRY)
@@ -354,4 +354,4 @@ BEGIN
   RAISE EXCEPTION 'Unknown partition %', in_partition;
 END
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;

lib-sql/functions/placex_triggers.sql

@@ -2,7 +2,7 @@
 --
 -- This file is part of Nominatim. (https://nominatim.org)
 --
--- Copyright (C) 2024 by the Nominatim developer community.
+-- Copyright (C) 2025 by the Nominatim developer community.
 -- For a full list of authors see the git log.
 
 -- Trigger functions for the placex table.
@@ -109,7 +109,7 @@ BEGIN
   RETURN result;
 END;
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION find_associated_street(poi_osm_type CHAR(1),
@@ -200,7 +200,7 @@ BEGIN
   RETURN result;
 END;
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 
 -- Find the parent road of a POI.
@@ -286,7 +286,7 @@ BEGIN
   RETURN parent_place_id;
 END;
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 -- Try to find a linked place for the given object.
 CREATE OR REPLACE FUNCTION find_linked_place(bnd placex)
@@ -404,7 +404,7 @@ BEGIN
   RETURN NULL;
 END;
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION create_poi_search_terms(obj_place_id BIGINT,
@@ -638,8 +638,10 @@ BEGIN
 
     -- Add it to the list of search terms
     {% if not db.reverse_only %}
-      nameaddress_vector := array_merge(nameaddress_vector,
-                                        location.keywords::integer[]);
+      IF location.rank_address != 11 AND location.rank_address != 5 THEN
+          nameaddress_vector := array_merge(nameaddress_vector,
+                                            location.keywords::integer[]);
+      END IF;
     {% endif %}
 
     INSERT INTO place_addressline (place_id, address_place_id, fromarea,
@@ -730,7 +732,7 @@ BEGIN
   IF NEW.rank_address between 2 and 27 THEN
     IF (ST_GeometryType(NEW.geometry) in ('ST_Polygon','ST_MultiPolygon') AND ST_IsValid(NEW.geometry)) THEN
       -- Performance: We just can't handle re-indexing for country level changes
-      IF (NEW.rank_address < 26 and st_area(NEW.geometry) < 1)
+      IF (NEW.rank_address < 26 and st_area(NEW.geometry) <= 2)
          OR (NEW.rank_address >= 26 and st_area(NEW.geometry) < 0.01)
       THEN
         -- mark items within the geometry for re-indexing
@@ -778,7 +780,7 @@ BEGIN
   SELECT count(*)>0 FROM pg_tables WHERE tablename = classtable and schemaname = current_schema() INTO result;
   IF result THEN
     EXECUTE 'INSERT INTO ' || classtable::regclass || ' (place_id, centroid) VALUES ($1,$2)' 
-    USING NEW.place_id, ST_Centroid(NEW.geometry);
+    USING NEW.place_id, NEW.centroid;
   END IF;
 
 {% endif %} -- not disable_diff_updates
@@ -960,9 +962,8 @@ BEGIN
           WHERE class = 'place' and rank_address between 1 and 23
                 and prank.address_rank >= NEW.rank_address
                 and ST_GeometryType(geometry) in ('ST_Polygon','ST_MultiPolygon') -- select right index
-                and geometry && NEW.geometry
-                and geometry ~ NEW.geometry -- needed because ST_Relate does not do bbox cover test
-                and ST_Relate(geometry, NEW.geometry, 'T*T***FF*') -- contains but not equal
+                and ST_Contains(geometry, NEW.geometry)
+                and not ST_Equals(geometry, NEW.geometry)
           ORDER BY prank.address_rank desc LIMIT 1
         LOOP
           NEW.rank_address := location.rank_address + 2;
@@ -983,9 +984,8 @@ BEGIN
                 and rank_address between 1 and 25 -- select right index
                 and ST_GeometryType(geometry) in ('ST_Polygon','ST_MultiPolygon') -- select right index
                 and prank.address_rank >= NEW.rank_address
-                and geometry && NEW.geometry
-                and geometry ~ NEW.geometry -- needed because ST_Relate does not do bbox cover test
-                and ST_Relate(geometry, NEW.geometry, 'T*T***FF*') -- contains but not equal
+                and ST_Contains(geometry, NEW.geometry)
+                and not ST_Equals(geometry, NEW.geometry)
           ORDER BY prank.address_rank desc LIMIT 1
         LOOP
           NEW.rank_address := location.rank_address + 2;

lib-sql/functions/ranking.sql

@@ -29,7 +29,7 @@ BEGIN
   RETURN 0.02;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 
 -- Return an approximate update radius according to the search rank.
@@ -60,7 +60,7 @@ BEGIN
   RETURN 0;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 -- Compute a base address rank from the extent of the given geometry.
 --
@@ -88,6 +88,10 @@ BEGIN
     area := area / 3;
   ELSIF country_code IN ('bo', 'ar', 'sd', 'mn', 'in', 'et', 'cd', 'mz', 'ly', 'cl', 'zm') THEN
     area := area / 2;
+  ELSIF country_code IN ('sg', 'ws', 'st', 'kn') THEN
+    area := area * 5;
+  ELSIF country_code IN ('dm', 'mt', 'lc', 'gg', 'sc', 'nr') THEN
+    area := area * 20;
   END IF;
 
   IF area > 1 THEN
@@ -107,7 +111,7 @@ BEGIN
    RETURN 23;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 
 -- Guess a ranking for postcodes from country and postcode format.
@@ -167,7 +171,7 @@ BEGIN
 
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 
 -- Get standard search and address rank for an object.
@@ -236,7 +240,7 @@ BEGIN
   END IF;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 CREATE OR REPLACE FUNCTION get_addr_tag_rank(key TEXT, country TEXT,
                                              OUT from_rank SMALLINT,
@@ -283,7 +287,7 @@ BEGIN
   END LOOP;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION weigh_search(search_vector INT[],
@@ -304,4 +308,4 @@ BEGIN
   RETURN def_weight;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;

lib-sql/functions/utils.sql

@@ -2,7 +2,7 @@
 --
 -- This file is part of Nominatim. (https://nominatim.org)
 --
--- Copyright (C) 2022 by the Nominatim developer community.
+-- Copyright (C) 2025 by the Nominatim developer community.
 -- For a full list of authors see the git log.
 
 -- Assorted helper functions for the triggers.
@@ -14,17 +14,17 @@ DECLARE
   geom_type TEXT;
 BEGIN
   geom_type := ST_GeometryType(place);
-  IF geom_type = ' ST_Point' THEN
+  IF geom_type = 'ST_Point' THEN
     RETURN place;
   END IF;
   IF geom_type = 'ST_LineString' THEN
-    RETURN ST_LineInterpolatePoint(place, 0.5);
+    RETURN ST_ReducePrecision(ST_LineInterpolatePoint(place, 0.5), 0.0000001);
   END IF;
 
-  RETURN ST_PointOnSurface(place);
+  RETURN ST_ReducePrecision(ST_PointOnSurface(place), 0.0000001);
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION geometry_sector(partition INTEGER, place GEOMETRY)
@@ -34,7 +34,7 @@ BEGIN
   RETURN (partition*1000000) + (500-ST_X(place)::INTEGER)*1000 + (500-ST_Y(place)::INTEGER);
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 
 
@@ -60,7 +60,7 @@ BEGIN
   RETURN r;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 -- Return the node members with a given label from a relation member list
 -- as a set.
@@ -88,7 +88,7 @@ BEGIN
   RETURN;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION get_rel_node_members(members JSONB, memberLabels TEXT[])
@@ -107,7 +107,7 @@ BEGIN
   RETURN;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 
 -- Copy 'name' to or from the default language.
@@ -136,7 +136,7 @@ BEGIN
   END IF;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 
 -- Find the nearest artificial postcode for the given geometry.
@@ -172,7 +172,7 @@ BEGIN
     RETURN outcode;
 END;
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION get_country_code(place geometry)
@@ -233,7 +233,7 @@ BEGIN
   RETURN NULL;
 END;
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION get_country_language_code(search_country_code VARCHAR(2))
@@ -251,7 +251,7 @@ BEGIN
   RETURN NULL;
 END;
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION get_partition(in_country_code VARCHAR(10))
@@ -268,7 +268,7 @@ BEGIN
   RETURN 0;
 END;
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 
 -- Find the parent of an address with addr:street/addr:place tag.
@@ -299,7 +299,7 @@ BEGIN
   RETURN parent_place_id;
 END;
 $$
-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION delete_location(OLD_place_id BIGINT)
@@ -337,7 +337,7 @@ BEGIN
                      ST_Project(geom::geography, radius, 3.9269908)::geometry));
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION add_location(place_id BIGINT, country_code varchar(2),
@@ -455,7 +455,7 @@ BEGIN
   RETURN;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION split_geometry(geometry GEOMETRY)
@@ -483,7 +483,7 @@ BEGIN
   RETURN;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 CREATE OR REPLACE FUNCTION simplify_large_polygons(geometry GEOMETRY)
   RETURNS GEOMETRY
@@ -497,7 +497,7 @@ BEGIN
   RETURN geometry;
 END;
 $$
-LANGUAGE plpgsql IMMUTABLE;
+LANGUAGE plpgsql IMMUTABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION place_force_delete(placeid BIGINT)

lib-sql/tokenizer/icu_tokenizer.sql

@@ -12,7 +12,7 @@ CREATE OR REPLACE FUNCTION token_get_name_search_tokens(info JSONB)
   RETURNS INTEGER[]
 AS $$
   SELECT (info->>'names')::INTEGER[]
-$$ LANGUAGE SQL IMMUTABLE STRICT;
+$$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 
 
 -- Get tokens for matching the place name against others.
@@ -22,7 +22,7 @@ CREATE OR REPLACE FUNCTION token_get_name_match_tokens(info JSONB)
   RETURNS INTEGER[]
 AS $$
   SELECT (info->>'names')::INTEGER[]
-$$ LANGUAGE SQL IMMUTABLE STRICT;
+$$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 
 
 -- Return the housenumber tokens applicable for the place.
@@ -30,7 +30,7 @@ CREATE OR REPLACE FUNCTION token_get_housenumber_search_tokens(info JSONB)
   RETURNS INTEGER[]
 AS $$
   SELECT (info->>'hnr_tokens')::INTEGER[]
-$$ LANGUAGE SQL IMMUTABLE STRICT;
+$$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 
 
 -- Return the housenumber in the form that it can be matched during search.
@@ -38,77 +38,77 @@ CREATE OR REPLACE FUNCTION token_normalized_housenumber(info JSONB)
   RETURNS TEXT
 AS $$
   SELECT info->>'hnr';
-$$ LANGUAGE SQL IMMUTABLE STRICT;
+$$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION token_is_street_address(info JSONB)
   RETURNS BOOLEAN
 AS $$
   SELECT info->>'street' is not null or info->>'place' is null;
-$$ LANGUAGE SQL IMMUTABLE;
+$$ LANGUAGE SQL IMMUTABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION token_has_addr_street(info JSONB)
   RETURNS BOOLEAN
 AS $$
   SELECT info->>'street' is not null and info->>'street' != '{}';
-$$ LANGUAGE SQL IMMUTABLE;
+$$ LANGUAGE SQL IMMUTABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION token_has_addr_place(info JSONB)
   RETURNS BOOLEAN
 AS $$
   SELECT info->>'place' is not null;
-$$ LANGUAGE SQL IMMUTABLE;
+$$ LANGUAGE SQL IMMUTABLE PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION token_matches_street(info JSONB, street_tokens INTEGER[])
   RETURNS BOOLEAN
 AS $$
   SELECT (info->>'street')::INTEGER[] && street_tokens
-$$ LANGUAGE SQL IMMUTABLE STRICT;
+$$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION token_matches_place(info JSONB, place_tokens INTEGER[])
   RETURNS BOOLEAN
 AS $$
   SELECT (info->>'place')::INTEGER[] <@ place_tokens
-$$ LANGUAGE SQL IMMUTABLE STRICT;
+$$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION token_addr_place_search_tokens(info JSONB)
   RETURNS INTEGER[]
 AS $$
   SELECT (info->>'place')::INTEGER[]
-$$ LANGUAGE SQL IMMUTABLE STRICT;
+$$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION token_get_address_keys(info JSONB)
   RETURNS SETOF TEXT
 AS $$
   SELECT * FROM jsonb_object_keys(info->'addr');
-$$ LANGUAGE SQL IMMUTABLE STRICT;
+$$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION token_get_address_search_tokens(info JSONB, key TEXT)
   RETURNS INTEGER[]
 AS $$
   SELECT (info->'addr'->>key)::INTEGER[];
-$$ LANGUAGE SQL IMMUTABLE STRICT;
+$$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION token_matches_address(info JSONB, key TEXT, tokens INTEGER[])
   RETURNS BOOLEAN
 AS $$
   SELECT (info->'addr'->>key)::INTEGER[] <@ tokens;
-$$ LANGUAGE SQL IMMUTABLE STRICT;
+$$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 
 
 CREATE OR REPLACE FUNCTION token_get_postcode(info JSONB)
   RETURNS TEXT
 AS $$
   SELECT info->>'postcode';
-$$ LANGUAGE SQL IMMUTABLE STRICT;
+$$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 
 
 -- Return token info that should be saved permanently in the database.
@@ -116,7 +116,7 @@ CREATE OR REPLACE FUNCTION token_strip_info(info JSONB)
   RETURNS JSONB
 AS $$
   SELECT NULL::JSONB;
-$$ LANGUAGE SQL IMMUTABLE STRICT;
+$$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 
 --------------- private functions ----------------------------------------------
 
@@ -128,16 +128,14 @@ DECLARE
   partial_terms TEXT[] = '{}'::TEXT[];
   term TEXT;
   term_id INTEGER;
-  term_count INTEGER;
 BEGIN
   SELECT min(word_id) INTO full_token
     FROM word WHERE word = norm_term and type = 'W';
 
   IF full_token IS NULL THEN
     full_token := nextval('seq_word');
-    INSERT INTO word (word_id, word_token, type, word, info)
-      SELECT full_token, lookup_term, 'W', norm_term,
-             json_build_object('count', 0)
+    INSERT INTO word (word_id, word_token, type, word)
+      SELECT full_token, lookup_term, 'W', norm_term
         FROM unnest(lookup_terms) as lookup_term;
   END IF;
 
@@ -150,14 +148,67 @@ BEGIN
 
   partial_tokens := '{}'::INT[];
   FOR term IN SELECT unnest(partial_terms) LOOP
-    SELECT min(word_id), max(info->>'count') INTO term_id, term_count
+    SELECT min(word_id) INTO term_id
       FROM word WHERE word_token = term and type = 'w';
 
     IF term_id IS NULL THEN
       term_id := nextval('seq_word');
-      term_count := 0;
-      INSERT INTO word (word_id, word_token, type, info)
-        VALUES (term_id, term, 'w', json_build_object('count', term_count));
+      INSERT INTO word (word_id, word_token, type)
+        VALUES (term_id, term, 'w');
+    END IF;
+
+    partial_tokens := array_merge(partial_tokens, ARRAY[term_id]);
+  END LOOP;
+END;
+$$
+LANGUAGE plpgsql;
+
+
+CREATE OR REPLACE FUNCTION getorcreate_full_word(norm_term TEXT,
+                                                 lookup_terms TEXT[],
+                                                 lookup_norm_terms TEXT[],
+                                                 OUT full_token INT,
+                                                 OUT partial_tokens INT[])
+  AS $$
+DECLARE
+  partial_terms TEXT[] = '{}'::TEXT[];
+  term TEXT;
+  term_id INTEGER;
+BEGIN
+  SELECT min(word_id) INTO full_token
+    FROM word WHERE word = norm_term and type = 'W';
+
+  IF full_token IS NULL THEN
+    full_token := nextval('seq_word');
+    IF lookup_norm_terms IS NULL THEN
+      INSERT INTO word (word_id, word_token, type, word)
+        SELECT full_token, lookup_term, 'W', norm_term
+          FROM unnest(lookup_terms) as lookup_term;
+    ELSE
+      INSERT INTO word (word_id, word_token, type, word, info)
+        SELECT full_token, t.lookup, 'W', norm_term,
+               CASE WHEN norm_term = t.norm THEN null
+               ELSE json_build_object('lookup', t.norm) END
+          FROM unnest(lookup_terms, lookup_norm_terms) as t(lookup, norm);
+    END IF;
+  END IF;
+
+  FOR term IN SELECT unnest(string_to_array(unnest(lookup_terms), ' ')) LOOP
+    term := trim(term);
+    IF NOT (ARRAY[term] <@ partial_terms) THEN
+      partial_terms := partial_terms || term;
+    END IF;
+  END LOOP;
+
+  partial_tokens := '{}'::INT[];
+  FOR term IN SELECT unnest(partial_terms) LOOP
+    SELECT min(word_id) INTO term_id
+      FROM word WHERE word_token = term and type = 'w';
+
+    IF term_id IS NULL THEN
+      term_id := nextval('seq_word');
+      INSERT INTO word (word_id, word_token, type)
+        VALUES (term_id, term, 'w');
     END IF;
 
     partial_tokens := array_merge(partial_tokens, ARRAY[term_id]);

nominatim-cli.py

@@ -3,7 +3,7 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Helper script for development to run nominatim from the source directory.
@@ -15,4 +15,4 @@ sys.path.insert(1, str((Path(__file__) / '..' / 'src').resolve()))
 
 from nominatim_db import cli
 
-exit(cli.nominatim(module_dir=None, osm2pgsql_path=None))
+exit(cli.nominatim())

packaging/nominatim-api/COPYING

@@ -1 +0,0 @@
-../../COPYING

packaging/nominatim-api/COPYING

@@ -0,0 +1,232 @@
+GNU GENERAL PUBLIC LICENSE
+Version 3, 29 June 2007
+
+Copyright © 2007 Free Software Foundation, Inc. <https://fsf.org/>
+
+Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed.
+
+Preamble
+
+The GNU General Public License is a free, copyleft license for software and other kinds of works.
+
+The licenses for most software and other practical works are designed to take away your freedom to share and change the works. By contrast, the GNU General Public License is intended to guarantee your freedom to share and change all versions of a program--to make sure it remains free software for all its users. We, the Free Software Foundation, use the GNU General Public License for most of our software; it applies also to any other work released this way by its authors. You can apply it to your programs, too.
+
+When we speak of free software, we are referring to freedom, not price. Our General Public Licenses are designed to make sure that you have the freedom to distribute copies of free software (and charge for them if you wish), that you receive source code or can get it if you want it, that you can change the software or use pieces of it in new free programs, and that you know you can do these things.
+
+To protect your rights, we need to prevent others from denying you these rights or asking you to surrender the rights. Therefore, you have certain responsibilities if you distribute copies of the software, or if you modify it: responsibilities to respect the freedom of others.
+
+For example, if you distribute copies of such a program, whether gratis or for a fee, you must pass on to the recipients the same freedoms that you received. You must make sure that they, too, receive or can get the source code. And you must show them these terms so they know their rights.
+
+Developers that use the GNU GPL protect your rights with two steps: (1) assert copyright on the software, and (2) offer you this License giving you legal permission to copy, distribute and/or modify it.
+
+For the developers' and authors' protection, the GPL clearly explains that there is no warranty for this free software. For both users' and authors' sake, the GPL requires that modified versions be marked as changed, so that their problems will not be attributed erroneously to authors of previous versions.
+
+Some devices are designed to deny users access to install or run modified versions of the software inside them, although the manufacturer can do so. This is fundamentally incompatible with the aim of protecting users' freedom to change the software. The systematic pattern of such abuse occurs in the area of products for individuals to use, which is precisely where it is most unacceptable. Therefore, we have designed this version of the GPL to prohibit the practice for those products. If such problems arise substantially in other domains, we stand ready to extend this provision to those domains in future versions of the GPL, as needed to protect the freedom of users.
+
+Finally, every program is threatened constantly by software patents. States should not allow patents to restrict development and use of software on general-purpose computers, but in those that do, we wish to avoid the special danger that patents applied to a free program could make it effectively proprietary. To prevent this, the GPL assures that patents cannot be used to render the program non-free.
+
+The precise terms and conditions for copying, distribution and modification follow.
+
+TERMS AND CONDITIONS
+
+0. Definitions.
+
+“This License” refers to version 3 of the GNU General Public License.
+
+“Copyright” also means copyright-like laws that apply to other kinds of works, such as semiconductor masks.
+
+“The Program” refers to any copyrightable work licensed under this License. Each licensee is addressed as “you”. “Licensees” and “recipients” may be individuals or organizations.
+
+To “modify” a work means to copy from or adapt all or part of the work in a fashion requiring copyright permission, other than the making of an exact copy. The resulting work is called a “modified version” of the earlier work or a work “based on” the earlier work.
+
+A “covered work” means either the unmodified Program or a work based on the Program.
+
+To “propagate” a work means to do anything with it that, without permission, would make you directly or secondarily liable for infringement under applicable copyright law, except executing it on a computer or modifying a private copy. Propagation includes copying, distribution (with or without modification), making available to the public, and in some countries other activities as well.
+
+To “convey” a work means any kind of propagation that enables other parties to make or receive copies. Mere interaction with a user through a computer network, with no transfer of a copy, is not conveying.
+
+An interactive user interface displays “Appropriate Legal Notices” to the extent that it includes a convenient and prominently visible feature that (1) displays an appropriate copyright notice, and (2) tells the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to view a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion.
+
+1. Source Code.
+The “source code” for a work means the preferred form of the work for making modifications to it. “Object code” means any non-source form of a work.
+
+A “Standard Interface” means an interface that either is an official standard defined by a recognized standards body, or, in the case of interfaces specified for a particular programming language, one that is widely used among developers working in that language.
+
+The “System Libraries” of an executable work include anything, other than the work as a whole, that (a) is included in the normal form of packaging a Major Component, but which is not part of that Major Component, and (b) serves only to enable use of the work with that Major Component, or to implement a Standard Interface for which an implementation is available to the public in source code form. A “Major Component”, in this context, means a major essential component (kernel, window system, and so on) of the specific operating system (if any) on which the executable work runs, or a compiler used to produce the work, or an object code interpreter used to run it.
+
+The “Corresponding Source” for a work in object code form means all the source code needed to generate, install, and (for an executable work) run the object code and to modify the work, including scripts to control those activities. However, it does not include the work's System Libraries, or general-purpose tools or generally available free programs which are used unmodified in performing those activities but which are not part of the work. For example, Corresponding Source includes interface definition files associated with source files for the work, and the source code for shared libraries and dynamically linked subprograms that the work is specifically designed to require, such as by intimate data communication or control flow between those subprograms and other parts of the work.
+
+The Corresponding Source need not include anything that users can regenerate automatically from other parts of the Corresponding Source.
+
+The Corresponding Source for a work in source code form is that same work.
+
+2. Basic Permissions.
+All rights granted under this License are granted for the term of copyright on the Program, and are irrevocable provided the stated conditions are met. This License explicitly affirms your unlimited permission to run the unmodified Program. The output from running a covered work is covered by this License only if the output, given its content, constitutes a covered work. This License acknowledges your rights of fair use or other equivalent, as provided by copyright law.
+
+You may make, run and propagate covered works that you do not convey, without conditions so long as your license otherwise remains in force. You may convey covered works to others for the sole purpose of having them make modifications exclusively for you, or provide you with facilities for running those works, provided that you comply with the terms of this License in conveying all material for which you do not control copyright. Those thus making or running the covered works for you must do so exclusively on your behalf, under your direction and control, on terms that prohibit them from making any copies of your copyrighted material outside their relationship with you.
+
+Conveying under any other circumstances is permitted solely under the conditions stated below. Sublicensing is not allowed; section 10 makes it unnecessary.
+
+3. Protecting Users' Legal Rights From Anti-Circumvention Law.
+No covered work shall be deemed part of an effective technological measure under any applicable law fulfilling obligations under article 11 of the WIPO copyright treaty adopted on 20 December 1996, or similar laws prohibiting or restricting circumvention of such measures.
+
+When you convey a covered work, you waive any legal power to forbid circumvention of technological measures to the extent such circumvention is effected by exercising rights under this License with respect to the covered work, and you disclaim any intention to limit operation or modification of the work as a means of enforcing, against the work's users, your or third parties' legal rights to forbid circumvention of technological measures.
+
+4. Conveying Verbatim Copies.
+You may convey verbatim copies of the Program's source code as you receive it, in any medium, provided that you conspicuously and appropriately publish on each copy an appropriate copyright notice; keep intact all notices stating that this License and any non-permissive terms added in accord with section 7 apply to the code; keep intact all notices of the absence of any warranty; and give all recipients a copy of this License along with the Program.
+
+You may charge any price or no price for each copy that you convey, and you may offer support or warranty protection for a fee.
+
+5. Conveying Modified Source Versions.
+You may convey a work based on the Program, or the modifications to produce it from the Program, in the form of source code under the terms of section 4, provided that you also meet all of these conditions:
+
+     a) The work must carry prominent notices stating that you modified it, and giving a relevant date.
+
+     b) The work must carry prominent notices stating that it is released under this License and any conditions added under section 7. This requirement modifies the requirement in section 4 to “keep intact all notices”.
+
+     c) You must license the entire work, as a whole, under this License to anyone who comes into possession of a copy. This License will therefore apply, along with any applicable section 7 additional terms, to the whole of the work, and all its parts, regardless of how they are packaged. This License gives no permission to license the work in any other way, but it does not invalidate such permission if you have separately received it.
+
+     d) If the work has interactive user interfaces, each must display Appropriate Legal Notices; however, if the Program has interactive interfaces that do not display Appropriate Legal Notices, your work need not make them do so.
+
+A compilation of a covered work with other separate and independent works, which are not by their nature extensions of the covered work, and which are not combined with it such as to form a larger program, in or on a volume of a storage or distribution medium, is called an “aggregate” if the compilation and its resulting copyright are not used to limit the access or legal rights of the compilation's users beyond what the individual works permit. Inclusion of a covered work in an aggregate does not cause this License to apply to the other parts of the aggregate.
+
+6. Conveying Non-Source Forms.
+You may convey a covered work in object code form under the terms of sections 4 and 5, provided that you also convey the machine-readable Corresponding Source under the terms of this License, in one of these ways:
+
+     a) Convey the object code in, or embodied in, a physical product (including a physical distribution medium), accompanied by the Corresponding Source fixed on a durable physical medium customarily used for software interchange.
+
+     b) Convey the object code in, or embodied in, a physical product (including a physical distribution medium), accompanied by a written offer, valid for at least three years and valid for as long as you offer spare parts or customer support for that product model, to give anyone who possesses the object code either (1) a copy of the Corresponding Source for all the software in the product that is covered by this License, on a durable physical medium customarily used for software interchange, for a price no more than your reasonable cost of physically performing this conveying of source, or (2) access to copy the Corresponding Source from a network server at no charge.
+
+     c) Convey individual copies of the object code with a copy of the written offer to provide the Corresponding Source. This alternative is allowed only occasionally and noncommercially, and only if you received the object code with such an offer, in accord with subsection 6b.
+
+     d) Convey the object code by offering access from a designated place (gratis or for a charge), and offer equivalent access to the Corresponding Source in the same way through the same place at no further charge. You need not require recipients to copy the Corresponding Source along with the object code. If the place to copy the object code is a network server, the Corresponding Source may be on a different server (operated by you or a third party) that supports equivalent copying facilities, provided you maintain clear directions next to the object code saying where to find the Corresponding Source. Regardless of what server hosts the Corresponding Source, you remain obligated to ensure that it is available for as long as needed to satisfy these requirements.
+
+     e) Convey the object code using peer-to-peer transmission, provided you inform other peers where the object code and Corresponding Source of the work are being offered to the general public at no charge under subsection 6d.
+
+A separable portion of the object code, whose source code is excluded from the Corresponding Source as a System Library, need not be included in conveying the object code work.
+
+A “User Product” is either (1) a “consumer product”, which means any tangible personal property which is normally used for personal, family, or household purposes, or (2) anything designed or sold for incorporation into a dwelling. In determining whether a product is a consumer product, doubtful cases shall be resolved in favor of coverage. For a particular product received by a particular user, “normally used” refers to a typical or common use of that class of product, regardless of the status of the particular user or of the way in which the particular user actually uses, or expects or is expected to use, the product. A product is a consumer product regardless of whether the product has substantial commercial, industrial or non-consumer uses, unless such uses represent the only significant mode of use of the product.
+
+“Installation Information” for a User Product means any methods, procedures, authorization keys, or other information required to install and execute modified versions of a covered work in that User Product from a modified version of its Corresponding Source. The information must suffice to ensure that the continued functioning of the modified object code is in no case prevented or interfered with solely because modification has been made.
+
+If you convey an object code work under this section in, or with, or specifically for use in, a User Product, and the conveying occurs as part of a transaction in which the right of possession and use of the User Product is transferred to the recipient in perpetuity or for a fixed term (regardless of how the transaction is characterized), the Corresponding Source conveyed under this section must be accompanied by the Installation Information. But this requirement does not apply if neither you nor any third party retains the ability to install modified object code on the User Product (for example, the work has been installed in ROM).
+
+The requirement to provide Installation Information does not include a requirement to continue to provide support service, warranty, or updates for a work that has been modified or installed by the recipient, or for the User Product in which it has been modified or installed. Access to a network may be denied when the modification itself materially and adversely affects the operation of the network or violates the rules and protocols for communication across the network.
+
+Corresponding Source conveyed, and Installation Information provided, in accord with this section must be in a format that is publicly documented (and with an implementation available to the public in source code form), and must require no special password or key for unpacking, reading or copying.
+
+7. Additional Terms.
+“Additional permissions” are terms that supplement the terms of this License by making exceptions from one or more of its conditions. Additional permissions that are applicable to the entire Program shall be treated as though they were included in this License, to the extent that they are valid under applicable law. If additional permissions apply only to part of the Program, that part may be used separately under those permissions, but the entire Program remains governed by this License without regard to the additional permissions.
+
+When you convey a copy of a covered work, you may at your option remove any additional permissions from that copy, or from any part of it. (Additional permissions may be written to require their own removal in certain cases when you modify the work.) You may place additional permissions on material, added by you to a covered work, for which you have or can give appropriate copyright permission.
+
+Notwithstanding any other provision of this License, for material you add to a covered work, you may (if authorized by the copyright holders of that material) supplement the terms of this License with terms:
+
+     a) Disclaiming warranty or limiting liability differently from the terms of sections 15 and 16 of this License; or
+
+     b) Requiring preservation of specified reasonable legal notices or author attributions in that material or in the Appropriate Legal Notices displayed by works containing it; or
+
+     c) Prohibiting misrepresentation of the origin of that material, or requiring that modified versions of such material be marked in reasonable ways as different from the original version; or
+
+     d) Limiting the use for publicity purposes of names of licensors or authors of the material; or
+
+     e) Declining to grant rights under trademark law for use of some trade names, trademarks, or service marks; or
+
+     f) Requiring indemnification of licensors and authors of that material by anyone who conveys the material (or modified versions of it) with contractual assumptions of liability to the recipient, for any liability that these contractual assumptions directly impose on those licensors and authors.
+
+All other non-permissive additional terms are considered “further restrictions” within the meaning of section 10. If the Program as you received it, or any part of it, contains a notice stating that it is governed by this License along with a term that is a further restriction, you may remove that term. If a license document contains a further restriction but permits relicensing or conveying under this License, you may add to a covered work material governed by the terms of that license document, provided that the further restriction does not survive such relicensing or conveying.
+
+If you add terms to a covered work in accord with this section, you must place, in the relevant source files, a statement of the additional terms that apply to those files, or a notice indicating where to find the applicable terms.
+
+Additional terms, permissive or non-permissive, may be stated in the form of a separately written license, or stated as exceptions; the above requirements apply either way.
+
+8. Termination.
+You may not propagate or modify a covered work except as expressly provided under this License. Any attempt otherwise to propagate or modify it is void, and will automatically terminate your rights under this License (including any patent licenses granted under the third paragraph of section 11).
+
+However, if you cease all violation of this License, then your license from a particular copyright holder is reinstated (a) provisionally, unless and until the copyright holder explicitly and finally terminates your license, and (b) permanently, if the copyright holder fails to notify you of the violation by some reasonable means prior to 60 days after the cessation.
+
+Moreover, your license from a particular copyright holder is reinstated permanently if the copyright holder notifies you of the violation by some reasonable means, this is the first time you have received notice of violation of this License (for any work) from that copyright holder, and you cure the violation prior to 30 days after your receipt of the notice.
+
+Termination of your rights under this section does not terminate the licenses of parties who have received copies or rights from you under this License. If your rights have been terminated and not permanently reinstated, you do not qualify to receive new licenses for the same material under section 10.
+
+9. Acceptance Not Required for Having Copies.
+You are not required to accept this License in order to receive or run a copy of the Program. Ancillary propagation of a covered work occurring solely as a consequence of using peer-to-peer transmission to receive a copy likewise does not require acceptance. However, nothing other than this License grants you permission to propagate or modify any covered work. These actions infringe copyright if you do not accept this License. Therefore, by modifying or propagating a covered work, you indicate your acceptance of this License to do so.
+
+10. Automatic Licensing of Downstream Recipients.
+Each time you convey a covered work, the recipient automatically receives a license from the original licensors, to run, modify and propagate that work, subject to this License. You are not responsible for enforcing compliance by third parties with this License.
+
+An “entity transaction” is a transaction transferring control of an organization, or substantially all assets of one, or subdividing an organization, or merging organizations. If propagation of a covered work results from an entity transaction, each party to that transaction who receives a copy of the work also receives whatever licenses to the work the party's predecessor in interest had or could give under the previous paragraph, plus a right to possession of the Corresponding Source of the work from the predecessor in interest, if the predecessor has it or can get it with reasonable efforts.
+
+You may not impose any further restrictions on the exercise of the rights granted or affirmed under this License. For example, you may not impose a license fee, royalty, or other charge for exercise of rights granted under this License, and you may not initiate litigation (including a cross-claim or counterclaim in a lawsuit) alleging that any patent claim is infringed by making, using, selling, offering for sale, or importing the Program or any portion of it.
+
+11. Patents.
+A “contributor” is a copyright holder who authorizes use under this License of the Program or a work on which the Program is based. The work thus licensed is called the contributor's “contributor version”.
+
+A contributor's “essential patent claims” are all patent claims owned or controlled by the contributor, whether already acquired or hereafter acquired, that would be infringed by some manner, permitted by this License, of making, using, or selling its contributor version, but do not include claims that would be infringed only as a consequence of further modification of the contributor version. For purposes of this definition, “control” includes the right to grant patent sublicenses in a manner consistent with the requirements of this License.
+
+Each contributor grants you a non-exclusive, worldwide, royalty-free patent license under the contributor's essential patent claims, to make, use, sell, offer for sale, import and otherwise run, modify and propagate the contents of its contributor version.
+
+In the following three paragraphs, a “patent license” is any express agreement or commitment, however denominated, not to enforce a patent (such as an express permission to practice a patent or covenant not to sue for patent infringement). To “grant” such a patent license to a party means to make such an agreement or commitment not to enforce a patent against the party.
+
+If you convey a covered work, knowingly relying on a patent license, and the Corresponding Source of the work is not available for anyone to copy, free of charge and under the terms of this License, through a publicly available network server or other readily accessible means, then you must either (1) cause the Corresponding Source to be so available, or (2) arrange to deprive yourself of the benefit of the patent license for this particular work, or (3) arrange, in a manner consistent with the requirements of this License, to extend the patent license to downstream recipients. “Knowingly relying” means you have actual knowledge that, but for the patent license, your conveying the covered work in a country, or your recipient's use of the covered work in a country, would infringe one or more identifiable patents in that country that you have reason to believe are valid.
+
+If, pursuant to or in connection with a single transaction or arrangement, you convey, or propagate by procuring conveyance of, a covered work, and grant a patent license to some of the parties receiving the covered work authorizing them to use, propagate, modify or convey a specific copy of the covered work, then the patent license you grant is automatically extended to all recipients of the covered work and works based on it.
+
+A patent license is “discriminatory” if it does not include within the scope of its coverage, prohibits the exercise of, or is conditioned on the non-exercise of one or more of the rights that are specifically granted under this License. You may not convey a covered work if you are a party to an arrangement with a third party that is in the business of distributing software, under which you make payment to the third party based on the extent of your activity of conveying the work, and under which the third party grants, to any of the parties who would receive the covered work from you, a discriminatory patent license (a) in connection with copies of the covered work conveyed by you (or copies made from those copies), or (b) primarily for and in connection with specific products or compilations that contain the covered work, unless you entered into that arrangement, or that patent license was granted, prior to 28 March 2007.
+
+Nothing in this License shall be construed as excluding or limiting any implied license or other defenses to infringement that may otherwise be available to you under applicable patent law.
+
+12. No Surrender of Others' Freedom.
+If conditions are imposed on you (whether by court order, agreement or otherwise) that contradict the conditions of this License, they do not excuse you from the conditions of this License. If you cannot convey a covered work so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not convey it at all. For example, if you agree to terms that obligate you to collect a royalty for further conveying from those to whom you convey the Program, the only way you could satisfy both those terms and this License would be to refrain entirely from conveying the Program.
+
+13. Use with the GNU Affero General Public License.
+Notwithstanding any other provision of this License, you have permission to link or combine any covered work with a work licensed under version 3 of the GNU Affero General Public License into a single combined work, and to convey the resulting work. The terms of this License will continue to apply to the part which is the covered work, but the special requirements of the GNU Affero General Public License, section 13, concerning interaction through a network will apply to the combination as such.
+
+14. Revised Versions of this License.
+The Free Software Foundation may publish revised and/or new versions of the GNU General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns.
+
+Each version is given a distinguishing version number. If the Program specifies that a certain numbered version of the GNU General Public License “or any later version” applies to it, you have the option of following the terms and conditions either of that numbered version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of the GNU General Public License, you may choose any version ever published by the Free Software Foundation.
+
+If the Program specifies that a proxy can decide which future versions of the GNU General Public License can be used, that proxy's public statement of acceptance of a version permanently authorizes you to choose that version for the Program.
+
+Later license versions may give you additional or different permissions. However, no additional obligations are imposed on any author or copyright holder as a result of your choosing to follow a later version.
+
+15. Disclaimer of Warranty.
+THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM “AS IS” WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
+
+16. Limitation of Liability.
+IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
+
+17. Interpretation of Sections 15 and 16.
+If the disclaimer of warranty and limitation of liability provided above cannot be given local legal effect according to their terms, reviewing courts shall apply local law that most closely approximates an absolute waiver of all civil liability in connection with the Program, unless a warranty or assumption of liability accompanies a copy of the Program in return for a fee.
+
+END OF TERMS AND CONDITIONS
+
+How to Apply These Terms to Your New Programs
+
+If you develop a new program, and you want it to be of the greatest possible use to the public, the best way to achieve this is to make it free software which everyone can redistribute and change under these terms.
+
+To do so, attach the following notices to the program. It is safest to attach them to the start of each source file to most effectively state the exclusion of warranty; and each file should have at least the “copyright” line and a pointer to where the full notice is found.
+
+     <one line to give the program's name and a brief idea of what it does.>
+     Copyright (C) <year>  <name of author>
+
+     This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
+
+     This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+
+     You should have received a copy of the GNU General Public License along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+Also add information on how to contact you by electronic and paper mail.
+
+If the program does terminal interaction, make it output a short notice like this when it starts in an interactive mode:
+
+     <program>  Copyright (C) <year>  <name of author>
+     This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+     This is free software, and you are welcome to redistribute it under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate parts of the General Public License. Of course, your program's commands might be different; for a GUI interface, you would use an “about box”.
+
+You should also get your employer (if you work as a programmer) or school, if any, to sign a “copyright disclaimer” for the program, if necessary. For more information on this, and how to apply and follow the GNU GPL, see <https://www.gnu.org/licenses/>.
+
+The GNU General Public License does not permit incorporating your program into proprietary programs. If your program is a subroutine library, you may consider it more useful to permit linking proprietary applications with the library. If this is what you want to do, use the GNU Lesser General Public License instead of this License. But first, please read <https://www.gnu.org/philosophy/why-not-lgpl.html>.

packaging/nominatim-api/pyproject.toml

@@ -2,7 +2,7 @@
 name = "nominatim-api"
 description = "A tool for building a database of OpenStreetMap for geocoding and for searching the database. Search library."
 readme = "README.md"
-requires-python = ">=3.7"
+requires-python = ">=3.9"
 license = 'GPL-3.0-or-later'
 maintainers = [
   { name = "Sarah Hoffmann", email = "lonvia@denofr.de" },

packaging/nominatim-db/COPYING

@@ -1 +0,0 @@
-../../COPYING

packaging/nominatim-db/COPYING

@@ -0,0 +1,232 @@
+GNU GENERAL PUBLIC LICENSE
+Version 3, 29 June 2007
+
+Copyright © 2007 Free Software Foundation, Inc. <https://fsf.org/>
+
+Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed.
+
+Preamble
+
+The GNU General Public License is a free, copyleft license for software and other kinds of works.
+
+The licenses for most software and other practical works are designed to take away your freedom to share and change the works. By contrast, the GNU General Public License is intended to guarantee your freedom to share and change all versions of a program--to make sure it remains free software for all its users. We, the Free Software Foundation, use the GNU General Public License for most of our software; it applies also to any other work released this way by its authors. You can apply it to your programs, too.
+
+When we speak of free software, we are referring to freedom, not price. Our General Public Licenses are designed to make sure that you have the freedom to distribute copies of free software (and charge for them if you wish), that you receive source code or can get it if you want it, that you can change the software or use pieces of it in new free programs, and that you know you can do these things.
+
+To protect your rights, we need to prevent others from denying you these rights or asking you to surrender the rights. Therefore, you have certain responsibilities if you distribute copies of the software, or if you modify it: responsibilities to respect the freedom of others.
+
+For example, if you distribute copies of such a program, whether gratis or for a fee, you must pass on to the recipients the same freedoms that you received. You must make sure that they, too, receive or can get the source code. And you must show them these terms so they know their rights.
+
+Developers that use the GNU GPL protect your rights with two steps: (1) assert copyright on the software, and (2) offer you this License giving you legal permission to copy, distribute and/or modify it.
+
+For the developers' and authors' protection, the GPL clearly explains that there is no warranty for this free software. For both users' and authors' sake, the GPL requires that modified versions be marked as changed, so that their problems will not be attributed erroneously to authors of previous versions.
+
+Some devices are designed to deny users access to install or run modified versions of the software inside them, although the manufacturer can do so. This is fundamentally incompatible with the aim of protecting users' freedom to change the software. The systematic pattern of such abuse occurs in the area of products for individuals to use, which is precisely where it is most unacceptable. Therefore, we have designed this version of the GPL to prohibit the practice for those products. If such problems arise substantially in other domains, we stand ready to extend this provision to those domains in future versions of the GPL, as needed to protect the freedom of users.
+
+Finally, every program is threatened constantly by software patents. States should not allow patents to restrict development and use of software on general-purpose computers, but in those that do, we wish to avoid the special danger that patents applied to a free program could make it effectively proprietary. To prevent this, the GPL assures that patents cannot be used to render the program non-free.
+
+The precise terms and conditions for copying, distribution and modification follow.
+
+TERMS AND CONDITIONS
+
+0. Definitions.
+
+“This License” refers to version 3 of the GNU General Public License.
+
+“Copyright” also means copyright-like laws that apply to other kinds of works, such as semiconductor masks.
+
+“The Program” refers to any copyrightable work licensed under this License. Each licensee is addressed as “you”. “Licensees” and “recipients” may be individuals or organizations.
+
+To “modify” a work means to copy from or adapt all or part of the work in a fashion requiring copyright permission, other than the making of an exact copy. The resulting work is called a “modified version” of the earlier work or a work “based on” the earlier work.
+
+A “covered work” means either the unmodified Program or a work based on the Program.
+
+To “propagate” a work means to do anything with it that, without permission, would make you directly or secondarily liable for infringement under applicable copyright law, except executing it on a computer or modifying a private copy. Propagation includes copying, distribution (with or without modification), making available to the public, and in some countries other activities as well.
+
+To “convey” a work means any kind of propagation that enables other parties to make or receive copies. Mere interaction with a user through a computer network, with no transfer of a copy, is not conveying.
+
+An interactive user interface displays “Appropriate Legal Notices” to the extent that it includes a convenient and prominently visible feature that (1) displays an appropriate copyright notice, and (2) tells the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to view a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion.
+
+1. Source Code.
+The “source code” for a work means the preferred form of the work for making modifications to it. “Object code” means any non-source form of a work.
+
+A “Standard Interface” means an interface that either is an official standard defined by a recognized standards body, or, in the case of interfaces specified for a particular programming language, one that is widely used among developers working in that language.
+
+The “System Libraries” of an executable work include anything, other than the work as a whole, that (a) is included in the normal form of packaging a Major Component, but which is not part of that Major Component, and (b) serves only to enable use of the work with that Major Component, or to implement a Standard Interface for which an implementation is available to the public in source code form. A “Major Component”, in this context, means a major essential component (kernel, window system, and so on) of the specific operating system (if any) on which the executable work runs, or a compiler used to produce the work, or an object code interpreter used to run it.
+
+The “Corresponding Source” for a work in object code form means all the source code needed to generate, install, and (for an executable work) run the object code and to modify the work, including scripts to control those activities. However, it does not include the work's System Libraries, or general-purpose tools or generally available free programs which are used unmodified in performing those activities but which are not part of the work. For example, Corresponding Source includes interface definition files associated with source files for the work, and the source code for shared libraries and dynamically linked subprograms that the work is specifically designed to require, such as by intimate data communication or control flow between those subprograms and other parts of the work.
+
+The Corresponding Source need not include anything that users can regenerate automatically from other parts of the Corresponding Source.
+
+The Corresponding Source for a work in source code form is that same work.
+
+2. Basic Permissions.
+All rights granted under this License are granted for the term of copyright on the Program, and are irrevocable provided the stated conditions are met. This License explicitly affirms your unlimited permission to run the unmodified Program. The output from running a covered work is covered by this License only if the output, given its content, constitutes a covered work. This License acknowledges your rights of fair use or other equivalent, as provided by copyright law.
+
+You may make, run and propagate covered works that you do not convey, without conditions so long as your license otherwise remains in force. You may convey covered works to others for the sole purpose of having them make modifications exclusively for you, or provide you with facilities for running those works, provided that you comply with the terms of this License in conveying all material for which you do not control copyright. Those thus making or running the covered works for you must do so exclusively on your behalf, under your direction and control, on terms that prohibit them from making any copies of your copyrighted material outside their relationship with you.
+
+Conveying under any other circumstances is permitted solely under the conditions stated below. Sublicensing is not allowed; section 10 makes it unnecessary.
+
+3. Protecting Users' Legal Rights From Anti-Circumvention Law.
+No covered work shall be deemed part of an effective technological measure under any applicable law fulfilling obligations under article 11 of the WIPO copyright treaty adopted on 20 December 1996, or similar laws prohibiting or restricting circumvention of such measures.
+
+When you convey a covered work, you waive any legal power to forbid circumvention of technological measures to the extent such circumvention is effected by exercising rights under this License with respect to the covered work, and you disclaim any intention to limit operation or modification of the work as a means of enforcing, against the work's users, your or third parties' legal rights to forbid circumvention of technological measures.
+
+4. Conveying Verbatim Copies.
+You may convey verbatim copies of the Program's source code as you receive it, in any medium, provided that you conspicuously and appropriately publish on each copy an appropriate copyright notice; keep intact all notices stating that this License and any non-permissive terms added in accord with section 7 apply to the code; keep intact all notices of the absence of any warranty; and give all recipients a copy of this License along with the Program.
+
+You may charge any price or no price for each copy that you convey, and you may offer support or warranty protection for a fee.
+
+5. Conveying Modified Source Versions.
+You may convey a work based on the Program, or the modifications to produce it from the Program, in the form of source code under the terms of section 4, provided that you also meet all of these conditions:
+
+     a) The work must carry prominent notices stating that you modified it, and giving a relevant date.
+
+     b) The work must carry prominent notices stating that it is released under this License and any conditions added under section 7. This requirement modifies the requirement in section 4 to “keep intact all notices”.
+
+     c) You must license the entire work, as a whole, under this License to anyone who comes into possession of a copy. This License will therefore apply, along with any applicable section 7 additional terms, to the whole of the work, and all its parts, regardless of how they are packaged. This License gives no permission to license the work in any other way, but it does not invalidate such permission if you have separately received it.
+
+     d) If the work has interactive user interfaces, each must display Appropriate Legal Notices; however, if the Program has interactive interfaces that do not display Appropriate Legal Notices, your work need not make them do so.
+
+A compilation of a covered work with other separate and independent works, which are not by their nature extensions of the covered work, and which are not combined with it such as to form a larger program, in or on a volume of a storage or distribution medium, is called an “aggregate” if the compilation and its resulting copyright are not used to limit the access or legal rights of the compilation's users beyond what the individual works permit. Inclusion of a covered work in an aggregate does not cause this License to apply to the other parts of the aggregate.
+
+6. Conveying Non-Source Forms.
+You may convey a covered work in object code form under the terms of sections 4 and 5, provided that you also convey the machine-readable Corresponding Source under the terms of this License, in one of these ways:
+
+     a) Convey the object code in, or embodied in, a physical product (including a physical distribution medium), accompanied by the Corresponding Source fixed on a durable physical medium customarily used for software interchange.
+
+     b) Convey the object code in, or embodied in, a physical product (including a physical distribution medium), accompanied by a written offer, valid for at least three years and valid for as long as you offer spare parts or customer support for that product model, to give anyone who possesses the object code either (1) a copy of the Corresponding Source for all the software in the product that is covered by this License, on a durable physical medium customarily used for software interchange, for a price no more than your reasonable cost of physically performing this conveying of source, or (2) access to copy the Corresponding Source from a network server at no charge.
+
+     c) Convey individual copies of the object code with a copy of the written offer to provide the Corresponding Source. This alternative is allowed only occasionally and noncommercially, and only if you received the object code with such an offer, in accord with subsection 6b.
+
+     d) Convey the object code by offering access from a designated place (gratis or for a charge), and offer equivalent access to the Corresponding Source in the same way through the same place at no further charge. You need not require recipients to copy the Corresponding Source along with the object code. If the place to copy the object code is a network server, the Corresponding Source may be on a different server (operated by you or a third party) that supports equivalent copying facilities, provided you maintain clear directions next to the object code saying where to find the Corresponding Source. Regardless of what server hosts the Corresponding Source, you remain obligated to ensure that it is available for as long as needed to satisfy these requirements.
+
+     e) Convey the object code using peer-to-peer transmission, provided you inform other peers where the object code and Corresponding Source of the work are being offered to the general public at no charge under subsection 6d.
+
+A separable portion of the object code, whose source code is excluded from the Corresponding Source as a System Library, need not be included in conveying the object code work.
+
+A “User Product” is either (1) a “consumer product”, which means any tangible personal property which is normally used for personal, family, or household purposes, or (2) anything designed or sold for incorporation into a dwelling. In determining whether a product is a consumer product, doubtful cases shall be resolved in favor of coverage. For a particular product received by a particular user, “normally used” refers to a typical or common use of that class of product, regardless of the status of the particular user or of the way in which the particular user actually uses, or expects or is expected to use, the product. A product is a consumer product regardless of whether the product has substantial commercial, industrial or non-consumer uses, unless such uses represent the only significant mode of use of the product.
+
+“Installation Information” for a User Product means any methods, procedures, authorization keys, or other information required to install and execute modified versions of a covered work in that User Product from a modified version of its Corresponding Source. The information must suffice to ensure that the continued functioning of the modified object code is in no case prevented or interfered with solely because modification has been made.
+
+If you convey an object code work under this section in, or with, or specifically for use in, a User Product, and the conveying occurs as part of a transaction in which the right of possession and use of the User Product is transferred to the recipient in perpetuity or for a fixed term (regardless of how the transaction is characterized), the Corresponding Source conveyed under this section must be accompanied by the Installation Information. But this requirement does not apply if neither you nor any third party retains the ability to install modified object code on the User Product (for example, the work has been installed in ROM).
+
+The requirement to provide Installation Information does not include a requirement to continue to provide support service, warranty, or updates for a work that has been modified or installed by the recipient, or for the User Product in which it has been modified or installed. Access to a network may be denied when the modification itself materially and adversely affects the operation of the network or violates the rules and protocols for communication across the network.
+
+Corresponding Source conveyed, and Installation Information provided, in accord with this section must be in a format that is publicly documented (and with an implementation available to the public in source code form), and must require no special password or key for unpacking, reading or copying.
+
+7. Additional Terms.
+“Additional permissions” are terms that supplement the terms of this License by making exceptions from one or more of its conditions. Additional permissions that are applicable to the entire Program shall be treated as though they were included in this License, to the extent that they are valid under applicable law. If additional permissions apply only to part of the Program, that part may be used separately under those permissions, but the entire Program remains governed by this License without regard to the additional permissions.
+
+When you convey a copy of a covered work, you may at your option remove any additional permissions from that copy, or from any part of it. (Additional permissions may be written to require their own removal in certain cases when you modify the work.) You may place additional permissions on material, added by you to a covered work, for which you have or can give appropriate copyright permission.
+
+Notwithstanding any other provision of this License, for material you add to a covered work, you may (if authorized by the copyright holders of that material) supplement the terms of this License with terms:
+
+     a) Disclaiming warranty or limiting liability differently from the terms of sections 15 and 16 of this License; or
+
+     b) Requiring preservation of specified reasonable legal notices or author attributions in that material or in the Appropriate Legal Notices displayed by works containing it; or
+
+     c) Prohibiting misrepresentation of the origin of that material, or requiring that modified versions of such material be marked in reasonable ways as different from the original version; or
+
+     d) Limiting the use for publicity purposes of names of licensors or authors of the material; or
+
+     e) Declining to grant rights under trademark law for use of some trade names, trademarks, or service marks; or
+
+     f) Requiring indemnification of licensors and authors of that material by anyone who conveys the material (or modified versions of it) with contractual assumptions of liability to the recipient, for any liability that these contractual assumptions directly impose on those licensors and authors.
+
+All other non-permissive additional terms are considered “further restrictions” within the meaning of section 10. If the Program as you received it, or any part of it, contains a notice stating that it is governed by this License along with a term that is a further restriction, you may remove that term. If a license document contains a further restriction but permits relicensing or conveying under this License, you may add to a covered work material governed by the terms of that license document, provided that the further restriction does not survive such relicensing or conveying.
+
+If you add terms to a covered work in accord with this section, you must place, in the relevant source files, a statement of the additional terms that apply to those files, or a notice indicating where to find the applicable terms.
+
+Additional terms, permissive or non-permissive, may be stated in the form of a separately written license, or stated as exceptions; the above requirements apply either way.
+
+8. Termination.
+You may not propagate or modify a covered work except as expressly provided under this License. Any attempt otherwise to propagate or modify it is void, and will automatically terminate your rights under this License (including any patent licenses granted under the third paragraph of section 11).
+
+However, if you cease all violation of this License, then your license from a particular copyright holder is reinstated (a) provisionally, unless and until the copyright holder explicitly and finally terminates your license, and (b) permanently, if the copyright holder fails to notify you of the violation by some reasonable means prior to 60 days after the cessation.
+
+Moreover, your license from a particular copyright holder is reinstated permanently if the copyright holder notifies you of the violation by some reasonable means, this is the first time you have received notice of violation of this License (for any work) from that copyright holder, and you cure the violation prior to 30 days after your receipt of the notice.
+
+Termination of your rights under this section does not terminate the licenses of parties who have received copies or rights from you under this License. If your rights have been terminated and not permanently reinstated, you do not qualify to receive new licenses for the same material under section 10.
+
+9. Acceptance Not Required for Having Copies.
+You are not required to accept this License in order to receive or run a copy of the Program. Ancillary propagation of a covered work occurring solely as a consequence of using peer-to-peer transmission to receive a copy likewise does not require acceptance. However, nothing other than this License grants you permission to propagate or modify any covered work. These actions infringe copyright if you do not accept this License. Therefore, by modifying or propagating a covered work, you indicate your acceptance of this License to do so.
+
+10. Automatic Licensing of Downstream Recipients.
+Each time you convey a covered work, the recipient automatically receives a license from the original licensors, to run, modify and propagate that work, subject to this License. You are not responsible for enforcing compliance by third parties with this License.
+
+An “entity transaction” is a transaction transferring control of an organization, or substantially all assets of one, or subdividing an organization, or merging organizations. If propagation of a covered work results from an entity transaction, each party to that transaction who receives a copy of the work also receives whatever licenses to the work the party's predecessor in interest had or could give under the previous paragraph, plus a right to possession of the Corresponding Source of the work from the predecessor in interest, if the predecessor has it or can get it with reasonable efforts.
+
+You may not impose any further restrictions on the exercise of the rights granted or affirmed under this License. For example, you may not impose a license fee, royalty, or other charge for exercise of rights granted under this License, and you may not initiate litigation (including a cross-claim or counterclaim in a lawsuit) alleging that any patent claim is infringed by making, using, selling, offering for sale, or importing the Program or any portion of it.
+
+11. Patents.
+A “contributor” is a copyright holder who authorizes use under this License of the Program or a work on which the Program is based. The work thus licensed is called the contributor's “contributor version”.
+
+A contributor's “essential patent claims” are all patent claims owned or controlled by the contributor, whether already acquired or hereafter acquired, that would be infringed by some manner, permitted by this License, of making, using, or selling its contributor version, but do not include claims that would be infringed only as a consequence of further modification of the contributor version. For purposes of this definition, “control” includes the right to grant patent sublicenses in a manner consistent with the requirements of this License.
+
+Each contributor grants you a non-exclusive, worldwide, royalty-free patent license under the contributor's essential patent claims, to make, use, sell, offer for sale, import and otherwise run, modify and propagate the contents of its contributor version.
+
+In the following three paragraphs, a “patent license” is any express agreement or commitment, however denominated, not to enforce a patent (such as an express permission to practice a patent or covenant not to sue for patent infringement). To “grant” such a patent license to a party means to make such an agreement or commitment not to enforce a patent against the party.
+
+If you convey a covered work, knowingly relying on a patent license, and the Corresponding Source of the work is not available for anyone to copy, free of charge and under the terms of this License, through a publicly available network server or other readily accessible means, then you must either (1) cause the Corresponding Source to be so available, or (2) arrange to deprive yourself of the benefit of the patent license for this particular work, or (3) arrange, in a manner consistent with the requirements of this License, to extend the patent license to downstream recipients. “Knowingly relying” means you have actual knowledge that, but for the patent license, your conveying the covered work in a country, or your recipient's use of the covered work in a country, would infringe one or more identifiable patents in that country that you have reason to believe are valid.
+
+If, pursuant to or in connection with a single transaction or arrangement, you convey, or propagate by procuring conveyance of, a covered work, and grant a patent license to some of the parties receiving the covered work authorizing them to use, propagate, modify or convey a specific copy of the covered work, then the patent license you grant is automatically extended to all recipients of the covered work and works based on it.
+
+A patent license is “discriminatory” if it does not include within the scope of its coverage, prohibits the exercise of, or is conditioned on the non-exercise of one or more of the rights that are specifically granted under this License. You may not convey a covered work if you are a party to an arrangement with a third party that is in the business of distributing software, under which you make payment to the third party based on the extent of your activity of conveying the work, and under which the third party grants, to any of the parties who would receive the covered work from you, a discriminatory patent license (a) in connection with copies of the covered work conveyed by you (or copies made from those copies), or (b) primarily for and in connection with specific products or compilations that contain the covered work, unless you entered into that arrangement, or that patent license was granted, prior to 28 March 2007.
+
+Nothing in this License shall be construed as excluding or limiting any implied license or other defenses to infringement that may otherwise be available to you under applicable patent law.
+
+12. No Surrender of Others' Freedom.
+If conditions are imposed on you (whether by court order, agreement or otherwise) that contradict the conditions of this License, they do not excuse you from the conditions of this License. If you cannot convey a covered work so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not convey it at all. For example, if you agree to terms that obligate you to collect a royalty for further conveying from those to whom you convey the Program, the only way you could satisfy both those terms and this License would be to refrain entirely from conveying the Program.
+
+13. Use with the GNU Affero General Public License.
+Notwithstanding any other provision of this License, you have permission to link or combine any covered work with a work licensed under version 3 of the GNU Affero General Public License into a single combined work, and to convey the resulting work. The terms of this License will continue to apply to the part which is the covered work, but the special requirements of the GNU Affero General Public License, section 13, concerning interaction through a network will apply to the combination as such.
+
+14. Revised Versions of this License.
+The Free Software Foundation may publish revised and/or new versions of the GNU General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns.
+
+Each version is given a distinguishing version number. If the Program specifies that a certain numbered version of the GNU General Public License “or any later version” applies to it, you have the option of following the terms and conditions either of that numbered version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of the GNU General Public License, you may choose any version ever published by the Free Software Foundation.
+
+If the Program specifies that a proxy can decide which future versions of the GNU General Public License can be used, that proxy's public statement of acceptance of a version permanently authorizes you to choose that version for the Program.
+
+Later license versions may give you additional or different permissions. However, no additional obligations are imposed on any author or copyright holder as a result of your choosing to follow a later version.
+
+15. Disclaimer of Warranty.
+THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM “AS IS” WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
+
+16. Limitation of Liability.
+IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
+
+17. Interpretation of Sections 15 and 16.
+If the disclaimer of warranty and limitation of liability provided above cannot be given local legal effect according to their terms, reviewing courts shall apply local law that most closely approximates an absolute waiver of all civil liability in connection with the Program, unless a warranty or assumption of liability accompanies a copy of the Program in return for a fee.
+
+END OF TERMS AND CONDITIONS
+
+How to Apply These Terms to Your New Programs
+
+If you develop a new program, and you want it to be of the greatest possible use to the public, the best way to achieve this is to make it free software which everyone can redistribute and change under these terms.
+
+To do so, attach the following notices to the program. It is safest to attach them to the start of each source file to most effectively state the exclusion of warranty; and each file should have at least the “copyright” line and a pointer to where the full notice is found.
+
+     <one line to give the program's name and a brief idea of what it does.>
+     Copyright (C) <year>  <name of author>
+
+     This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
+
+     This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
+
+     You should have received a copy of the GNU General Public License along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+Also add information on how to contact you by electronic and paper mail.
+
+If the program does terminal interaction, make it output a short notice like this when it starts in an interactive mode:
+
+     <program>  Copyright (C) <year>  <name of author>
+     This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
+     This is free software, and you are welcome to redistribute it under certain conditions; type `show c' for details.
+
+The hypothetical commands `show w' and `show c' should show the appropriate parts of the General Public License. Of course, your program's commands might be different; for a GUI interface, you would use an “about box”.
+
+You should also get your employer (if you work as a programmer) or school, if any, to sign a “copyright disclaimer” for the program, if necessary. For more information on this, and how to apply and follow the GNU GPL, see <https://www.gnu.org/licenses/>.
+
+The GNU General Public License does not permit incorporating your program into proprietary programs. If your program is a subroutine library, you may consider it more useful to permit linking proprietary applications with the library. If this is what you want to do, use the GNU Lesser General Public License instead of this License. But first, please read <https://www.gnu.org/philosophy/why-not-lgpl.html>.

packaging/nominatim-db/pyproject.toml

@@ -2,7 +2,7 @@
 name = "nominatim-db"
 description = "A tool for building a database of OpenStreetMap for geocoding and for searching the database. Database backend."
 readme = "README.md"
-requires-python = ">=3.7"
+requires-python = ">=3.9"
 license = 'GPL-3.0-or-later'
 maintainers = [
   { name = "Sarah Hoffmann", email = "lonvia@denofr.de" },
@@ -19,7 +19,6 @@ dependencies = [
     "python-dotenv",
     "jinja2",
     "pyYAML>=5.1",
-    "datrie",
     "psutil",
     "PyICU"
 ]

packaging/nominatim-db/scripts/nominatim

@@ -2,4 +2,4 @@
 
 from nominatim_db import cli
 
-exit(cli.nominatim(osm2pgsql_path=None))
+exit(cli.nominatim())

settings/address-levels.json

@@ -216,6 +216,14 @@
       }
   }
 },
+{ "countries" : ["sa"],
+  "tags" : {
+      "place" : {
+          "province" : 12,
+          "municipality" : 18
+      }
+  }
+},
 { "countries" : ["sk"],
   "tags" : {
       "boundary" : {

settings/country_settings.yaml

@@ -944,7 +944,7 @@ kp:
 # South Korea (대한민국)
 kr:
     partition: 49
-    languages: ko, en
+    languages: ko
     names: !include country-names/kr.yaml
     postcode:
       pattern: "ddddd"
@@ -1809,7 +1809,8 @@ us:
     languages: en
     names: !include country-names/us.yaml
     postcode:
-      pattern: "ddddd"
+      pattern: "(ddddd)(?:-dddd)?"
+      output: \1
 
 
 # Uruguay (Uruguay)

settings/env.defaults

@@ -5,7 +5,6 @@
 # Database connection string.
 # Add host, port, user etc through additional semicolon-separated attributes.
 # e.g. ;host=...;port=...;user=...;password=...
-# Changing this variable requires to run 'nominatim refresh --website'.
 NOMINATIM_DATABASE_DSN="pgsql:dbname=nominatim"
 
 # Database web user.
@@ -36,11 +35,11 @@ NOMINATIM_TOKENIZER_CONFIG=
 
 # Search in the Tiger house number data for the US.
 # Note: The tables must already exist or queries will throw errors.
-# Changing this value requires to run ./utils/setup --create-functions --setup-website.
+# Changing this value requires to run ./utils/setup --create-functions.
 NOMINATIM_USE_US_TIGER_DATA=no
 
 # Search in the auxiliary housenumber table.
-# Changing this value requires to run ./utils/setup --create-functions --setup-website.
+# Changing this value requires to run ./utils/setup --create-functions.
 NOMINATIM_USE_AUX_LOCATION_DATA=no
 
 # Proxy settings
@@ -143,8 +142,7 @@ NOMINATIM_REPLICATION_RECHECK_INTERVAL=60
 
 ### API settings
 #
-# The following settings configure the API responses. You must rerun
-# 'nominatim refresh --website' after changing any of them.
+# The following settings configure the API responses.
 
 # Send permissive CORS access headers.
 # When enabled, send CORS headers to allow access to everybody.
@@ -192,16 +190,17 @@ NOMINATIM_REQUEST_TIMEOUT=60
 # to geocode" instead.
 NOMINATIM_SEARCH_WITHIN_COUNTRIES=False
 
+# Specifies the order in which different name tags are used.
+# The values in this list determine the preferred order of name variants,
+# including language-specific names.
+# Comma-separated list, where :XX stands for language-specific tags
+# (e.g. name:en) and no :XX stands for general tags (e.g. name).
+NOMINATIM_OUTPUT_NAMES=name:XX,name,brand,official_name:XX,short_name:XX,official_name,short_name,ref
+
 ### Log settings
 #
 # The following options allow to enable logging of API requests.
-# You must rerun 'nominatim refresh --website' after changing any of them.
 #
-# Enable logging of requests into the DB.
-# The request will be logged into the new_query_log table.
-# You should set up a cron job that regularly clears out this table.
-NOMINATIM_LOG_DB=no
-
 # Enable logging of requests into a file.
 # To enable logging set this setting to the file to log to.
 NOMINATIM_LOG_FILE=

settings/icu-rules/variants-ca.yaml

@@ -4,7 +4,7 @@
     -  aparcament -> aparc
     -  apartament -> apmt
     -  apartat -> apt
-    - àtic -> àt
+    -  àtic -> àt
     -  autopista -> auto
     -  autopista -> autop
     -  autovia -> autov
@@ -19,7 +19,6 @@
     -  biblioteca -> bibl
     -  bloc -> bl
     -  carrer -> c
-    -  carrer -> c/
     -  carreró -> cró
     -  carretera -> ctra
     -  cantonada -> cant
@@ -58,7 +57,6 @@
     -  número -> n
     -  sense número -> s/n
     -  parada -> par
-    -  parcel·la -> parc
     -  passadís -> pdís
     -  passatge -> ptge
     -  passeig -> pg

settings/icu-rules/variants-en.yaml

@@ -1,438 +1,393 @@
 # Source: https://wiki.openstreetmap.org/wiki/Name_finder:Abbreviations#English
+# Source: https://pe.usps.com/text/pub28/28apc_002.htm
 - lang: en
   words:
     -  Access -> Accs
     -  Air Force Base -> AFB
     -  Air National Guard Base -> ANGB
     -  Airport -> Aprt
-    -  Alley -> Al
-    -  Alley -> All
-    -  Alley -> Ally
-    -  Alley -> Aly
+    -  Alley -> Al,All,Ally,Aly
     -  Alleyway -> Alwy
     -  Amble -> Ambl
+    -  Anex -> Anx
     -  Apartments -> Apts
-    -  Approach -> Apch
-    -  Approach -> App
+    -  Approach -> Apch,App
     -  Arcade -> Arc
     -  Arterial -> Artl
     -  Artery -> Arty
-    -  Avenue -> Av
-    -  Avenue -> Ave
+    -  Avenue -> Av,Ave
     -  Back -> Bk
     -  Banan -> Ba
-    -  Basin -> Basn
-    -  Basin -> Bsn
+    -  Basin -> Basn,Bsn
+    -  Bayou -> Byu
     -  Beach -> Bch
-    -  Bend -> Bend
     -  Bend -> Bnd
     -  Block -> Blk
+    -  Bluff -> Blf
+    -  Bluffs -> Blfs
     -  Boardwalk -> Bwlk
-    -  Boulevard -> Blvd
-    -  Boulevard -> Bvd
+    -  Bottom -> Btm
+    -  Boulevard -> Blvd,Bvd
     -  Boundary -> Bdy
     -  Bowl -> Bl
     -  Brace -> Br
     -  Brae -> Br
-    -  Brae -> Brae
+    -  Branch -> Br
     -  Break -> Brk
-    -  Bridge -> Bdge
-    -  Bridge -> Br
-    -  Bridge -> Brdg
-    -  Bridge -> Bri
-    -  Broadway -> Bdwy
-    -  Broadway -> Bway
-    -  Broadway -> Bwy
+    -  Bridge$ -> Bdge,Br,Brdg,Brg,Bri
+    -  Broadway -> Bdwy,Bway,Bwy
     -  Brook -> Brk
+    -  Brooks -> Brks
     -  Brow -> Brw
-    -  Brow -> Brow
-    -  Buildings -> Bldgs
-    -  Buildings -> Bldngs
+    -  Buildings -> Bldgs,Bldngs
     -  Business -> Bus
-    -  Bypass -> Bps
-    -  Bypass -> Byp
-    -  Bypass -> Bypa
+    -  Burg -> Bg
+    -  Burgs -> Bgs
+    -  Bypass -> Bps,Byp,Bypa
     -  Byway -> Bywy
+    -  Camp -> Cp
+    -  Canyon -> Cyn
+    -  Cape -> Cpe
     -  Caravan -> Cvn
-    -  Causeway -> Caus
-    -  Causeway -> Cswy
-    -  Causeway -> Cway
-    -  Center -> Cen
-    -  Center -> Ctr
+    -  Causeway -> Caus,Cswy,Cway
+    -  Center,Centre -> Cen,Ctr
+    -  Centers -> Ctrs
     -  Central -> Ctrl
-    -  Centre -> Cen
-    -  Centre -> Ctr
     -  Centreway -> Cnwy
     -  Chase -> Ch
     -  Church -> Ch
     -  Circle -> Cir
-    -  Circuit -> Cct
-    -  Circuit -> Ci
-    -  Circus -> Crc
-    -  Circus -> Crcs
+    -  Circles -> Cirs
+    -  Circuit -> Cct,Ci
+    -  Circus -> Crc,Crcs
     -  City -> Cty
+    -  Cliff -> Clf
+    -  Cliffs -> Clfs
     -  Close -> Cl
-    -  Common -> Cmn
-    -  Common -> Comm
+    -  Club -> Clb
+    -  Common -> Cmn,Comm
+    -  Commons -> Cmns
     -  Community -> Comm
     -  Concourse -> Cnc
     -  Concourse -> Con
     -  Copse -> Cps
-    -  Corner -> Cnr
-    -  Corner -> Crn
+    -  Corner -> Cor,Cnr,Crn
+    -  Corners -> Cors
     -  Corso -> Cso
     -  Cottages -> Cotts
     -  County -> Co
     -  County Road -> CR
     -  County Route -> CR
-    -  Court -> Crt
-    -  Court -> Ct
+    -  Course -> Crse
+    -  Court -> Crt,Ct
+    -  Courts -> Cts
     -  Courtyard -> Cyd
     -  Courtyard -> Ctyd
-    -  Cove -> Ce
-    -  Cove -> Cov
-    -  Cove -> Cove
-    -  Cove -> Cv
-    -  Creek -> Ck
-    -  Creek -> Cr
-    -  Creek -> Crk
+    -  Cove$ -> Ce,Cov,Cv
+    -  Coves -> Cvs
+    -  Creek$ -> Ck,Cr,Crk
     -  Crescent -> Cr
     -  Crescent -> Cres
-    -  Crest -> Crst
-    -  Crest -> Cst
+    -  Crest -> Crst,Cst
     -  Croft -> Cft
-    -  Cross -> Cs
-    -  Cross -> Crss
-    -  Crossing -> Crsg
-    -  Crossing -> Csg
-    -  Crossing -> Xing
-    -  Crossroad -> Crd
+    -  Cross -> Cs,Crss
+    -  Crossing -> Crsg,Csg,Xing
+    -  Crossroad -> Crd,Xrd
+    -  Crossroads -> Xrds
     -  Crossway -> Cowy
-    -  Cul-de-sac -> Cds
-    -  Cul-de-sac -> Csac
-    -  Curve -> Cve
+    -  Cul-de-sac -> Cds,Csac
+    -  Curve -> Cve,Curv
     -  Cutting -> Cutt
     -  Dale -> Dle
-    -  Dale -> Dale
+    -  Dam -> Dm
     -  Deviation -> Devn
-    -  Dip -> Dip
     -  Distributor -> Dstr
+    -  Divide -> Dv
     -  Down -> Dn
     -  Downs -> Dn
-    -  Drive -> Dr
-    -  Drive -> Drv
-    -  Drive -> Dv
+    -  Drive -> Dr,Drv,Dv
+    -  Drives -> Drs
     -  Drive-In => Drive-In # prevent abbreviation here
-    -  Driveway -> Drwy
-    -  Driveway -> Dvwy
-    -  Driveway -> Dwy
+    -  Driveway -> Drwy,Dvwy,Dwy
     -  East -> E
     -  Edge -> Edg
-    -  Edge -> Edge
     -  Elbow -> Elb
-    -  End -> End
     -  Entrance -> Ent
     -  Esplanade -> Esp
     -  Estate -> Est
-    -  Expressway -> Exp
-    -  Expressway -> Expy
-    -  Expressway -> Expwy
-    -  Expressway -> Xway
+    -  Estates -> Ests
+    -  Expressway -> Exp,Expy,Expwy,Xway
     -  Extension -> Ex
-    -  Fairway -> Fawy
-    -  Fairway -> Fy
+    -  Extensions -> Exts
+    -  Fairway -> Fawy,Fy
+    -  Falls -> Fls
     -  Father -> Fr
-    -  Ferry -> Fy
-    -  Field -> Fd
+    -  Ferry -> Fy,Fry
+    -  Field -> Fd,Fld
+    -  Fields -> Flds
     -  Fire Track -> Ftrk
     -  Firetrail -> Fit
-    -  Flat -> Fl
-    -  Flat -> Flat
+    -  Flat -> Fl,Flt
+    -  Flats -> Flts
     -  Follow -> Folw
     -  Footway -> Ftwy
+    -  Ford -> Frd
+    -  Fords -> Frds
     -  Foreshore -> Fshr
+    -  Forest -> Frst
     -  Forest Service Road -> FSR
+    -  Forge -> Frg
+    -  Forges -> Frgs
     -  Formation -> Form
+    -  Fork -> Frk
+    -  Forks -> Frks
     -  Fort -> Ft
-    -  Freeway -> Frwy
-    -  Freeway -> Fwy
+    -  Freeway -> Frwy,Fwy
     -  Front -> Frnt
-    -  Frontage -> Fr
-    -  Frontage -> Frtg
-    -  Gap -> Gap
+    -  Frontage -> Fr,Frtg
     -  Garden -> Gdn
-    -  Gardens -> Gdn
-    -  Gardens -> Gdns
-    -  Gate -> Ga
-    -  Gate -> Gte
-    -  Gates -> Ga
-    -  Gates -> Gte
-    -  Gateway -> Gwy
+    -  Gardens -> Gdn,Gdns
+    -  Gate,Gates -> Ga,Gte
+    -  Gateway -> Gwy,Gtwy
     -  George -> Geo
-    -  Glade -> Gl
-    -  Glade -> Gld
-    -  Glade -> Glde
+    -  Glade$ -> Gl,Gld,Glde
     -  Glen -> Gln
-    -  Glen -> Glen
+    -  Glens -> Glns
     -  Grange -> Gra
-    -  Green -> Gn
-    -  Green -> Grn
+    -  Green -> Gn,Grn
+    -  Greens -> Grns
     -  Ground -> Grnd
-    -  Grove -> Gr
-    -  Grove -> Gro
+    -  Grove$ -> Gr,Gro,Grv
+    -  Groves -> Grvs
     -  Grovet -> Gr
     -  Gully -> Gly
-    -  Harbor -> Hbr
-    -  Harbour -> Hbr
+    -  Harbor -> Hbr,Harbour
+    -  Harbors -> Hbrs
+    -  Harbour -> Hbr,Harbor
     -  Haven -> Hvn
     -  Head -> Hd
     -  Heads -> Hd
-    -  Heights -> Hgts
-    -  Heights -> Ht
-    -  Heights -> Hts
+    -  Heights -> Hgts,Ht,Hts
     -  High School -> HS
-    -  Highroad -> Hird
-    -  Highroad -> Hrd
+    -  Highroad -> Hird,Hrd
     -  Highway -> Hwy
-    -  Hill -> Hill
     -  Hill -> Hl
-    -  Hills -> Hl
-    -  Hills -> Hls
+    -  Hills -> Hl,Hls
+    -  Hollow -> Holw
     -  Hospital -> Hosp
-    -  House -> Ho
-    -  House -> Hse
+    -  House -> Ho,Hse
     -  Industrial -> Ind
+    -  Inlet -> Inlt
     -  Interchange -> Intg
     -  International -> Intl
-    -  Island -> I
-    -  Island -> Is
-    -  Junction -> Jctn
-    -  Junction -> Jnc
+    -  Island -> I,Is
+    -  Islands -> Iss
+    -  Junction -> Jct,Jctn,Jnc
+    -  Junctions -> Jcts
     -  Junior -> Jr
-    -  Key -> Key
+    -  Key -> Ky
+    -  Keys -> Kys
+    -  Knoll -> Knl
+    -  Knolls -> Knls
     -  Lagoon -> Lgn
-    -  Lakes -> L
-    -  Landing -> Ldg
-    -  Lane -> La
-    -  Lane -> Lane
-    -  Lane -> Ln
+    -  Lake -> Lk
+    -  Lakes -> L,Lks
+    -  Landing -> Ldg,Lndg
+    -  Lane -> La,Ln
     -  Laneway -> Lnwy
-    -  Line -> Line
+    -  Light -> Lgt
+    -  Lights -> Lgts
     -  Line -> Ln
-    -  Link -> Link
     -  Link -> Lk
-    -  Little -> Lit
-    -  Little -> Lt
+    -  Little -> Lit,Lt
+    -  Loaf -> Lf
+    -  Lock -> Lck
+    -  Locks -> Lcks
     -  Lodge -> Ldg
     -  Lookout -> Lkt
-    -  Loop -> Loop
     -  Loop -> Lp
-    -  Lower -> Low
-    -  Lower -> Lr
-    -  Lower -> Lwr
-    -  Mall -> Mall
+    -  Lower -> Low,Lr,Lwr
     -  Mall -> Ml
     -  Manor -> Mnr
+    -  Manors -> Mnrs
     -  Mansions -> Mans
     -  Market -> Mkt
     -  Meadow -> Mdw
-    -  Meadows -> Mdw
-    -  Meadows -> Mdws
+    -  Meadows -> Mdw,Mdws
     -  Mead -> Md
-    -  Meander -> Mdr
-    -  Meander -> Mndr
-    -  Meander -> Mr
+    -  Meander -> Mdr,Mndr,Mr
     -  Medical -> Med
     -  Memorial -> Mem
-    -  Mews -> Mews
     -  Mews -> Mw
     -  Middle -> Mid
     -  Middle School -> MS
     -  Mile -> Mi
     -  Military -> Mil
-    -  Motorway -> Mtwy
-    -  Motorway -> Mwy
+    -  Mill -> Ml
+    -  Mills -> Mls
+    -  Mission -> Msn
+    -  Motorway -> Mtwy,Mwy
     -  Mount -> Mt
     -  Mountain -> Mtn
-    -  Mountains -> Mtn
+    -  Mountains$ -> Mtn,Mtns
     -  Municipal -> Mun
     -  Museum -> Mus
     -  National Park -> NP
     -  National Recreation Area -> NRA
     -  National Wildlife Refuge Area -> NWRA
+    -  Neck -> Nck
     -  Nook -> Nk
-    -  Nook -> Nook
     -  North -> N
     -  Northeast -> NE
     -  Northwest -> NW
-    -  Outlook -> Out
-    -  Outlook -> Otlk
+    -  Orchard -> Orch
+    -  Outlook -> Out,Otlk
+    -  Overpass -> Opas
     -  Parade -> Pde
     -  Paradise -> Pdse
-    -  Park -> Park
     -  Park -> Pk
     -  Parklands -> Pkld
-    -  Parkway -> Pkwy
-    -  Parkway -> Pky
-    -  Parkway -> Pwy
-    -  Pass -> Pass
+    -  Parkway -> Pkwy,Pky,Pwy
+    -  Parkways -> Pkwy
     -  Pass -> Ps
     -  Passage -> Psge
-    -  Path -> Path
-    -  Pathway -> Phwy
-    -  Pathway -> Pway
-    -  Pathway -> Pwy
+    -  Pathway -> Phwy,Pway,Pwy
     -  Piazza -> Piaz
     -  Pike -> Pk
+    -  Pine -> Pne
+    -  Pines -> Pnes
     -  Place -> Pl
-    -  Plain -> Pl
-    -  Plains -> Pl
+    -  Plain -> Pl,Pln
+    -  Plains -> Pl,Plns
     -  Plateau -> Plat
-    -  Plaza -> Pl
-    -  Plaza -> Plz
-    -  Plaza -> Plza
+    -  Plaza -> Pl,Plz,Plza
     -  Pocket -> Pkt
-    -  Point -> Pnt
-    -  Point -> Pt
-    -  Port -> Port
-    -  Port -> Pt
+    -  Point -> Pnt,Pt
+    -  Points -> Pts
+    -  Port -> Prt,Pt
+    -  Ports -> Prts
     -  Post Office -> PO
+    -  Prairie -> Pr
     -  Precinct -> Pct
-    -  Promenade -> Prm
-    -  Promenade -> Prom
-    -  Quad -> Quad
+    -  Promenade -> Prm,Prom
     -  Quadrangle -> Qdgl
-    -  Quadrant -> Qdrt
-    -  Quadrant -> Qd
+    -  Quadrant -> Qdrt,Qd
     -  Quay -> Qy
     -  Quays -> Qy
     -  Quays -> Qys
+    -  Radial -> Radl
     -  Ramble -> Ra
     -  Ramble -> Rmbl
-    -  Range -> Rge
-    -  Range -> Rnge
+    -  Ranch -> Rnch
+    -  Range -> Rge,Rnge
+    -  Rapid -> Rpd
+    -  Rapids -> Rpds
     -  Reach -> Rch
     -  Reservation -> Res
     -  Reserve -> Res
     -  Reservoir -> Res
-    -  Rest -> Rest
     -  Rest -> Rst
-    -  Retreat -> Rt
-    -  Retreat -> Rtt
+    -  Retreat -> Rt,Rtt
     -  Return -> Rtn
-    -  Ridge -> Rdg
-    -  Ridge -> Rdge
+    -  Ridge -> Rdg,Rdge
+    -  Ridges -> Rdgs
     -  Ridgeway -> Rgwy
     -  Right of Way -> Rowy
     -  Rise -> Ri
-    -  Rise -> Rise
-    -  River -> R
-    -  River -> Riv
-    -  River -> Rvr
+    -  ^River -> R,Riv,Rvr
+    -  River$ -> R,Riv,Rvr
     -  Riverway -> Rvwy
     -  Riviera -> Rvra
     -  Road -> Rd
     -  Roads -> Rds
     -  Roadside -> Rdsd
-    -  Roadway -> Rdwy
-    -  Roadway -> Rdy
-    -  Robert -> Robt
+    -  Roadway -> Rdwy,Rdy
     -  Rocks -> Rks
     -  Ronde -> Rnde
     -  Rosebowl -> Rsbl
     -  Rotary -> Rty
     -  Round -> Rnd
-    -  Route -> Rt
-    -  Route -> Rte
-    -  Row -> Row
-    -  Rue -> Rue
-    -  Run -> Run
+    -  Route -> Rt,Rte
     -  Saint -> St
     -  Saints -> SS
     -  Senior -> Sr
-    -  Serviceway -> Swy
-    -  Serviceway -> Svwy
+    -  Serviceway -> Swy,Svwy
+    -  Shoal -> Shl
+    -  Shore -> Shr
+    -  Shores -> Shrs
     -  Shunt -> Shun
     -  Siding -> Sdng
     -  Sister -> Sr
+    -  Skyway -> Skwy
     -  Slope -> Slpe
     -  Sound -> Snd
-    -  South -> S
-    -  South -> Sth
+    -  South -> S,Sth
     -  Southeast -> SE
     -  Southwest -> SW
-    -  Spur -> Spur
+    -  Spring -> Spg
+    -  Springs -> Spgs
+    -  Spurs -> Spur
     -  Square -> Sq
+    -  Squares -> Sqs
     -  Stairway -> Strwy
-    -  State Highway -> SH
-    -  State Highway -> SHwy
+    -  State Highway -> SH,SHwy
     -  State Route -> SR
-    -  Station -> Sta
-    -  Station -> Stn
-    -  Strand -> Sd
-    -  Strand -> Stra
+    -  Station -> Sta,Stn
+    -  Strand -> Sd,Stra
+    -  Stravenue -> Stra
+    -  Stream -> Strm
     -  Street -> St
+    -  Streets -> Sts
     -  Strip -> Strp
     -  Subway -> Sbwy
+    -  Summit -> Smt
     -  Tarn -> Tn
-    -  Tarn -> Tarn
     -  Terminal -> Term
-    -  Terrace -> Tce
-    -  Terrace -> Ter
-    -  Terrace -> Terr
-    -  Thoroughfare -> Thfr
-    -  Thoroughfare -> Thor
-    -  Tollway -> Tlwy
-    -  Tollway -> Twy
-    -  Top -> Top
-    -  Tor -> Tor
+    -  Terrace -> Tce,Ter,Terr
+    -  Thoroughfare -> Thfr,Thor
+    -  Throughway -> Trwy
+    -  Tollway -> Tlwy,Twy
     -  Towers -> Twrs
     -  Township -> Twp
     -  Trace -> Trce
-    -  Track -> Tr
-    -  Track -> Trk
+    -  Track -> Tr,Trak,Trk
+    -  Trafficway -> Trfy
     -  Trail -> Trl
     -  Trailer -> Trlr
     -  Triangle -> Tri
     -  Trunkway -> Tkwy
-    -  Tunnel -> Tun
-    -  Turn -> Tn
-    -  Turn -> Trn
-    -  Turn -> Turn
-    -  Turnpike -> Tpk
-    -  Turnpike -> Tpke
-    -  Underpass -> Upas
-    -  Underpass -> Ups
-    -  University -> Uni
-    -  University -> Univ
+    -  Tunnel -> Tun,Tunl
+    -  Turn -> Tn,Trn
+    -  Turnpike -> Tpk,Tpke
+    -  Underpass -> Upas,Ups
+    -  Union -> Un
+    -  Unions -> Uns
+    -  University -> Uni,Univ
     -  Upper -> Up
     -  Upper -> Upr
     -  Vale -> Va
-    -  Vale -> Vale
+    -  Valley -> Vly
     -  Valley -> Vy
-    -  Viaduct -> Vdct
-    -  Viaduct -> Via
-    -  Viaduct -> Viad
+    -  Valleys -> Vlys
+    -  Viaduct$ -> Vdct,Via,Viad
     -  View -> Vw
-    -  View -> View
-    -  Village -> Vill
+    -  Views -> Vws
+    -  Village -> Vill,Vlg
+    -  Villages -> Vlgs
     -  Villas -> Vlls
-    -  Vista -> Vst
-    -  Vista -> Vsta
-    -  Walk -> Walk
-    -  Walk -> Wk
-    -  Walk -> Wlk
-    -  Walkway -> Wkwy
-    -  Walkway -> Wky
+    -  Ville -> Vl
+    -  Vista -> Vis,Vst,Vsta
+    -  Walk -> Wk,Wlk
+    -  Walks -> Walk
+    -  Walkway -> Wkwy,Wky
     -  Waters -> Wtr
-    -  Way -> Way
     -  Way -> Wy
+    -  Well -> Wl
+    -  Wells -> Wls
     -  West -> W
     -  Wharf -> Whrf
     -  William -> Wm
     -  Wynd -> Wyn
-    -  Wynd -> Wynd
-    -  Yard -> Yard
     -  Yard -> Yd
 - lang: en
   country: ca

settings/icu-rules/variants-es.yaml

@@ -30,7 +30,6 @@
     -  Bloque -> Blq
     -  Bulevar -> Blvr
     -  Boulevard -> Blvd
-    -  Calle -> C/
     -  Calle -> C
     -  Calle -> Cl
     -  Calleja -> Cllja

settings/icu-rules/variants-fr.yaml

@@ -3,20 +3,16 @@
   words:
     -  Abbaye -> ABE
     -  Agglomération -> AGL
-    -  Aire -> AIRE
     -  Aires -> AIRE
     -  Allée -> ALL
-    -  Allée -> All
     -  Allées -> ALL
     -  Ancien chemin -> ACH
     -  Ancienne route -> ART
     -  Anciennes routes -> ART
-    -  Anse -> ANSE
     -  Arcade -> ARC
     -  Arcades -> ARC
     -  Autoroute -> AUT
     -  Avenue -> AV
-    -  Avenue -> Av
     -  Barrière -> BRE
     -  Barrières -> BRE
     -  Bas chemin -> BCH
@@ -28,16 +24,11 @@
     -  Berges -> BER
     -  Bois -> BOIS
     -  Boucle -> BCLE
-    -  Boulevard -> Bd
     -  Boulevard -> BD
     -  Bourg -> BRG
     -  Butte -> BUT
-    -  Cité -> CITE
     -  Cités -> CITE
-    -  Côte -> COTE
     -  Côteau -> COTE
-    -  Cale -> CALE
-    -  Camp -> CAMP
     -  Campagne -> CGNE
     -  Camping -> CPG
     -  Carreau -> CAU
@@ -56,17 +47,13 @@
     -  Chaussées -> CHS
     -  Chemin -> Ch
     -  Chemin -> CHE
-    -  Chemin -> Che
     -  Chemin vicinal -> CHV
     -  Cheminement -> CHEM
     -  Cheminements -> CHEM
     -  Chemins -> CHE
     -  Chemins vicinaux -> CHV
-    -  Chez -> CHEZ
     -  Château -> CHT
     -  Cloître -> CLOI
-    -  Clos -> CLOS
-    -  Col -> COL
     -  Colline -> COLI
     -  Collines -> COLI
     -  Contour -> CTR
@@ -74,9 +61,7 @@
     -  Corniches -> COR
     -  Cottage -> COTT
     -  Cottages -> COTT
-    -  Cour -> COUR
     -  Cours -> CRS
-    -  Cours -> Crs
     -  Darse -> DARS
     -  Degré -> DEG
     -  Degrés -> DEG
@@ -87,11 +72,8 @@
     -  Domaine -> DOM
     -  Domaines -> DOM
     -  Écluse -> ECL
-    -  Écluse -> ÉCL
     -  Écluses -> ECL
-    -  Écluses -> ÉCL
     -  Église -> EGL
-    -  Église -> ÉGL
     -  Enceinte -> EN
     -  Enclave -> ENV
     -  Enclos -> ENC
@@ -100,21 +82,16 @@
     -  Espace -> ESPA
     -  Esplanade -> ESP
     -  Esplanades -> ESP
-    -  Étang -> ETANG
-    -  Étang -> ÉTANG
     -  Faubourg -> FG
-    -  Faubourg -> Fg
     -  Ferme -> FRM
     -  Fermes -> FRM
     -  Fontaine -> FON
-    -  Fort -> FORT
     -  Forum -> FORM
     -  Fosse -> FOS
     -  Fosses -> FOS
     -  Foyer -> FOYR
     -  Galerie -> GAL
     -  Galeries -> GAL
-    -  Gare -> GARE
     -  Garenne -> GARN
     -  Grand boulevard -> GBD
     -  Grand ensemble -> GDEN
@@ -134,13 +111,9 @@
     -  Haut chemin -> HCH
     -  Hauts chemins -> HCH
     -  Hippodrome -> HIP
-    -  HLM -> HLM
-    -  Île -> ILE
-    -  Île -> ÎLE
     -  Immeuble -> IMM
     -  Immeubles -> IMM
     -  Impasse -> IMP
-    -  Impasse -> Imp
     -  Impasses -> IMP
     -  Jardin -> JARD
     -  Jardins -> JARD
@@ -150,13 +123,11 @@
     -  Lieu-dit -> LD
     -  Lotissement -> LOT
     -  Lotissements -> LOT
-    -  Mail -> MAIL
     -  Maison forestière -> MF
     -  Manoir -> MAN
     -  Marche -> MAR
     -  Marches -> MAR
     -  Maréchal -> MAL
-    -  Mas -> MAS
     -  Monseigneur -> Mgr
     -  Mont -> Mt
     -  Montée -> MTE
@@ -168,13 +139,9 @@
     -  Métro -> MÉT
     -  Nouvelle route -> NTE
     -  Palais -> PAL
-    -  Parc -> PARC
-    -  Parcs -> PARC
     -  Parking -> PKG
     -  Parvis -> PRV
     -  Passage -> PAS
-    -  Passage -> Pas
-    -  Passage -> Pass
     -  Passage à niveau -> PN
     -  Passe -> PASS
     -  Passerelle -> PLE
@@ -191,19 +158,14 @@
     -  Petite rue -> PTR
     -  Petites allées -> PTA
     -  Place -> PL
-    -  Place -> Pl
     -  Placis -> PLCI
     -  Plage -> PLAG
     -  Plages -> PLAG
     -  Plaine -> PLN
-    -  Plan -> PLAN
     -  Plateau -> PLT
     -  Plateaux -> PLT
     -  Pointe -> PNT
-    -  Pont -> PONT
-    -  Ponts -> PONT
     -  Porche -> PCH
-    -  Port -> PORT
     -  Porte -> PTE
     -  Portique -> PORQ
     -  Portiques -> PORQ
@@ -211,25 +173,19 @@
     -  Pourtour -> POUR
     -  Presqu’île -> PRQ
     -  Promenade -> PROM
-    -  Promenade -> Prom
-    -  Pré -> PRE
-    -  Pré -> PRÉ
     -  Périphérique -> PERI
     -  Péristyle -> PSTY
     -  Quai -> QU
-    -  Quai -> Qu
     -  Quartier -> QUA
     -  Raccourci -> RAC
     -  Raidillon -> RAID
     -  Rampe -> RPE
     -  Rempart -> REM
-    -  Roc -> ROC
     -  Rocade -> ROC
     -  Rond point -> RPT
     -  Roquet -> ROQT
     -  Rotonde -> RTD
     -  Route -> RTE
-    -  Route -> Rte
     -  Routes -> RTE
     -  Rue -> R
     -  Rue -> R
@@ -245,7 +201,6 @@
     -  Sentier -> SEN
     -  Sentiers -> SEN
     -  Square -> SQ
-    -  Square -> Sq
     -  Stade -> STDE
     -  Station -> STA
     -  Terrain -> TRN
@@ -254,13 +209,11 @@
     -  Terre plein -> TPL
     -  Tertre -> TRT
     -  Tertres -> TRT
-    -  Tour -> TOUR
     -  Traverse -> TRA
     -  Vallon -> VAL
     -  Vallée -> VAL
     -  Venelle -> VEN
     -  Venelles -> VEN
-    -  Via -> VIA
     -  Vieille route -> VTE
     -  Vieux chemin -> VCHE
     -  Villa -> VLA
@@ -269,7 +222,6 @@
     -  Villas -> VLA
     -  Voie -> VOI
     -  Voies -> VOI
-    -  Zone -> ZONE
     -  Zone artisanale -> ZA
     -  Zone d'aménagement concerté -> ZAC
     -  Zone d'aménagement différé -> ZAD
@@ -289,7 +241,6 @@
     -  Esplanade -> ESPL
     -  Passage -> PASS
     -  Plateau -> PLAT
-    -  Rang -> RANG
     -  Rond-point -> RDPT
     -  Sentier -> SENT
     -  Subdivision -> SUBDIV

settings/icu-rules/variants-it.yaml

@@ -29,7 +29,6 @@
     -  Prima -> I
     -  Primo -> I
     -  Primo -> 1
-    -  Primo -> 1°
     -  Quarta -> IV
     -  Quarto -> IV
     -  Quattro -> IV

settings/icu-rules/variants-no.yaml

@@ -1,11 +1,10 @@
 # Source: https://wiki.openstreetmap.org/wiki/Name_finder:Abbreviations#Norsk_-_Norwegian
-- lang: no
+- lang: "no"
   words:
     # convert between Nynorsk and Bookmal here
-    -  vei, veg => v,vn,vei,veg
-    -  veien, vegen -> v,vn,veien,vegen
-    -  gate -> g,gt
+    -  ~vei, ~veg -> v,vei,veg
+    -  ~veien, ~vegen -> vn,veien,vegen
     # convert between the two female forms
-    -  gaten, gata => g,gt,gaten,gata
+    -  gate, gaten, gata -> g,gt
     -  plass, plassen -> pl
     -  sving, svingen -> sv

settings/icu-rules/variants-ru.yaml

@@ -1,14 +1,128 @@
 # Source: https://wiki.openstreetmap.org/wiki/Name_finder:Abbreviations#.D0.A0.D1.83.D1.81.D1.81.D0.BA.D0.B8.D0.B9_-_Russian
+# Source: https://www.plantarium.ru/page/help/topic/abbreviations.html
+# Source: https://dic.academic.ru/dic.nsf/ruwiki/1871310
 - lang: ru
   words:
+    -  Академик, Академика -> Ак
+    -  акционерное общество -> АО
     -  аллея -> ал
+    -  архипелаг -> арх
+    -  атомная электростанция -> АЭС
+    -  аэродром -> аэрд
+    -  аэропорт -> аэрп
+    -  Башкирский, Башкирская, Башкирское, Башкирские -> Баш, Башк, Башкир
+    -  Белый, Белая, Белое. Белые -> Бел
+    -  болото -> бол
+    -  больница -> больн
+    -  Большой, Большая, Большое, Большие -> Б, Бол
+    -  брод -> бр
     -  бульвар -> бул
+    -  бухта -> бух
+    -  бывший, бывшая, бывшее, бывшие -> бывш
+    -  Великий, Великая, Великое, Великие -> Вел
+    -  Верхний, Верхняя, Верхнее, Верхние -> В, Верх
+    -  водокачка -> вдкч
+    -  водопад -> вдп
+    -  водохранилище -> вдхр
+    -  вокзал -> вкз, вокз
+    -  Восточный, Восточная, Восточное, Восточные -> В, Вост
+    -  вулкан -> влк
+    -  гидроэлектростанция -> ГЭС
+    -  гора -> г
+    -  город -> г
+    -  дворец культуры, дом культуры -> ДК
+    -  дворец спорта -> ДС
+    -  деревня -> д, дер
+    -  детский оздоровительный лагерь -> ДОЛ
+    -  дом -> д
+    -  дом отдыха -> Д О
+    -  железная дорога -> ж д
+    -  железнодорожный, железнодорожная, железнодорожное -> ж-д
+    -  железобетонных изделий -> ЖБИ
+    -  жилой комплекс -> ЖК
+    -  завод -> з-д
+    -  закрытое административно-территориальное образование -> ЗАТО
+    -  залив -> зал
+    -  Западный, Западная, Западное, Западные -> З, Зап, Запад
+    -  заповедник -> запов
+    -  имени -> им
+    -  институт -> инст
+    -  исправительная колония -> ИК
+    -  километр -> км
+    -  Красный, Красная, Красное, Красные -> Кр, Крас
+    -  лагерь -> лаг
+    -  Левый, Левая,Левое, Левые -> Л, Лев
+    -  ледник -> ледн
+    -  лесничество -> леснич
+    -  лесной, лесная, лесное -> лес
+    -  линия электропередачи -> ЛЭП
+    -  Малый, Малая, Малое, Малые -> М, Мал
+    -  Мордовский, Мордовская, Мордовское, Мордовские -> Мордов
+    -  морской, морская, морское -> мор
+    -  Московский, Московская, Московское, Московские -> Мос, Моск
+    -  мыс -> м
     -  набережная -> наб
+    -  Нижний, Нижняя, Нижнее, Нижние -> Ниж, Н
+    -  Новый, Новая, Новое, Новые -> Нов, Н
+    -  обгонный пункт -> обг п
+    -  область -> обл
+    -  озеро -> оз
+    -  особо охраняемая природная территория -> ООПТ
+    -  остановочный пункт -> о п
+    -  остров -> о
+    -  острова -> о-ва
+    -  парк культуры и отдыха -> ПКиО
+    -  перевал -> пер
     -  переулок -> пер
+    -  пещера -> пещ
+    -  пионерский лагерь -> пионерлаг
+    -  платформа -> пл, платф
     -  площадь -> пл
+    -  подсобное хозяйство -> подсоб хоз
+    -  полуостров -> п-ов
+    -  посёлок -> пос, п
+    -  посёлок городского типа -> п г т, пгт
+    -  Правый, Правая, Правое, Правые -> П, Пр, Прав
     -  проезд -> пр
     -  проспект -> просп
-    -  шоссе -> ш
+    -  пруд -> пр
+    -  пустыня -> пуст
+    -  разъезд -> рзд
+    -  район -> р-н
+    -  резинотехнических изделий -> РТИ
+    -  река -> р
+    -  речной, речная, речное -> реч, речн
+    -  Российский, Российская, Российское, Российские -> Рос
+    -  Русский, Русская, Русское, Русские -> Рус, Русск
+    -  ручей -> руч
+    -  садовое некоммерческое товарищество -> СНТ
+    -  садовые участки -> сад уч
+    -  санаторий -> сан
+    -  сарай -> сар
+    -  Северный, Северная, Северное, Северные -> С, Сев
+    -  село -> с
+    -  Сибирский, Сибирская, Сибирское, Сибирские -> Сиб
+    -  Советский, Советская, Советское, Советские -> Сов
+    -  совхоз -> свх
+    -  Сортировочный, Сортировочная, Сортировочное, Сортировочные -> Сорт
+    -  станция -> ст
+    -  Старый, Старая, Среднее, Средние -> Ср
+    -  Татарский, Татарская, Татарское, Татарские -> Тат, Татар
+    -  теплоэлекстростанция -> ТЭС
+    -  теплоэлектроцентраль -> ТЭЦ
+    -  техникум -> техн
+    -  тоннель, туннель -> тун
     -  тупик -> туп
     -  улица -> ул
-    -  область -> обл
+    -  Уральский, Уральская, Уральское, Уральские -> Ур, Урал
+    -  урочище -> ур
+    -  хозяйство -> хоз, хоз-во
+    -  хребет -> хр
+    -  хутор -> хут
+    -  Чёрный, Чёрная, Чёрное, Чёрные -> Черн
+    -  Чувашский, Чувашская, Чувашское, Чувашские -> Чуваш
+    -  шахта -> шах
+    -  школа -> шк
+    -  шоссе -> ш
+    -  элеватор -> элев
+    -  Южный, Южная, Южное, Южные -> Ю, Юж, Южн

settings/icu_tokenizer.yaml

@@ -46,7 +46,7 @@ sanitizers:
     - step: strip-brace-terms
     - step: tag-analyzer-by-language
       filter-kind: [".*name.*"]
-      whitelist: [bg,ca,cs,da,de,el,en,es,et,eu,fi,fr,gl,hu,it,ja,mg,ms,nl,no,pl,pt,ro,ru,sk,sl,sv,tr,uk,vi]
+      whitelist: [bg,ca,cs,da,de,el,en,es,et,eu,fi,fr,gl,hu,it,ja,mg,ms,nl,"no",pl,pt,ro,ru,sk,sl,sv,tr,uk,vi]
       use-defaults: all
       mode: append
     - step: tag-japanese
@@ -158,7 +158,7 @@ token-analysis:
       mode: variant-only
       variants:
           - !include icu-rules/variants-nl.yaml
-    - id: no
+    - id: "no"
       analyzer: generic
       mode: variant-only
       variants:

src/nominatim_api/core.py

@@ -26,7 +26,7 @@ from .connection import SearchConnection
 from .status import get_status, StatusResult
 from .lookup import get_places, get_detailed_place
 from .reverse import ReverseGeocoder
-from .search import ForwardGeocoder, Phrase, PhraseType, make_query_analyzer
+from . import search as nsearch
 from . import types as ntyp
 from .results import DetailedResult, ReverseResult, SearchResults
 
@@ -207,7 +207,7 @@ class NominatimAPIAsync:
         async with self.begin() as conn:
             conn.set_query_timeout(self.query_timeout)
             if details.keywords:
-                await make_query_analyzer(conn)
+                await nsearch.make_query_analyzer(conn)
             return await get_detailed_place(conn, place, details)
 
     async def lookup(self, places: Sequence[ntyp.PlaceRef], **params: Any) -> SearchResults:
@@ -219,7 +219,7 @@ class NominatimAPIAsync:
         async with self.begin() as conn:
             conn.set_query_timeout(self.query_timeout)
             if details.keywords:
-                await make_query_analyzer(conn)
+                await nsearch.make_query_analyzer(conn)
             return await get_places(conn, places, details)
 
     async def reverse(self, coord: ntyp.AnyPoint, **params: Any) -> Optional[ReverseResult]:
@@ -237,7 +237,7 @@ class NominatimAPIAsync:
         async with self.begin() as conn:
             conn.set_query_timeout(self.query_timeout)
             if details.keywords:
-                await make_query_analyzer(conn)
+                await nsearch.make_query_analyzer(conn)
             geocoder = ReverseGeocoder(conn, details,
                                        self.reverse_restrict_to_country_area)
             return await geocoder.lookup(coord)
@@ -251,10 +251,10 @@ class NominatimAPIAsync:
 
         async with self.begin() as conn:
             conn.set_query_timeout(self.query_timeout)
-            geocoder = ForwardGeocoder(conn, ntyp.SearchDetails.from_kwargs(params),
-                                       self.config.get_int('REQUEST_TIMEOUT')
-                                       if self.config.REQUEST_TIMEOUT else None)
-            phrases = [Phrase(PhraseType.NONE, p.strip()) for p in query.split(',')]
+            geocoder = nsearch.ForwardGeocoder(conn, ntyp.SearchDetails.from_kwargs(params),
+                                               self.config.get_int('REQUEST_TIMEOUT')
+                                               if self.config.REQUEST_TIMEOUT else None)
+            phrases = [nsearch.Phrase(nsearch.PHRASE_ANY, p.strip()) for p in query.split(',')]
             return await geocoder.lookup(phrases)
 
     async def search_address(self, amenity: Optional[str] = None,
@@ -271,22 +271,22 @@ class NominatimAPIAsync:
             conn.set_query_timeout(self.query_timeout)
             details = ntyp.SearchDetails.from_kwargs(params)
 
-            phrases: List[Phrase] = []
+            phrases: List[nsearch.Phrase] = []
 
             if amenity:
-                phrases.append(Phrase(PhraseType.AMENITY, amenity))
+                phrases.append(nsearch.Phrase(nsearch.PHRASE_AMENITY, amenity))
             if street:
-                phrases.append(Phrase(PhraseType.STREET, street))
+                phrases.append(nsearch.Phrase(nsearch.PHRASE_STREET, street))
             if city:
-                phrases.append(Phrase(PhraseType.CITY, city))
+                phrases.append(nsearch.Phrase(nsearch.PHRASE_CITY, city))
             if county:
-                phrases.append(Phrase(PhraseType.COUNTY, county))
+                phrases.append(nsearch.Phrase(nsearch.PHRASE_COUNTY, county))
             if state:
-                phrases.append(Phrase(PhraseType.STATE, state))
+                phrases.append(nsearch.Phrase(nsearch.PHRASE_STATE, state))
             if postalcode:
-                phrases.append(Phrase(PhraseType.POSTCODE, postalcode))
+                phrases.append(nsearch.Phrase(nsearch.PHRASE_POSTCODE, postalcode))
             if country:
-                phrases.append(Phrase(PhraseType.COUNTRY, country))
+                phrases.append(nsearch.Phrase(nsearch.PHRASE_COUNTRY, country))
 
             if not phrases:
                 raise UsageError('Nothing to search for.')
@@ -304,14 +304,14 @@ class NominatimAPIAsync:
             else:
                 details.restrict_min_max_rank(4, 4)
 
-            if 'layers' not in params:
+            if details.layers is None:
                 details.layers = ntyp.DataLayer.ADDRESS
                 if amenity:
                     details.layers |= ntyp.DataLayer.POI
 
-            geocoder = ForwardGeocoder(conn, details,
-                                       self.config.get_int('REQUEST_TIMEOUT')
-                                       if self.config.REQUEST_TIMEOUT else None)
+            geocoder = nsearch.ForwardGeocoder(conn, details,
+                                               self.config.get_int('REQUEST_TIMEOUT')
+                                               if self.config.REQUEST_TIMEOUT else None)
             return await geocoder.lookup(phrases)
 
     async def search_category(self, categories: List[Tuple[str, str]],
@@ -328,15 +328,15 @@ class NominatimAPIAsync:
         async with self.begin() as conn:
             conn.set_query_timeout(self.query_timeout)
             if near_query:
-                phrases = [Phrase(PhraseType.NONE, p) for p in near_query.split(',')]
+                phrases = [nsearch.Phrase(nsearch.PHRASE_ANY, p) for p in near_query.split(',')]
             else:
                 phrases = []
                 if details.keywords:
-                    await make_query_analyzer(conn)
+                    await nsearch.make_query_analyzer(conn)
 
-            geocoder = ForwardGeocoder(conn, details,
-                                       self.config.get_int('REQUEST_TIMEOUT')
-                                       if self.config.REQUEST_TIMEOUT else None)
+            geocoder = nsearch.ForwardGeocoder(conn, details,
+                                               self.config.get_int('REQUEST_TIMEOUT')
+                                               if self.config.REQUEST_TIMEOUT else None)
             return await geocoder.lookup_pois(categories, phrases)
 
 

src/nominatim_api/localization.py

@@ -8,6 +8,7 @@
 Helper functions for localizing names of results.
 """
 from typing import Mapping, List, Optional
+from .config import Configuration
 
 import re
 
@@ -20,14 +21,18 @@ class Locales:
     """
 
     def __init__(self, langs: Optional[List[str]] = None):
+        self.config = Configuration(None)
         self.languages = langs or []
         self.name_tags: List[str] = []
 
-        # Build the list of supported tags. It is currently hard-coded.
-        self._add_lang_tags('name')
-        self._add_tags('name', 'brand')
-        self._add_lang_tags('official_name', 'short_name')
-        self._add_tags('official_name', 'short_name', 'ref')
+        parts = self.config.OUTPUT_NAMES.split(',')
+
+        for part in parts:
+            part = part.strip()
+            if part.endswith(":XX"):
+                self._add_lang_tags(part[:-3])
+            else:
+                self._add_tags(part)
 
     def __bool__(self) -> bool:
         return len(self.languages) > 0

src/nominatim_api/logging.py

@@ -2,7 +2,7 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Functions for specialised logging with HTML output.
@@ -106,9 +106,6 @@ class BaseLogger:
                 except TypeError:
                     return sqlstr
 
-            # Fixes an odd issue with Python 3.7 where percentages are not
-            # quoted correctly.
-            sqlstr = re.sub(r'%(?!\()', '%%', sqlstr)
             sqlstr = re.sub(r'__\[POSTCOMPILE_([^]]*)\]', r'%(\1)s', sqlstr)
             return sqlstr % params
 
@@ -342,7 +339,8 @@ HTML_HEADER: str = """<!DOCTYPE html>
   <title>Nominatim - Debug</title>
   <style>
 """ + \
-    (HtmlFormatter(nobackground=True).get_style_defs('.highlight') if CODE_HIGHLIGHT else '') + \
+    (HtmlFormatter(nobackground=True).get_style_defs('.highlight')  # type: ignore[no-untyped-call]
+     if CODE_HIGHLIGHT else '') + \
     """
     h2 { font-size: x-large }
 

src/nominatim_api/query_preprocessing/normalize.py

@@ -27,5 +27,5 @@ def create(config: QueryConfig) -> QueryProcessingFunc:
 
     return lambda phrases: list(
         filter(lambda p: p.text,
-               (Phrase(p.ptype, cast(str, normalizer.transliterate(p.text)))
+               (Phrase(p.ptype, cast(str, normalizer.transliterate(p.text)).strip('-: '))
                 for p in phrases)))

src/nominatim_api/query_preprocessing/regex_replace.py

@@ -0,0 +1,52 @@
+# SPDX-License-Identifier: GPL-3.0-or-later
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2025 by the Nominatim developer community.
+# For a full list of authors see the git log.
+"""
+This preprocessor replaces values in a given input based on pre-defined regex rules.
+
+Arguments:
+    pattern: Regex pattern to be applied on the input
+    replace: The string that it is to be replaced with
+"""
+from typing import List
+import re
+
+from .config import QueryConfig
+from .base import QueryProcessingFunc
+from ..search.query import Phrase
+
+
+class _GenericPreprocessing:
+    """Perform replacements to input phrases using custom regex patterns."""
+
+    def __init__(self, config: QueryConfig) -> None:
+        """Initialise the _GenericPreprocessing class with patterns from the ICU config file."""
+        self.config = config
+
+        match_patterns = self.config.get('replacements', 'Key not found')
+        self.compiled_patterns = [
+            (re.compile(item['pattern']), item['replace']) for item in match_patterns
+            ]
+
+    def split_phrase(self, phrase: Phrase) -> Phrase:
+        """This function performs replacements on the given text using regex patterns."""
+        for item in self.compiled_patterns:
+            phrase.text = item[0].sub(item[1], phrase.text)
+
+        return phrase
+
+    def __call__(self, phrases: List[Phrase]) -> List[Phrase]:
+        """
+        Return the final Phrase list.
+        Returns an empty list if there is nothing left after split_phrase.
+        """
+        result = [p for p in map(self.split_phrase, phrases) if p.text.strip()]
+        return result
+
+
+def create(config: QueryConfig) -> QueryProcessingFunc:
+    """ Create a function for generic preprocessing."""
+    return _GenericPreprocessing(config)

src/nominatim_api/search/__init__.py

@@ -9,5 +9,12 @@ Module for forward search.
 """
 from .geocoder import (ForwardGeocoder as ForwardGeocoder)
 from .query import (Phrase as Phrase,
-                    PhraseType as PhraseType)
+                    PHRASE_ANY as PHRASE_ANY,
+                    PHRASE_AMENITY as PHRASE_AMENITY,
+                    PHRASE_STREET as PHRASE_STREET,
+                    PHRASE_CITY as PHRASE_CITY,
+                    PHRASE_COUNTY as PHRASE_COUNTY,
+                    PHRASE_STATE as PHRASE_STATE,
+                    PHRASE_POSTCODE as PHRASE_POSTCODE,
+                    PHRASE_COUNTRY as PHRASE_COUNTRY)
 from .query_analyzer_factory import (make_query_analyzer as make_query_analyzer)

src/nominatim_api/search/db_search_builder.py

@@ -2,7 +2,7 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Conversion from token assignment to an abstract DB search.
@@ -11,7 +11,7 @@ from typing import Optional, List, Tuple, Iterator, Dict
 import heapq
 
 from ..types import SearchDetails, DataLayer
-from .query import QueryStruct, Token, TokenType, TokenRange, BreakType
+from . import query as qmod
 from .token_assignment import TokenAssignment
 from . import db_search_fields as dbf
 from . import db_searches as dbs
@@ -51,7 +51,7 @@ class SearchBuilder:
     """ Build the abstract search queries from token assignments.
     """
 
-    def __init__(self, query: QueryStruct, details: SearchDetails) -> None:
+    def __init__(self, query: qmod.QueryStruct, details: SearchDetails) -> None:
         self.query = query
         self.details = details
 
@@ -97,7 +97,7 @@ class SearchBuilder:
                 builder = self.build_poi_search(sdata)
             elif assignment.housenumber:
                 hnr_tokens = self.query.get_tokens(assignment.housenumber,
-                                                   TokenType.HOUSENUMBER)
+                                                   qmod.TOKEN_HOUSENUMBER)
                 builder = self.build_housenumber_search(sdata, hnr_tokens, assignment.address)
             else:
                 builder = self.build_special_search(sdata, assignment.address,
@@ -128,7 +128,7 @@ class SearchBuilder:
             yield dbs.PoiSearch(sdata)
 
     def build_special_search(self, sdata: dbf.SearchData,
-                             address: List[TokenRange],
+                             address: List[qmod.TokenRange],
                              is_category: bool) -> Iterator[dbs.AbstractSearch]:
         """ Build abstract search queries for searches that do not involve
             a named place.
@@ -146,45 +146,51 @@ class SearchBuilder:
             if address:
                 sdata.lookups = [dbf.FieldLookup('nameaddress_vector',
                                                  [t.token for r in address
-                                                  for t in self.query.get_partials_list(r)],
+                                                  for t in self.query.iter_partials(r)],
                                                  lookups.Restrict)]
-                penalty += 0.2
             yield dbs.PostcodeSearch(penalty, sdata)
 
-    def build_housenumber_search(self, sdata: dbf.SearchData, hnrs: List[Token],
-                                 address: List[TokenRange]) -> Iterator[dbs.AbstractSearch]:
+    def build_housenumber_search(self, sdata: dbf.SearchData, hnrs: List[qmod.Token],
+                                 address: List[qmod.TokenRange]) -> Iterator[dbs.AbstractSearch]:
         """ Build a simple address search for special entries where the
             housenumber is the main name token.
         """
-        sdata.lookups = [dbf.FieldLookup('name_vector', [t.token for t in hnrs], lookups.LookupAny)]
-        expected_count = sum(t.count for t in hnrs)
-
-        partials = {t.token: t.addr_count for trange in address
-                    for t in self.query.get_partials_list(trange)}
+        partials = dbf.CountedTokenIDs((t for trange in address
+                                        for t in self.query.iter_partials(trange)),
+                                       'addr_count')
 
         if not partials:
             # can happen when none of the partials is indexed
             return
 
-        if expected_count < 8000:
-            sdata.lookups.append(dbf.FieldLookup('nameaddress_vector',
-                                                 list(partials), lookups.Restrict))
-        elif len(partials) != 1 or list(partials.values())[0] < 10000:
-            sdata.lookups.append(dbf.FieldLookup('nameaddress_vector',
-                                                 list(partials), lookups.LookupAll))
+        expected_count = sum(t.count for t in hnrs)
+        hnr_tokens = [t.token for t in hnrs]
+
+        if expected_count < 10000:
+            sdata.lookups = [dbf.FieldLookup('name_vector', hnr_tokens, lookups.LookupAny),
+                             dbf.FieldLookup('nameaddress_vector',
+                                             partials.get_tokens(),
+                                             lookups.Restrict)]
         else:
-            addr_fulls = [t.token for t
-                          in self.query.get_tokens(address[0], TokenType.WORD)]
-            if len(addr_fulls) > 5:
-                return
-            sdata.lookups.append(
-                dbf.FieldLookup('nameaddress_vector', addr_fulls, lookups.LookupAny))
+            split = partials.get_num_lookup_tokens(20000, 5)
+            if split > 0:
+                sdata.lookups = partials.split_lookup(split, 'nameaddress_vector')
+                sdata.lookups.append(
+                    dbf.FieldLookup('name_vector', hnr_tokens, lookups.Restrict))
+            else:
+                addr_fulls = [t.token for t in
+                              self.query.get_tokens(address[0], qmod.TOKEN_WORD)]
+                if len(addr_fulls) > 5:
+                    return
+                sdata.lookups = [
+                    dbf.FieldLookup('name_vector', hnr_tokens, lookups.LookupAny),
+                    dbf.FieldLookup('nameaddress_vector', addr_fulls, lookups.LookupAny)]
 
         sdata.housenumbers = dbf.WeightedStrings([], [])
-        yield dbs.PlaceSearch(0.05, sdata, expected_count)
+        yield dbs.PlaceSearch(0.05, sdata, expected_count, True)
 
     def build_name_search(self, sdata: dbf.SearchData,
-                          name: TokenRange, address: List[TokenRange],
+                          name: qmod.TokenRange, address: List[qmod.TokenRange],
                           is_category: bool) -> Iterator[dbs.AbstractSearch]:
         """ Build abstract search queries for simple name or address searches.
         """
@@ -195,48 +201,102 @@ class SearchBuilder:
                 sdata.rankings.append(ranking)
             for penalty, count, lookup in self.yield_lookups(name, address):
                 sdata.lookups = lookup
-                yield dbs.PlaceSearch(penalty + name_penalty, sdata, count)
+                if sdata.housenumbers:
+                    yield dbs.AddressSearch(penalty + name_penalty, sdata, count, bool(address))
+                else:
+                    yield dbs.PlaceSearch(penalty + name_penalty, sdata, count, bool(address))
 
-    def yield_lookups(self, name: TokenRange, address: List[TokenRange]
+    def yield_lookups(self, name: qmod.TokenRange, address: List[qmod.TokenRange]
                       ) -> Iterator[Tuple[float, int, List[dbf.FieldLookup]]]:
         """ Yield all variants how the given name and address should best
             be searched for. This takes into account how frequent the terms
             are and tries to find a lookup that optimizes index use.
         """
+        name_partials = dbf.CountedTokenIDs(self.query.iter_partials(name))
+        addr_partials = dbf.CountedTokenIDs((t for r in address
+                                             for t in self.query.iter_partials(r)),
+                                            'addr_count')
+
+        if not addr_partials:
+            yield from self.yield_name_only_lookups(name_partials, name)
+        else:
+            yield from self.yield_address_lookups(name_partials, addr_partials, name)
+
+    def yield_name_only_lookups(self, partials: dbf.CountedTokenIDs, name: qmod.TokenRange
+                                ) -> Iterator[Tuple[float, int, List[dbf.FieldLookup]]]:
+        """ Yield the best lookup for a name-only search.
+        """
+        split = partials.get_num_lookup_tokens(30000, 6)
+
+        if split > 0:
+            yield 0.0, partials.expected_for_all_search(5), \
+                  partials.split_lookup(split, 'name_vector')
+        else:
+            # lots of results expected: try lookup by full names first
+            name_fulls = list(filter(lambda t: t.count < 50000,
+                                     self.query.get_tokens(name, qmod.TOKEN_WORD)))
+            if name_fulls:
+                yield 0.0, sum(t.count for t in name_fulls), \
+                      dbf.lookup_by_any_name([t.token for t in name_fulls], [], [])
+
+            # look the name up by its partials
+            exp_count = partials.expected_for_all_search(5)
+            if exp_count < 50000:
+                yield 1.0, exp_count, \
+                      [dbf.FieldLookup('name_vector', partials.get_tokens(), lookups.LookupAll)]
+
+    def yield_address_lookups(self, name_partials: dbf.CountedTokenIDs,
+                              addr_partials: dbf.CountedTokenIDs, name: qmod.TokenRange,
+                              ) -> Iterator[Tuple[float, int, List[dbf.FieldLookup]]]:
         penalty = 0.0  # extra penalty
-        name_partials = {t.token: t for t in self.query.get_partials_list(name)}
 
-        addr_partials = [t for r in address for t in self.query.get_partials_list(r)]
-        addr_tokens = list({t.token for t in addr_partials})
+        name_split = name_partials.get_num_lookup_tokens(20000, 6)
+        addr_split = addr_partials.get_num_lookup_tokens(10000, 3)
 
-        exp_count = min(t.count for t in name_partials.values()) / (2**(len(name_partials) - 1))
+        if name_split < 0 and addr_split < 0:
+            # Partial term too frequent. Try looking up by rare full names first.
+            name_fulls = self.query.get_tokens(name, qmod.TOKEN_WORD)
+            if name_fulls:
+                fulls_count = sum(t.count for t in name_fulls)
 
-        if (len(name_partials) > 3 or exp_count < 8000):
-            yield penalty, exp_count, dbf.lookup_by_names(list(name_partials.keys()), addr_tokens)
-            return
+                if fulls_count < 80000:
+                    yield 0.0, fulls_count, \
+                          dbf.lookup_by_any_name([t.token for t in name_fulls],
+                                                 addr_partials.get_tokens(),
+                                                 [])
+                    penalty += 0.2
+            penalty += 0.4
 
-        addr_count = min(t.addr_count for t in addr_partials) if addr_partials else 30000
-        # Partial term to frequent. Try looking up by rare full names first.
-        name_fulls = self.query.get_tokens(name, TokenType.WORD)
-        if name_fulls:
-            fulls_count = sum(t.count for t in name_fulls)
+            name_split = name_partials.get_num_lookup_tokens(50000, 10)
+            addr_split = addr_partials.get_num_lookup_tokens(30000, 5)
 
-            if fulls_count < 50000 or addr_count < 30000:
-                yield penalty, fulls_count / (2**len(addr_tokens)), \
-                    self.get_full_name_ranking(name_fulls, addr_partials,
-                                               fulls_count > 30000 / max(1, len(addr_tokens)))
+        if name_split > 0 \
+           and (addr_split < 0 or name_partials.min_count() <= addr_partials.min_count()):
+            # lookup by name
+            lookup = name_partials.split_lookup(name_split, 'name_vector')
+            lookup.append(dbf.FieldLookup('nameaddress_vector',
+                                          addr_partials.get_tokens(), lookups.Restrict))
+            yield penalty, name_partials.expected_for_all_search(5), lookup
+        elif addr_split > 0:
+            # lookup by address
+            lookup = addr_partials.split_lookup(addr_split, 'nameaddress_vector')
+            lookup.append(dbf.FieldLookup('name_vector',
+                                          name_partials.get_tokens(), lookups.Restrict))
+            yield penalty, addr_partials.expected_for_all_search(3), lookup
+        elif len(name_partials) > 1:
+            penalty += 0.5
+            # To catch remaining results, lookup by name and address
+            # We only do this if there is a reasonable number of results expected.
+            exp_count = min(name_partials.min_count(), addr_partials.min_count())
+            exp_count = int(exp_count / (min(3, len(name_partials)) + min(3, len(addr_partials))))
+            if exp_count < 50000:
+                lookup = name_partials.split_lookup(3, 'name_vector')
+                lookup.extend(addr_partials.split_lookup(3, 'nameaddress_vector'))
 
-        # To catch remaining results, lookup by name and address
-        # We only do this if there is a reasonable number of results expected.
-        exp_count = exp_count / (2**len(addr_tokens)) if addr_tokens else exp_count
-        if exp_count < 10000 and addr_count < 20000:
-            penalty += 0.35 * max(1 if name_fulls else 0.1,
-                                  5 - len(name_partials) - len(addr_tokens))
-            yield penalty, exp_count, \
-                self.get_name_address_ranking(list(name_partials.keys()), addr_partials)
+                yield penalty, exp_count, lookup
 
     def get_name_address_ranking(self, name_tokens: List[int],
-                                 addr_partials: List[Token]) -> List[dbf.FieldLookup]:
+                                 addr_partials: List[qmod.Token]) -> List[dbf.FieldLookup]:
         """ Create a ranking expression looking up by name and address.
         """
         lookup = [dbf.FieldLookup('name_vector', name_tokens, lookups.LookupAll)]
@@ -258,23 +318,16 @@ class SearchBuilder:
 
         return lookup
 
-    def get_full_name_ranking(self, name_fulls: List[Token], addr_partials: List[Token],
+    def get_full_name_ranking(self, name_fulls: List[qmod.Token], addr_partials: List[qmod.Token],
                               use_lookup: bool) -> List[dbf.FieldLookup]:
         """ Create a ranking expression with full name terms and
             additional address lookup. When 'use_lookup' is true, then
             address lookups will use the index, when the occurrences are not
             too many.
         """
-        # At this point drop unindexed partials from the address.
-        # This might yield wrong results, nothing we can do about that.
         if use_lookup:
             addr_restrict_tokens = []
-            addr_lookup_tokens = []
-            for t in addr_partials:
-                if t.addr_count > 20000:
-                    addr_restrict_tokens.append(t.token)
-                else:
-                    addr_lookup_tokens.append(t.token)
+            addr_lookup_tokens = [t.token for t in addr_partials]
         else:
             addr_restrict_tokens = [t.token for t in addr_partials]
             addr_lookup_tokens = []
@@ -282,19 +335,22 @@ class SearchBuilder:
         return dbf.lookup_by_any_name([t.token for t in name_fulls],
                                       addr_restrict_tokens, addr_lookup_tokens)
 
-    def get_name_ranking(self, trange: TokenRange,
+    def get_name_ranking(self, trange: qmod.TokenRange,
                          db_field: str = 'name_vector') -> dbf.FieldRanking:
         """ Create a ranking expression for a name term in the given range.
         """
-        name_fulls = self.query.get_tokens(trange, TokenType.WORD)
-        ranks = [dbf.RankedTokens(t.penalty, [t.token]) for t in name_fulls]
+        name_fulls = self.query.get_tokens(trange, qmod.TOKEN_WORD)
+        full_word_penalty = self.query.get_in_word_penalty(trange)
+        ranks = [dbf.RankedTokens(t.penalty + full_word_penalty, [t.token])
+                 for t in name_fulls]
         ranks.sort(key=lambda r: r.penalty)
         # Fallback, sum of penalty for partials
-        name_partials = self.query.get_partials_list(trange)
-        default = sum(t.penalty for t in name_partials) + 0.2
+        default = sum(t.penalty for t in self.query.iter_partials(trange)) + 0.2
+        default += sum(n.word_break_penalty
+                       for n in self.query.nodes[trange.start + 1:trange.end])
         return dbf.FieldRanking(db_field, default, ranks)
 
-    def get_addr_ranking(self, trange: TokenRange) -> dbf.FieldRanking:
+    def get_addr_ranking(self, trange: qmod.TokenRange) -> dbf.FieldRanking:
         """ Create a list of ranking expressions for an address term
             for the given ranges.
         """
@@ -303,36 +359,40 @@ class SearchBuilder:
         ranks: List[dbf.RankedTokens] = []
 
         while todo:
-            neglen, pos, rank = heapq.heappop(todo)
+            _, pos, rank = heapq.heappop(todo)
+            # partial node
+            partial = self.query.nodes[pos].partial
+            if partial is not None:
+                if pos + 1 < trange.end:
+                    penalty = rank.penalty + partial.penalty \
+                              + self.query.nodes[pos + 1].word_break_penalty
+                    heapq.heappush(todo, (-(pos + 1), pos + 1,
+                                   dbf.RankedTokens(penalty, rank.tokens)))
+                else:
+                    ranks.append(dbf.RankedTokens(rank.penalty + partial.penalty,
+                                                  rank.tokens))
+            # full words
             for tlist in self.query.nodes[pos].starting:
-                if tlist.ttype in (TokenType.PARTIAL, TokenType.WORD):
+                if tlist.ttype == qmod.TOKEN_WORD:
                     if tlist.end < trange.end:
-                        chgpenalty = PENALTY_WORDCHANGE[self.query.nodes[tlist.end].btype]
-                        if tlist.ttype == TokenType.PARTIAL:
-                            penalty = rank.penalty + chgpenalty \
-                                      + max(t.penalty for t in tlist.tokens)
-                            heapq.heappush(todo, (neglen - 1, tlist.end,
-                                                  dbf.RankedTokens(penalty, rank.tokens)))
-                        else:
-                            for t in tlist.tokens:
-                                heapq.heappush(todo, (neglen - 1, tlist.end,
-                                                      rank.with_token(t, chgpenalty)))
+                        chgpenalty = self.query.nodes[tlist.end].word_break_penalty \
+                                     + self.query.get_in_word_penalty(
+                                            qmod.TokenRange(pos, tlist.end))
+                        for t in tlist.tokens:
+                            heapq.heappush(todo, (-tlist.end, tlist.end,
+                                                  rank.with_token(t, chgpenalty)))
                     elif tlist.end == trange.end:
-                        if tlist.ttype == TokenType.PARTIAL:
-                            ranks.append(dbf.RankedTokens(rank.penalty
-                                                          + max(t.penalty for t in tlist.tokens),
-                                                          rank.tokens))
-                        else:
-                            ranks.extend(rank.with_token(t, 0.0) for t in tlist.tokens)
-                        if len(ranks) >= 10:
-                            # Too many variants, bail out and only add
-                            # Worst-case Fallback: sum of penalty of partials
-                            name_partials = self.query.get_partials_list(trange)
-                            default = sum(t.penalty for t in name_partials) + 0.2
-                            ranks.append(dbf.RankedTokens(rank.penalty + default, []))
-                            # Bail out of outer loop
-                            todo.clear()
-                            break
+                        ranks.extend(rank.with_token(t, 0.0) for t in tlist.tokens)
+
+            if len(ranks) >= 10:
+                # Too many variants, bail out and only add
+                # Worst-case Fallback: sum of penalty of partials
+                default = sum(t.penalty for t in self.query.iter_partials(trange)) + 0.2
+                default += sum(n.word_break_penalty
+                               for n in self.query.nodes[trange.start + 1:trange.end])
+                ranks.append(dbf.RankedTokens(rank.penalty + default, []))
+                # Bail out of outer loop
+                break
 
         ranks.sort(key=lambda r: len(r.tokens))
         default = ranks[0].penalty + 0.3
@@ -352,58 +412,54 @@ class SearchBuilder:
             if not tokens:
                 return None
             sdata.set_strings('countries', tokens)
+            sdata.penalty += self.query.get_in_word_penalty(assignment.country)
         elif self.details.countries:
             sdata.countries = dbf.WeightedStrings(self.details.countries,
                                                   [0.0] * len(self.details.countries))
         if assignment.housenumber:
             sdata.set_strings('housenumbers',
                               self.query.get_tokens(assignment.housenumber,
-                                                    TokenType.HOUSENUMBER))
+                                                    qmod.TOKEN_HOUSENUMBER))
+            sdata.penalty += self.query.get_in_word_penalty(assignment.housenumber)
         if assignment.postcode:
             sdata.set_strings('postcodes',
                               self.query.get_tokens(assignment.postcode,
-                                                    TokenType.POSTCODE))
+                                                    qmod.TOKEN_POSTCODE))
+            sdata.penalty += self.query.get_in_word_penalty(assignment.postcode)
         if assignment.qualifier:
             tokens = self.get_qualifier_tokens(assignment.qualifier)
             if not tokens:
                 return None
             sdata.set_qualifiers(tokens)
+            sdata.penalty += self.query.get_in_word_penalty(assignment.qualifier)
         elif self.details.categories:
             sdata.qualifiers = dbf.WeightedCategories(self.details.categories,
                                                       [0.0] * len(self.details.categories))
 
         if assignment.address:
-            if not assignment.name and assignment.housenumber:
-                # housenumber search: the first item needs to be handled like
-                # a name in ranking or penalties are not comparable with
-                # normal searches.
-                sdata.set_ranking([self.get_name_ranking(assignment.address[0],
-                                                         db_field='nameaddress_vector')]
-                                  + [self.get_addr_ranking(r) for r in assignment.address[1:]])
-            else:
-                sdata.set_ranking([self.get_addr_ranking(r) for r in assignment.address])
+            sdata.set_ranking([self.get_addr_ranking(r) for r in assignment.address])
         else:
             sdata.rankings = []
 
         return sdata
 
-    def get_country_tokens(self, trange: TokenRange) -> List[Token]:
+    def get_country_tokens(self, trange: qmod.TokenRange) -> List[qmod.Token]:
         """ Return the list of country tokens for the given range,
             optionally filtered by the country list from the details
             parameters.
         """
-        tokens = self.query.get_tokens(trange, TokenType.COUNTRY)
+        tokens = self.query.get_tokens(trange, qmod.TOKEN_COUNTRY)
         if self.details.countries:
             tokens = [t for t in tokens if t.lookup_word in self.details.countries]
 
         return tokens
 
-    def get_qualifier_tokens(self, trange: TokenRange) -> List[Token]:
+    def get_qualifier_tokens(self, trange: qmod.TokenRange) -> List[qmod.Token]:
         """ Return the list of qualifier tokens for the given range,
             optionally filtered by the qualifier list from the details
             parameters.
         """
-        tokens = self.query.get_tokens(trange, TokenType.QUALIFIER)
+        tokens = self.query.get_tokens(trange, qmod.TOKEN_QUALIFIER)
         if self.details.categories:
             tokens = [t for t in tokens if t.get_category() in self.details.categories]
 
@@ -416,7 +472,7 @@ class SearchBuilder:
         """
         if assignment.near_item:
             tokens: Dict[Tuple[str, str], float] = {}
-            for t in self.query.get_tokens(assignment.near_item, TokenType.NEAR_ITEM):
+            for t in self.query.get_tokens(assignment.near_item, qmod.TOKEN_NEAR_ITEM):
                 cat = t.get_category()
                 # The category of a near search will be that of near_item.
                 # Thus, if search is restricted to a category parameter,
@@ -427,14 +483,3 @@ class SearchBuilder:
             return dbf.WeightedCategories(list(tokens.keys()), list(tokens.values()))
 
         return None
-
-
-PENALTY_WORDCHANGE = {
-    BreakType.START: 0.0,
-    BreakType.END: 0.0,
-    BreakType.PHRASE: 0.0,
-    BreakType.SOFT_PHRASE: 0.0,
-    BreakType.WORD: 0.1,
-    BreakType.PART: 0.2,
-    BreakType.TOKEN: 0.4
-}

src/nominatim_api/search/db_search_fields.py

@@ -7,7 +7,7 @@
 """
 Data structures for more complex fields in abstract search descriptions.
 """
-from typing import List, Tuple, Iterator, Dict, Type
+from typing import List, Tuple, Iterator, Dict, Type, cast
 import dataclasses
 
 import sqlalchemy as sa
@@ -18,6 +18,66 @@ from .query import Token
 from . import db_search_lookups as lookups
 
 
+class CountedTokenIDs:
+    """ A list of token IDs with their respective counts, sorted
+        from least frequent to most frequent.
+
+        If a token count is one, then statistics are likely to be unavaible
+        and a relatively high count is assumed instead.
+    """
+
+    def __init__(self, tokens: Iterator[Token], count_column: str = 'count'):
+        self.tokens = list({(cast(int, getattr(t, count_column)), t.token) for t in tokens})
+        self.tokens.sort(key=lambda t: t[0] if t[0] > 1 else 100000)
+
+    def __len__(self) -> int:
+        return len(self.tokens)
+
+    def get_num_lookup_tokens(self, limit: int, fac: int) -> int:
+        """ Suggest the number of tokens to be used for an index lookup.
+            The idea here is to use as few items as possible while making
+            sure the number of rows returned stays below 'limit' which
+            makes recheck of the returned rows more expensive than adding
+            another item for the index lookup. 'fac' is the factor by which
+            the limit is increased every time a lookup item is added.
+
+            If the list of tokens doesn't seem suitable at all for index
+            lookup, -1 is returned.
+        """
+        length = len(self.tokens)
+        min_count = self.tokens[0][0]
+        if min_count == 1:
+            return min(length, 3)  # no statistics available, use index
+
+        for i in range(min(length, 3)):
+            if min_count < limit:
+                return i + 1
+            limit = limit * fac
+
+        return -1
+
+    def min_count(self) -> int:
+        return self.tokens[0][0]
+
+    def expected_for_all_search(self, fac: int = 5) -> int:
+        return int(self.tokens[0][0] / (fac**(len(self.tokens) - 1)))
+
+    def get_tokens(self) -> List[int]:
+        return [t[1] for t in self.tokens]
+
+    def get_head_tokens(self, num_tokens: int) -> List[int]:
+        return [t[1] for t in self.tokens[:num_tokens]]
+
+    def get_tail_tokens(self, first: int) -> List[int]:
+        return [t[1] for t in self.tokens[first:]]
+
+    def split_lookup(self, split: int, column: str) -> 'List[FieldLookup]':
+        lookup = [FieldLookup(column, self.get_head_tokens(split), lookups.LookupAll)]
+        if split < len(self.tokens):
+            lookup.append(FieldLookup(column, self.get_tail_tokens(split), lookups.Restrict))
+        return lookup
+
+
 @dataclasses.dataclass
 class WeightedStrings:
     """ A list of strings together with a penalty.

src/nominatim_api/search/db_searches.py

@@ -1,863 +0,0 @@
-# SPDX-License-Identifier: GPL-3.0-or-later
-#
-# This file is part of Nominatim. (https://nominatim.org)
-#
-# Copyright (C) 2024 by the Nominatim developer community.
-# For a full list of authors see the git log.
-"""
-Implementation of the actual database accesses for forward search.
-"""
-from typing import List, Tuple, AsyncIterator, Dict, Any, Callable, cast
-import abc
-
-import sqlalchemy as sa
-
-from ..typing import SaFromClause, SaScalarSelect, SaColumn, \
-                     SaExpression, SaSelect, SaLambdaSelect, SaRow, SaBind
-from ..sql.sqlalchemy_types import Geometry, IntArray
-from ..connection import SearchConnection
-from ..types import SearchDetails, DataLayer, GeometryFormat, Bbox
-from .. import results as nres
-from .db_search_fields import SearchData, WeightedCategories
-
-
-def no_index(expr: SaColumn) -> SaColumn:
-    """ Wrap the given expression, so that the query planner will
-        refrain from using the expression for index lookup.
-    """
-    return sa.func.coalesce(sa.null(), expr)
-
-
-def _details_to_bind_params(details: SearchDetails) -> Dict[str, Any]:
-    """ Create a dictionary from search parameters that can be used
-        as bind parameter for SQL execute.
-    """
-    return {'limit': details.max_results,
-            'min_rank': details.min_rank,
-            'max_rank': details.max_rank,
-            'viewbox': details.viewbox,
-            'viewbox2': details.viewbox_x2,
-            'near': details.near,
-            'near_radius': details.near_radius,
-            'excluded': details.excluded,
-            'countries': details.countries}
-
-
-LIMIT_PARAM: SaBind = sa.bindparam('limit')
-MIN_RANK_PARAM: SaBind = sa.bindparam('min_rank')
-MAX_RANK_PARAM: SaBind = sa.bindparam('max_rank')
-VIEWBOX_PARAM: SaBind = sa.bindparam('viewbox', type_=Geometry)
-VIEWBOX2_PARAM: SaBind = sa.bindparam('viewbox2', type_=Geometry)
-NEAR_PARAM: SaBind = sa.bindparam('near', type_=Geometry)
-NEAR_RADIUS_PARAM: SaBind = sa.bindparam('near_radius')
-COUNTRIES_PARAM: SaBind = sa.bindparam('countries')
-
-
-def filter_by_area(sql: SaSelect, t: SaFromClause,
-                   details: SearchDetails, avoid_index: bool = False) -> SaSelect:
-    """ Apply SQL statements for filtering by viewbox and near point,
-        if applicable.
-    """
-    if details.near is not None and details.near_radius is not None:
-        if details.near_radius < 0.1 and not avoid_index:
-            sql = sql.where(t.c.geometry.within_distance(NEAR_PARAM, NEAR_RADIUS_PARAM))
-        else:
-            sql = sql.where(t.c.geometry.ST_Distance(NEAR_PARAM) <= NEAR_RADIUS_PARAM)
-    if details.viewbox is not None and details.bounded_viewbox:
-        sql = sql.where(t.c.geometry.intersects(VIEWBOX_PARAM,
-                                                use_index=not avoid_index and
-                                                details.viewbox.area < 0.2))
-
-    return sql
-
-
-def _exclude_places(t: SaFromClause) -> Callable[[], SaExpression]:
-    return lambda: t.c.place_id.not_in(sa.bindparam('excluded'))
-
-
-def _select_placex(t: SaFromClause) -> SaSelect:
-    return sa.select(t.c.place_id, t.c.osm_type, t.c.osm_id, t.c.name,
-                     t.c.class_, t.c.type,
-                     t.c.address, t.c.extratags,
-                     t.c.housenumber, t.c.postcode, t.c.country_code,
-                     t.c.wikipedia,
-                     t.c.parent_place_id, t.c.rank_address, t.c.rank_search,
-                     t.c.linked_place_id, t.c.admin_level,
-                     t.c.centroid,
-                     t.c.geometry.ST_Expand(0).label('bbox'))
-
-
-def _add_geometry_columns(sql: SaLambdaSelect, col: SaColumn, details: SearchDetails) -> SaSelect:
-    out = []
-
-    if details.geometry_simplification > 0.0:
-        col = sa.func.ST_SimplifyPreserveTopology(col, details.geometry_simplification)
-
-    if details.geometry_output & GeometryFormat.GEOJSON:
-        out.append(sa.func.ST_AsGeoJSON(col, 7).label('geometry_geojson'))
-    if details.geometry_output & GeometryFormat.TEXT:
-        out.append(sa.func.ST_AsText(col).label('geometry_text'))
-    if details.geometry_output & GeometryFormat.KML:
-        out.append(sa.func.ST_AsKML(col, 7).label('geometry_kml'))
-    if details.geometry_output & GeometryFormat.SVG:
-        out.append(sa.func.ST_AsSVG(col, 0, 7).label('geometry_svg'))
-
-    return sql.add_columns(*out)
-
-
-def _make_interpolation_subquery(table: SaFromClause, inner: SaFromClause,
-                                 numerals: List[int], details: SearchDetails) -> SaScalarSelect:
-    all_ids = sa.func.ArrayAgg(table.c.place_id)
-    sql = sa.select(all_ids).where(table.c.parent_place_id == inner.c.place_id)
-
-    if len(numerals) == 1:
-        sql = sql.where(sa.between(numerals[0], table.c.startnumber, table.c.endnumber))\
-                 .where((numerals[0] - table.c.startnumber) % table.c.step == 0)
-    else:
-        sql = sql.where(sa.or_(
-                *(sa.and_(sa.between(n, table.c.startnumber, table.c.endnumber),
-                          (n - table.c.startnumber) % table.c.step == 0)
-                  for n in numerals)))
-
-    if details.excluded:
-        sql = sql.where(_exclude_places(table))
-
-    return sql.scalar_subquery()
-
-
-def _filter_by_layer(table: SaFromClause, layers: DataLayer) -> SaColumn:
-    orexpr: List[SaExpression] = []
-    if layers & DataLayer.ADDRESS and layers & DataLayer.POI:
-        orexpr.append(no_index(table.c.rank_address).between(1, 30))
-    elif layers & DataLayer.ADDRESS:
-        orexpr.append(no_index(table.c.rank_address).between(1, 29))
-        orexpr.append(sa.func.IsAddressPoint(table))
-    elif layers & DataLayer.POI:
-        orexpr.append(sa.and_(no_index(table.c.rank_address) == 30,
-                              table.c.class_.not_in(('place', 'building'))))
-
-    if layers & DataLayer.MANMADE:
-        exclude = []
-        if not layers & DataLayer.RAILWAY:
-            exclude.append('railway')
-        if not layers & DataLayer.NATURAL:
-            exclude.extend(('natural', 'water', 'waterway'))
-        orexpr.append(sa.and_(table.c.class_.not_in(tuple(exclude)),
-                              no_index(table.c.rank_address) == 0))
-    else:
-        include = []
-        if layers & DataLayer.RAILWAY:
-            include.append('railway')
-        if layers & DataLayer.NATURAL:
-            include.extend(('natural', 'water', 'waterway'))
-        orexpr.append(sa.and_(table.c.class_.in_(tuple(include)),
-                              no_index(table.c.rank_address) == 0))
-
-    if len(orexpr) == 1:
-        return orexpr[0]
-
-    return sa.or_(*orexpr)
-
-
-def _interpolated_position(table: SaFromClause, nr: SaColumn) -> SaColumn:
-    pos = sa.cast(nr - table.c.startnumber, sa.Float) / (table.c.endnumber - table.c.startnumber)
-    return sa.case(
-            (table.c.endnumber == table.c.startnumber, table.c.linegeo.ST_Centroid()),
-            else_=table.c.linegeo.ST_LineInterpolatePoint(pos)).label('centroid')
-
-
-async def _get_placex_housenumbers(conn: SearchConnection,
-                                   place_ids: List[int],
-                                   details: SearchDetails) -> AsyncIterator[nres.SearchResult]:
-    t = conn.t.placex
-    sql = _select_placex(t).add_columns(t.c.importance)\
-                           .where(t.c.place_id.in_(place_ids))
-
-    if details.geometry_output:
-        sql = _add_geometry_columns(sql, t.c.geometry, details)
-
-    for row in await conn.execute(sql):
-        result = nres.create_from_placex_row(row, nres.SearchResult)
-        assert result
-        result.bbox = Bbox.from_wkb(row.bbox)
-        yield result
-
-
-def _int_list_to_subquery(inp: List[int]) -> 'sa.Subquery':
-    """ Create a subselect that returns the given list of integers
-        as rows in the column 'nr'.
-    """
-    vtab = sa.func.JsonArrayEach(sa.type_coerce(inp, sa.JSON))\
-             .table_valued(sa.column('value', type_=sa.JSON))
-    return sa.select(sa.cast(sa.cast(vtab.c.value, sa.Text), sa.Integer).label('nr')).subquery()
-
-
-async def _get_osmline(conn: SearchConnection, place_ids: List[int],
-                       numerals: List[int],
-                       details: SearchDetails) -> AsyncIterator[nres.SearchResult]:
-    t = conn.t.osmline
-
-    values = _int_list_to_subquery(numerals)
-    sql = sa.select(t.c.place_id, t.c.osm_id,
-                    t.c.parent_place_id, t.c.address,
-                    values.c.nr.label('housenumber'),
-                    _interpolated_position(t, values.c.nr),
-                    t.c.postcode, t.c.country_code)\
-            .where(t.c.place_id.in_(place_ids))\
-            .join(values, values.c.nr.between(t.c.startnumber, t.c.endnumber))
-
-    if details.geometry_output:
-        sub = sql.subquery()
-        sql = _add_geometry_columns(sa.select(sub), sub.c.centroid, details)
-
-    for row in await conn.execute(sql):
-        result = nres.create_from_osmline_row(row, nres.SearchResult)
-        assert result
-        yield result
-
-
-async def _get_tiger(conn: SearchConnection, place_ids: List[int],
-                     numerals: List[int], osm_id: int,
-                     details: SearchDetails) -> AsyncIterator[nres.SearchResult]:
-    t = conn.t.tiger
-    values = _int_list_to_subquery(numerals)
-    sql = sa.select(t.c.place_id, t.c.parent_place_id,
-                    sa.literal('W').label('osm_type'),
-                    sa.literal(osm_id).label('osm_id'),
-                    values.c.nr.label('housenumber'),
-                    _interpolated_position(t, values.c.nr),
-                    t.c.postcode)\
-            .where(t.c.place_id.in_(place_ids))\
-            .join(values, values.c.nr.between(t.c.startnumber, t.c.endnumber))
-
-    if details.geometry_output:
-        sub = sql.subquery()
-        sql = _add_geometry_columns(sa.select(sub), sub.c.centroid, details)
-
-    for row in await conn.execute(sql):
-        result = nres.create_from_tiger_row(row, nres.SearchResult)
-        assert result
-        yield result
-
-
-class AbstractSearch(abc.ABC):
-    """ Encapuslation of a single lookup in the database.
-    """
-    SEARCH_PRIO: int = 2
-
-    def __init__(self, penalty: float) -> None:
-        self.penalty = penalty
-
-    @abc.abstractmethod
-    async def lookup(self, conn: SearchConnection,
-                     details: SearchDetails) -> nres.SearchResults:
-        """ Find results for the search in the database.
-        """
-
-
-class NearSearch(AbstractSearch):
-    """ Category search of a place type near the result of another search.
-    """
-    def __init__(self, penalty: float, categories: WeightedCategories,
-                 search: AbstractSearch) -> None:
-        super().__init__(penalty)
-        self.search = search
-        self.categories = categories
-
-    async def lookup(self, conn: SearchConnection,
-                     details: SearchDetails) -> nres.SearchResults:
-        """ Find results for the search in the database.
-        """
-        results = nres.SearchResults()
-        base = await self.search.lookup(conn, details)
-
-        if not base:
-            return results
-
-        base.sort(key=lambda r: (r.accuracy, r.rank_search))
-        max_accuracy = base[0].accuracy + 0.5
-        if base[0].rank_address == 0:
-            min_rank = 0
-            max_rank = 0
-        elif base[0].rank_address < 26:
-            min_rank = 1
-            max_rank = min(25, base[0].rank_address + 4)
-        else:
-            min_rank = 26
-            max_rank = 30
-        base = nres.SearchResults(r for r in base
-                                  if (r.source_table == nres.SourceTable.PLACEX
-                                      and r.accuracy <= max_accuracy
-                                      and r.bbox and r.bbox.area < 20
-                                      and r.rank_address >= min_rank
-                                      and r.rank_address <= max_rank))
-
-        if base:
-            baseids = [b.place_id for b in base[:5] if b.place_id]
-
-            for category, penalty in self.categories:
-                await self.lookup_category(results, conn, baseids, category, penalty, details)
-                if len(results) >= details.max_results:
-                    break
-
-        return results
-
-    async def lookup_category(self, results: nres.SearchResults,
-                              conn: SearchConnection, ids: List[int],
-                              category: Tuple[str, str], penalty: float,
-                              details: SearchDetails) -> None:
-        """ Find places of the given category near the list of
-            place ids and add the results to 'results'.
-        """
-        table = await conn.get_class_table(*category)
-
-        tgeom = conn.t.placex.alias('pgeom')
-
-        if table is None:
-            # No classtype table available, do a simplified lookup in placex.
-            table = conn.t.placex
-            sql = sa.select(table.c.place_id,
-                            sa.func.min(tgeom.c.centroid.ST_Distance(table.c.centroid))
-                              .label('dist'))\
-                    .join(tgeom, table.c.geometry.intersects(tgeom.c.centroid.ST_Expand(0.01)))\
-                    .where(table.c.class_ == category[0])\
-                    .where(table.c.type == category[1])
-        else:
-            # Use classtype table. We can afford to use a larger
-            # radius for the lookup.
-            sql = sa.select(table.c.place_id,
-                            sa.func.min(tgeom.c.centroid.ST_Distance(table.c.centroid))
-                              .label('dist'))\
-                    .join(tgeom,
-                          table.c.centroid.ST_CoveredBy(
-                              sa.case((sa.and_(tgeom.c.rank_address > 9,
-                                               tgeom.c.geometry.is_area()),
-                                       tgeom.c.geometry),
-                                      else_=tgeom.c.centroid.ST_Expand(0.05))))
-
-        inner = sql.where(tgeom.c.place_id.in_(ids))\
-                   .group_by(table.c.place_id).subquery()
-
-        t = conn.t.placex
-        sql = _select_placex(t).add_columns((-inner.c.dist).label('importance'))\
-                               .join(inner, inner.c.place_id == t.c.place_id)\
-                               .order_by(inner.c.dist)
-
-        sql = sql.where(no_index(t.c.rank_address).between(MIN_RANK_PARAM, MAX_RANK_PARAM))
-        if details.countries:
-            sql = sql.where(t.c.country_code.in_(COUNTRIES_PARAM))
-        if details.excluded:
-            sql = sql.where(_exclude_places(t))
-        if details.layers is not None:
-            sql = sql.where(_filter_by_layer(t, details.layers))
-
-        sql = sql.limit(LIMIT_PARAM)
-        for row in await conn.execute(sql, _details_to_bind_params(details)):
-            result = nres.create_from_placex_row(row, nres.SearchResult)
-            assert result
-            result.accuracy = self.penalty + penalty
-            result.bbox = Bbox.from_wkb(row.bbox)
-            results.append(result)
-
-
-class PoiSearch(AbstractSearch):
-    """ Category search in a geographic area.
-    """
-    def __init__(self, sdata: SearchData) -> None:
-        super().__init__(sdata.penalty)
-        self.qualifiers = sdata.qualifiers
-        self.countries = sdata.countries
-
-    async def lookup(self, conn: SearchConnection,
-                     details: SearchDetails) -> nres.SearchResults:
-        """ Find results for the search in the database.
-        """
-        bind_params = _details_to_bind_params(details)
-        t = conn.t.placex
-
-        rows: List[SaRow] = []
-
-        if details.near and details.near_radius is not None and details.near_radius < 0.2:
-            # simply search in placex table
-            def _base_query() -> SaSelect:
-                return _select_placex(t) \
-                           .add_columns((-t.c.centroid.ST_Distance(NEAR_PARAM))
-                                        .label('importance'))\
-                           .where(t.c.linked_place_id == None) \
-                           .where(t.c.geometry.within_distance(NEAR_PARAM, NEAR_RADIUS_PARAM)) \
-                           .order_by(t.c.centroid.ST_Distance(NEAR_PARAM)) \
-                           .limit(LIMIT_PARAM)
-
-            classtype = self.qualifiers.values
-            if len(classtype) == 1:
-                cclass, ctype = classtype[0]
-                sql: SaLambdaSelect = sa.lambda_stmt(
-                    lambda: _base_query().where(t.c.class_ == cclass)
-                                         .where(t.c.type == ctype))
-            else:
-                sql = _base_query().where(sa.or_(*(sa.and_(t.c.class_ == cls, t.c.type == typ)
-                                                   for cls, typ in classtype)))
-
-            if self.countries:
-                sql = sql.where(t.c.country_code.in_(self.countries.values))
-
-            if details.viewbox is not None and details.bounded_viewbox:
-                sql = sql.where(t.c.geometry.intersects(VIEWBOX_PARAM))
-
-            rows.extend(await conn.execute(sql, bind_params))
-        else:
-            # use the class type tables
-            for category in self.qualifiers.values:
-                table = await conn.get_class_table(*category)
-                if table is not None:
-                    sql = _select_placex(t)\
-                               .add_columns(t.c.importance)\
-                               .join(table, t.c.place_id == table.c.place_id)\
-                               .where(t.c.class_ == category[0])\
-                               .where(t.c.type == category[1])
-
-                    if details.viewbox is not None and details.bounded_viewbox:
-                        sql = sql.where(table.c.centroid.intersects(VIEWBOX_PARAM))
-
-                    if details.near and details.near_radius is not None:
-                        sql = sql.order_by(table.c.centroid.ST_Distance(NEAR_PARAM))\
-                                 .where(table.c.centroid.within_distance(NEAR_PARAM,
-                                                                         NEAR_RADIUS_PARAM))
-
-                    if self.countries:
-                        sql = sql.where(t.c.country_code.in_(self.countries.values))
-
-                    sql = sql.limit(LIMIT_PARAM)
-                    rows.extend(await conn.execute(sql, bind_params))
-
-        results = nres.SearchResults()
-        for row in rows:
-            result = nres.create_from_placex_row(row, nres.SearchResult)
-            assert result
-            result.accuracy = self.penalty + self.qualifiers.get_penalty((row.class_, row.type))
-            result.bbox = Bbox.from_wkb(row.bbox)
-            results.append(result)
-
-        return results
-
-
-class CountrySearch(AbstractSearch):
-    """ Search for a country name or country code.
-    """
-    SEARCH_PRIO = 0
-
-    def __init__(self, sdata: SearchData) -> None:
-        super().__init__(sdata.penalty)
-        self.countries = sdata.countries
-
-    async def lookup(self, conn: SearchConnection,
-                     details: SearchDetails) -> nres.SearchResults:
-        """ Find results for the search in the database.
-        """
-        t = conn.t.placex
-
-        ccodes = self.countries.values
-        sql = _select_placex(t)\
-            .add_columns(t.c.importance)\
-            .where(t.c.country_code.in_(ccodes))\
-            .where(t.c.rank_address == 4)
-
-        if details.geometry_output:
-            sql = _add_geometry_columns(sql, t.c.geometry, details)
-
-        if details.excluded:
-            sql = sql.where(_exclude_places(t))
-
-        sql = filter_by_area(sql, t, details)
-
-        results = nres.SearchResults()
-        for row in await conn.execute(sql, _details_to_bind_params(details)):
-            result = nres.create_from_placex_row(row, nres.SearchResult)
-            assert result
-            result.accuracy = self.penalty + self.countries.get_penalty(row.country_code, 5.0)
-            result.bbox = Bbox.from_wkb(row.bbox)
-            results.append(result)
-
-        if not results:
-            results = await self.lookup_in_country_table(conn, details)
-
-        if results:
-            details.min_rank = min(5, details.max_rank)
-            details.max_rank = min(25, details.max_rank)
-
-        return results
-
-    async def lookup_in_country_table(self, conn: SearchConnection,
-                                      details: SearchDetails) -> nres.SearchResults:
-        """ Look up the country in the fallback country tables.
-        """
-        # Avoid the fallback search when this is a more search. Country results
-        # usually are in the first batch of results and it is not possible
-        # to exclude these fallbacks.
-        if details.excluded:
-            return nres.SearchResults()
-
-        t = conn.t.country_name
-        tgrid = conn.t.country_grid
-
-        sql = sa.select(tgrid.c.country_code,
-                        tgrid.c.geometry.ST_Centroid().ST_Collect().ST_Centroid()
-                             .label('centroid'),
-                        tgrid.c.geometry.ST_Collect().ST_Expand(0).label('bbox'))\
-                .where(tgrid.c.country_code.in_(self.countries.values))\
-                .group_by(tgrid.c.country_code)
-
-        sql = filter_by_area(sql, tgrid, details, avoid_index=True)
-
-        sub = sql.subquery('grid')
-
-        sql = sa.select(t.c.country_code,
-                        t.c.name.merge(t.c.derived_name).label('name'),
-                        sub.c.centroid, sub.c.bbox)\
-                .join(sub, t.c.country_code == sub.c.country_code)
-
-        if details.geometry_output:
-            sql = _add_geometry_columns(sql, sub.c.centroid, details)
-
-        results = nres.SearchResults()
-        for row in await conn.execute(sql, _details_to_bind_params(details)):
-            result = nres.create_from_country_row(row, nres.SearchResult)
-            assert result
-            result.bbox = Bbox.from_wkb(row.bbox)
-            result.accuracy = self.penalty + self.countries.get_penalty(row.country_code, 5.0)
-            results.append(result)
-
-        return results
-
-
-class PostcodeSearch(AbstractSearch):
-    """ Search for a postcode.
-    """
-    def __init__(self, extra_penalty: float, sdata: SearchData) -> None:
-        super().__init__(sdata.penalty + extra_penalty)
-        self.countries = sdata.countries
-        self.postcodes = sdata.postcodes
-        self.lookups = sdata.lookups
-        self.rankings = sdata.rankings
-
-    async def lookup(self, conn: SearchConnection,
-                     details: SearchDetails) -> nres.SearchResults:
-        """ Find results for the search in the database.
-        """
-        t = conn.t.postcode
-        pcs = self.postcodes.values
-
-        sql = sa.select(t.c.place_id, t.c.parent_place_id,
-                        t.c.rank_search, t.c.rank_address,
-                        t.c.postcode, t.c.country_code,
-                        t.c.geometry.label('centroid'))\
-                .where(t.c.postcode.in_(pcs))
-
-        if details.geometry_output:
-            sql = _add_geometry_columns(sql, t.c.geometry, details)
-
-        penalty: SaExpression = sa.literal(self.penalty)
-
-        if details.viewbox is not None and not details.bounded_viewbox:
-            penalty += sa.case((t.c.geometry.intersects(VIEWBOX_PARAM), 0.0),
-                               (t.c.geometry.intersects(VIEWBOX2_PARAM), 0.5),
-                               else_=1.0)
-
-        if details.near is not None:
-            sql = sql.order_by(t.c.geometry.ST_Distance(NEAR_PARAM))
-
-        sql = filter_by_area(sql, t, details)
-
-        if self.countries:
-            sql = sql.where(t.c.country_code.in_(self.countries.values))
-
-        if details.excluded:
-            sql = sql.where(_exclude_places(t))
-
-        if self.lookups:
-            assert len(self.lookups) == 1
-            tsearch = conn.t.search_name
-            sql = sql.where(tsearch.c.place_id == t.c.parent_place_id)\
-                     .where((tsearch.c.name_vector + tsearch.c.nameaddress_vector)
-                            .contains(sa.type_coerce(self.lookups[0].tokens,
-                                                     IntArray)))
-
-        for ranking in self.rankings:
-            penalty += ranking.sql_penalty(conn.t.search_name)
-        penalty += sa.case(*((t.c.postcode == v, p) for v, p in self.postcodes),
-                           else_=1.0)
-
-        sql = sql.add_columns(penalty.label('accuracy'))
-        sql = sql.order_by('accuracy').limit(LIMIT_PARAM)
-
-        results = nres.SearchResults()
-        for row in await conn.execute(sql, _details_to_bind_params(details)):
-            p = conn.t.placex
-            placex_sql = _select_placex(p)\
-                .add_columns(p.c.importance)\
-                .where(sa.text("""class = 'boundary'
-                                  AND type = 'postal_code'
-                                  AND osm_type = 'R'"""))\
-                .where(p.c.country_code == row.country_code)\
-                .where(p.c.postcode == row.postcode)\
-                .limit(1)
-
-            if details.geometry_output:
-                placex_sql = _add_geometry_columns(placex_sql, p.c.geometry, details)
-
-            for prow in await conn.execute(placex_sql, _details_to_bind_params(details)):
-                result = nres.create_from_placex_row(prow, nres.SearchResult)
-                if result is not None:
-                    result.bbox = Bbox.from_wkb(prow.bbox)
-                break
-            else:
-                result = nres.create_from_postcode_row(row, nres.SearchResult)
-
-            assert result
-            if result.place_id not in details.excluded:
-                result.accuracy = row.accuracy
-                results.append(result)
-
-        return results
-
-
-class PlaceSearch(AbstractSearch):
-    """ Generic search for an address or named place.
-    """
-    SEARCH_PRIO = 1
-
-    def __init__(self, extra_penalty: float, sdata: SearchData, expected_count: int) -> None:
-        super().__init__(sdata.penalty + extra_penalty)
-        self.countries = sdata.countries
-        self.postcodes = sdata.postcodes
-        self.housenumbers = sdata.housenumbers
-        self.qualifiers = sdata.qualifiers
-        self.lookups = sdata.lookups
-        self.rankings = sdata.rankings
-        self.expected_count = expected_count
-
-    def _inner_search_name_cte(self, conn: SearchConnection,
-                               details: SearchDetails) -> 'sa.CTE':
-        """ Create a subquery that preselects the rows in the search_name
-            table.
-        """
-        t = conn.t.search_name
-
-        penalty: SaExpression = sa.literal(self.penalty)
-        for ranking in self.rankings:
-            penalty += ranking.sql_penalty(t)
-
-        sql = sa.select(t.c.place_id, t.c.search_rank, t.c.address_rank,
-                        t.c.country_code, t.c.centroid,
-                        t.c.name_vector, t.c.nameaddress_vector,
-                        sa.case((t.c.importance > 0, t.c.importance),
-                                else_=0.40001-(sa.cast(t.c.search_rank, sa.Float())/75))
-                          .label('importance'),
-                        penalty.label('penalty'))
-
-        for lookup in self.lookups:
-            sql = sql.where(lookup.sql_condition(t))
-
-        if self.countries:
-            sql = sql.where(t.c.country_code.in_(self.countries.values))
-
-        if self.postcodes:
-            # if a postcode is given, don't search for state or country level objects
-            sql = sql.where(t.c.address_rank > 9)
-            if self.expected_count > 10000:
-                # Many results expected. Restrict by postcode.
-                tpc = conn.t.postcode
-                sql = sql.where(sa.select(tpc.c.postcode)
-                                  .where(tpc.c.postcode.in_(self.postcodes.values))
-                                  .where(t.c.centroid.within_distance(tpc.c.geometry, 0.4))
-                                  .exists())
-
-        if details.viewbox is not None:
-            if details.bounded_viewbox:
-                sql = sql.where(t.c.centroid
-                                   .intersects(VIEWBOX_PARAM,
-                                               use_index=details.viewbox.area < 0.2))
-            elif not self.postcodes and not self.housenumbers and self.expected_count >= 10000:
-                sql = sql.where(t.c.centroid
-                                   .intersects(VIEWBOX2_PARAM,
-                                               use_index=details.viewbox.area < 0.5))
-
-        if details.near is not None and details.near_radius is not None:
-            if details.near_radius < 0.1:
-                sql = sql.where(t.c.centroid.within_distance(NEAR_PARAM,
-                                                             NEAR_RADIUS_PARAM))
-            else:
-                sql = sql.where(t.c.centroid
-                                 .ST_Distance(NEAR_PARAM) < NEAR_RADIUS_PARAM)
-
-        if self.housenumbers:
-            sql = sql.where(t.c.address_rank.between(16, 30))
-        else:
-            if details.excluded:
-                sql = sql.where(_exclude_places(t))
-            if details.min_rank > 0:
-                sql = sql.where(sa.or_(t.c.address_rank >= MIN_RANK_PARAM,
-                                       t.c.search_rank >= MIN_RANK_PARAM))
-            if details.max_rank < 30:
-                sql = sql.where(sa.or_(t.c.address_rank <= MAX_RANK_PARAM,
-                                       t.c.search_rank <= MAX_RANK_PARAM))
-
-        inner = sql.limit(10000).order_by(sa.desc(sa.text('importance'))).subquery()
-
-        sql = sa.select(inner.c.place_id, inner.c.search_rank, inner.c.address_rank,
-                        inner.c.country_code, inner.c.centroid, inner.c.importance,
-                        inner.c.penalty)
-
-        # If the query is not an address search or has a geographic preference,
-        # preselect most important items to restrict the number of places
-        # that need to be looked up in placex.
-        if not self.housenumbers\
-           and (details.viewbox is None or details.bounded_viewbox)\
-           and (details.near is None or details.near_radius is not None)\
-           and not self.qualifiers:
-            sql = sql.add_columns(sa.func.first_value(inner.c.penalty - inner.c.importance)
-                                    .over(order_by=inner.c.penalty - inner.c.importance)
-                                    .label('min_penalty'))
-
-            inner = sql.subquery()
-
-            sql = sa.select(inner.c.place_id, inner.c.search_rank, inner.c.address_rank,
-                            inner.c.country_code, inner.c.centroid, inner.c.importance,
-                            inner.c.penalty)\
-                    .where(inner.c.penalty - inner.c.importance < inner.c.min_penalty + 0.5)
-
-        return sql.cte('searches')
-
-    async def lookup(self, conn: SearchConnection,
-                     details: SearchDetails) -> nres.SearchResults:
-        """ Find results for the search in the database.
-        """
-        t = conn.t.placex
-        tsearch = self._inner_search_name_cte(conn, details)
-
-        sql = _select_placex(t).join(tsearch, t.c.place_id == tsearch.c.place_id)
-
-        if details.geometry_output:
-            sql = _add_geometry_columns(sql, t.c.geometry, details)
-
-        penalty: SaExpression = tsearch.c.penalty
-
-        if self.postcodes:
-            tpc = conn.t.postcode
-            pcs = self.postcodes.values
-
-            pc_near = sa.select(sa.func.min(tpc.c.geometry.ST_Distance(t.c.centroid)))\
-                        .where(tpc.c.postcode.in_(pcs))\
-                        .scalar_subquery()
-            penalty += sa.case((t.c.postcode.in_(pcs), 0.0),
-                               else_=sa.func.coalesce(pc_near, cast(SaColumn, 2.0)))
-
-        if details.viewbox is not None and not details.bounded_viewbox:
-            penalty += sa.case((t.c.geometry.intersects(VIEWBOX_PARAM, use_index=False), 0.0),
-                               (t.c.geometry.intersects(VIEWBOX2_PARAM, use_index=False), 0.5),
-                               else_=1.0)
-
-        if details.near is not None:
-            sql = sql.add_columns((-tsearch.c.centroid.ST_Distance(NEAR_PARAM))
-                                  .label('importance'))
-            sql = sql.order_by(sa.desc(sa.text('importance')))
-        else:
-            sql = sql.order_by(penalty - tsearch.c.importance)
-            sql = sql.add_columns(tsearch.c.importance)
-
-        sql = sql.add_columns(penalty.label('accuracy'))\
-                 .order_by(sa.text('accuracy'))
-
-        if self.housenumbers:
-            hnr_list = '|'.join(self.housenumbers.values)
-            inner = sql.where(sa.or_(tsearch.c.address_rank < 30,
-                                     sa.func.RegexpWord(hnr_list, t.c.housenumber)))\
-                       .subquery()
-
-            # Housenumbers from placex
-            thnr = conn.t.placex.alias('hnr')
-            pid_list = sa.func.ArrayAgg(thnr.c.place_id)
-            place_sql = sa.select(pid_list)\
-                          .where(thnr.c.parent_place_id == inner.c.place_id)\
-                          .where(sa.func.RegexpWord(hnr_list, thnr.c.housenumber))\
-                          .where(thnr.c.linked_place_id == None)\
-                          .where(thnr.c.indexed_status == 0)
-
-            if details.excluded:
-                place_sql = place_sql.where(thnr.c.place_id.not_in(sa.bindparam('excluded')))
-            if self.qualifiers:
-                place_sql = place_sql.where(self.qualifiers.sql_restrict(thnr))
-
-            numerals = [int(n) for n in self.housenumbers.values
-                        if n.isdigit() and len(n) < 8]
-            interpol_sql: SaColumn
-            tiger_sql: SaColumn
-            if numerals and \
-               (not self.qualifiers or ('place', 'house') in self.qualifiers.values):
-                # Housenumbers from interpolations
-                interpol_sql = _make_interpolation_subquery(conn.t.osmline, inner,
-                                                            numerals, details)
-                # Housenumbers from Tiger
-                tiger_sql = sa.case((inner.c.country_code == 'us',
-                                     _make_interpolation_subquery(conn.t.tiger, inner,
-                                                                  numerals, details)
-                                     ), else_=None)
-            else:
-                interpol_sql = sa.null()
-                tiger_sql = sa.null()
-
-            unsort = sa.select(inner, place_sql.scalar_subquery().label('placex_hnr'),
-                               interpol_sql.label('interpol_hnr'),
-                               tiger_sql.label('tiger_hnr')).subquery('unsort')
-            sql = sa.select(unsort)\
-                    .order_by(sa.case((unsort.c.placex_hnr != None, 1),
-                                      (unsort.c.interpol_hnr != None, 2),
-                                      (unsort.c.tiger_hnr != None, 3),
-                                      else_=4),
-                              unsort.c.accuracy)
-        else:
-            sql = sql.where(t.c.linked_place_id == None)\
-                     .where(t.c.indexed_status == 0)
-            if self.qualifiers:
-                sql = sql.where(self.qualifiers.sql_restrict(t))
-            if details.layers is not None:
-                sql = sql.where(_filter_by_layer(t, details.layers))
-
-        sql = sql.limit(LIMIT_PARAM)
-
-        results = nres.SearchResults()
-        for row in await conn.execute(sql, _details_to_bind_params(details)):
-            result = nres.create_from_placex_row(row, nres.SearchResult)
-            assert result
-            result.bbox = Bbox.from_wkb(row.bbox)
-            result.accuracy = row.accuracy
-            if self.housenumbers and row.rank_address < 30:
-                if row.placex_hnr:
-                    subs = _get_placex_housenumbers(conn, row.placex_hnr, details)
-                elif row.interpol_hnr:
-                    subs = _get_osmline(conn, row.interpol_hnr, numerals, details)
-                elif row.tiger_hnr:
-                    subs = _get_tiger(conn, row.tiger_hnr, numerals, row.osm_id, details)
-                else:
-                    subs = None
-
-                if subs is not None:
-                    async for sub in subs:
-                        assert sub.housenumber
-                        sub.accuracy = result.accuracy
-                        if not any(nr in self.housenumbers.values
-                                   for nr in sub.housenumber.split(';')):
-                            sub.accuracy += 0.6
-                        results.append(sub)
-
-                # Only add the street as a result, if it meets all other
-                # filter conditions.
-                if (not details.excluded or result.place_id not in details.excluded)\
-                   and (not self.qualifiers or result.category in self.qualifiers.values)\
-                   and result.rank_address >= details.min_rank:
-                    result.accuracy += 1.0  # penalty for missing housenumber
-                    results.append(result)
-            else:
-                results.append(result)
-
-        return results

src/nominatim_api/search/db_searches/__init__.py

@@ -0,0 +1,17 @@
+# SPDX-License-Identifier: GPL-3.0-or-later
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2025 by the Nominatim developer community.
+# For a full list of authors see the git log.
+"""
+Module implementing the actual database accesses for forward search.
+"""
+
+from .base import AbstractSearch as AbstractSearch
+from .near_search import NearSearch as NearSearch
+from .poi_search import PoiSearch as PoiSearch
+from .country_search import CountrySearch as CountrySearch
+from .postcode_search import PostcodeSearch as PostcodeSearch
+from .place_search import PlaceSearch as PlaceSearch
+from .address_search import AddressSearch as AddressSearch

src/nominatim_api/search/db_searches/address_search.py

@@ -0,0 +1,360 @@
+# SPDX-License-Identifier: GPL-3.0-or-later
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2025 by the Nominatim developer community.
+# For a full list of authors see the git log.
+"""
+Implementation of search for an address (search with housenumber).
+"""
+from typing import cast, List, AsyncIterator
+
+import sqlalchemy as sa
+
+from . import base
+from ...typing import SaBind, SaExpression, SaColumn, SaFromClause, SaScalarSelect
+from ...types import SearchDetails, Bbox
+from ...sql.sqlalchemy_types import Geometry
+from ...connection import SearchConnection
+from ... import results as nres
+from ..db_search_fields import SearchData
+
+
+LIMIT_PARAM: SaBind = sa.bindparam('limit')
+MIN_RANK_PARAM: SaBind = sa.bindparam('min_rank')
+MAX_RANK_PARAM: SaBind = sa.bindparam('max_rank')
+VIEWBOX_PARAM: SaBind = sa.bindparam('viewbox', type_=Geometry)
+VIEWBOX2_PARAM: SaBind = sa.bindparam('viewbox2', type_=Geometry)
+NEAR_PARAM: SaBind = sa.bindparam('near', type_=Geometry)
+NEAR_RADIUS_PARAM: SaBind = sa.bindparam('near_radius')
+COUNTRIES_PARAM: SaBind = sa.bindparam('countries')
+
+
+def _int_list_to_subquery(inp: List[int]) -> 'sa.Subquery':
+    """ Create a subselect that returns the given list of integers
+        as rows in the column 'nr'.
+    """
+    vtab = sa.func.JsonArrayEach(sa.type_coerce(inp, sa.JSON))\
+             .table_valued(sa.column('value', type_=sa.JSON))
+    return sa.select(sa.cast(sa.cast(vtab.c.value, sa.Text), sa.Integer).label('nr')).subquery()
+
+
+def _interpolated_position(table: SaFromClause, nr: SaColumn) -> SaColumn:
+    pos = sa.cast(nr - table.c.startnumber, sa.Float) / (table.c.endnumber - table.c.startnumber)
+    return sa.case(
+            (table.c.endnumber == table.c.startnumber, table.c.linegeo.ST_Centroid()),
+            else_=table.c.linegeo.ST_LineInterpolatePoint(pos)).label('centroid')
+
+
+def _make_interpolation_subquery(table: SaFromClause, inner: SaFromClause,
+                                 numerals: List[int], details: SearchDetails) -> SaScalarSelect:
+    all_ids = sa.func.ArrayAgg(table.c.place_id)
+    sql = sa.select(all_ids).where(table.c.parent_place_id == inner.c.place_id)
+
+    if len(numerals) == 1:
+        sql = sql.where(sa.between(numerals[0], table.c.startnumber, table.c.endnumber))\
+                 .where((numerals[0] - table.c.startnumber) % table.c.step == 0)
+    else:
+        sql = sql.where(sa.or_(
+                *(sa.and_(sa.between(n, table.c.startnumber, table.c.endnumber),
+                          (n - table.c.startnumber) % table.c.step == 0)
+                  for n in numerals)))
+
+    if details.excluded:
+        sql = sql.where(base.exclude_places(table))
+
+    return sql.scalar_subquery()
+
+
+async def _get_placex_housenumbers(conn: SearchConnection,
+                                   place_ids: List[int],
+                                   details: SearchDetails) -> AsyncIterator[nres.SearchResult]:
+    t = conn.t.placex
+    sql = base.select_placex(t).add_columns(t.c.importance)\
+                               .where(t.c.place_id.in_(place_ids))
+
+    if details.geometry_output:
+        sql = base.add_geometry_columns(sql, t.c.geometry, details)
+
+    for row in await conn.execute(sql):
+        result = nres.create_from_placex_row(row, nres.SearchResult)
+        assert result
+        result.bbox = Bbox.from_wkb(row.bbox)
+        yield result
+
+
+async def _get_osmline(conn: SearchConnection, place_ids: List[int],
+                       numerals: List[int],
+                       details: SearchDetails) -> AsyncIterator[nres.SearchResult]:
+    t = conn.t.osmline
+
+    values = _int_list_to_subquery(numerals)
+    sql = sa.select(t.c.place_id, t.c.osm_id,
+                    t.c.parent_place_id, t.c.address,
+                    values.c.nr.label('housenumber'),
+                    _interpolated_position(t, values.c.nr),
+                    t.c.postcode, t.c.country_code)\
+            .where(t.c.place_id.in_(place_ids))\
+            .join(values, values.c.nr.between(t.c.startnumber, t.c.endnumber))
+
+    if details.geometry_output:
+        sub = sql.subquery()
+        sql = base.add_geometry_columns(sa.select(sub), sub.c.centroid, details)
+
+    for row in await conn.execute(sql):
+        result = nres.create_from_osmline_row(row, nres.SearchResult)
+        assert result
+        yield result
+
+
+async def _get_tiger(conn: SearchConnection, place_ids: List[int],
+                     numerals: List[int], osm_id: int,
+                     details: SearchDetails) -> AsyncIterator[nres.SearchResult]:
+    t = conn.t.tiger
+    values = _int_list_to_subquery(numerals)
+    sql = sa.select(t.c.place_id, t.c.parent_place_id,
+                    sa.literal('W').label('osm_type'),
+                    sa.literal(osm_id).label('osm_id'),
+                    values.c.nr.label('housenumber'),
+                    _interpolated_position(t, values.c.nr),
+                    t.c.postcode)\
+            .where(t.c.place_id.in_(place_ids))\
+            .join(values, values.c.nr.between(t.c.startnumber, t.c.endnumber))
+
+    if details.geometry_output:
+        sub = sql.subquery()
+        sql = base.add_geometry_columns(sa.select(sub), sub.c.centroid, details)
+
+    for row in await conn.execute(sql):
+        result = nres.create_from_tiger_row(row, nres.SearchResult)
+        assert result
+        yield result
+
+
+class AddressSearch(base.AbstractSearch):
+    """ Generic search for an address or named place.
+    """
+    SEARCH_PRIO = 1
+
+    def __init__(self, extra_penalty: float, sdata: SearchData,
+                 expected_count: int, has_address_terms: bool) -> None:
+        assert sdata.housenumbers
+        super().__init__(sdata.penalty + extra_penalty)
+        self.countries = sdata.countries
+        self.postcodes = sdata.postcodes
+        self.housenumbers = sdata.housenumbers
+        self.qualifiers = sdata.qualifiers
+        self.lookups = sdata.lookups
+        self.rankings = sdata.rankings
+        self.expected_count = expected_count
+        self.has_address_terms = has_address_terms
+
+    def _inner_search_name_cte(self, conn: SearchConnection,
+                               details: SearchDetails) -> 'sa.CTE':
+        """ Create a subquery that preselects the rows in the search_name
+            table.
+        """
+        t = conn.t.search_name
+
+        penalty: SaExpression = sa.literal(self.penalty)
+        for ranking in self.rankings:
+            penalty += ranking.sql_penalty(t)
+
+        sql = sa.select(t.c.place_id, t.c.search_rank, t.c.address_rank,
+                        t.c.country_code, t.c.centroid,
+                        t.c.name_vector, t.c.nameaddress_vector,
+                        sa.case((t.c.importance > 0, t.c.importance),
+                                else_=0.40001-(sa.cast(t.c.search_rank, sa.Float())/75))
+                          .label('importance'),
+                        penalty.label('penalty'))
+
+        for lookup in self.lookups:
+            sql = sql.where(lookup.sql_condition(t))
+
+        if self.countries:
+            sql = sql.where(t.c.country_code.in_(self.countries.values))
+
+        if self.postcodes:
+            if self.expected_count > 10000:
+                tpc = conn.t.postcode
+                sql = sql.where(sa.select(tpc.c.postcode)
+                                  .where(tpc.c.postcode.in_(self.postcodes.values))
+                                  .where(tpc.c.country_code == t.c.country_code)
+                                  .where(t.c.centroid.within_distance(tpc.c.geometry, 0.4))
+                                  .exists())
+
+        if details.viewbox is not None:
+            if details.bounded_viewbox:
+                sql = sql.where(t.c.centroid
+                                   .intersects(VIEWBOX_PARAM,
+                                               use_index=details.viewbox.area < 0.2))
+
+        if details.near is not None and details.near_radius is not None:
+            if details.near_radius < 0.1:
+                sql = sql.where(t.c.centroid.within_distance(NEAR_PARAM,
+                                                             NEAR_RADIUS_PARAM))
+            else:
+                sql = sql.where(t.c.centroid
+                                 .ST_Distance(NEAR_PARAM) < NEAR_RADIUS_PARAM)
+
+        if self.has_address_terms:
+            sql = sql.where(t.c.address_rank.between(16, 30))
+        else:
+            # If no further address terms are given, then the base street must
+            # be in the name. No search for named POIs with the given house number.
+            sql = sql.where(t.c.address_rank.between(16, 27))
+
+        inner = sql.limit(10000).order_by(sa.desc(sa.text('importance'))).subquery()
+
+        sql = sa.select(inner.c.place_id, inner.c.search_rank, inner.c.address_rank,
+                        inner.c.country_code, inner.c.centroid, inner.c.importance,
+                        inner.c.penalty)
+
+        return sql.cte('searches')
+
+    async def lookup(self, conn: SearchConnection,
+                     details: SearchDetails) -> nres.SearchResults:
+        """ Find results for the search in the database.
+        """
+        t = conn.t.placex
+        tsearch = self._inner_search_name_cte(conn, details)
+
+        sql = base.select_placex(t).join(tsearch, t.c.place_id == tsearch.c.place_id)
+
+        if details.geometry_output:
+            sql = base.add_geometry_columns(sql, t.c.geometry, details)
+
+        penalty: SaExpression = tsearch.c.penalty
+
+        if self.postcodes:
+            tpc = conn.t.postcode
+            pcs = self.postcodes.values
+
+            pc_near = sa.select(sa.func.min(tpc.c.geometry.ST_Distance(t.c.centroid)
+                                            * (tpc.c.rank_search - 19)))\
+                        .where(tpc.c.postcode.in_(pcs))\
+                        .where(tpc.c.country_code == t.c.country_code)\
+                        .scalar_subquery()
+            penalty += sa.case((t.c.postcode.in_(pcs), 0.0),
+                               else_=sa.func.coalesce(pc_near, cast(SaColumn, 2.0)))
+
+        if details.viewbox is not None and not details.bounded_viewbox:
+            penalty += sa.case((t.c.geometry.intersects(VIEWBOX_PARAM, use_index=False), 0.0),
+                               (t.c.geometry.intersects(VIEWBOX2_PARAM, use_index=False), 0.5),
+                               else_=1.0)
+
+        if details.near is not None:
+            sql = sql.add_columns((-tsearch.c.centroid.ST_Distance(NEAR_PARAM))
+                                  .label('importance'))
+            sql = sql.order_by(sa.desc(sa.text('importance')))
+        else:
+            sql = sql.order_by(penalty - tsearch.c.importance)
+            sql = sql.add_columns(tsearch.c.importance)
+
+        sql = sql.add_columns(penalty.label('accuracy'))\
+                 .order_by(sa.text('accuracy'))
+
+        hnr_list = '|'.join(self.housenumbers.values)
+
+        if self.has_address_terms:
+            sql = sql.where(sa.or_(tsearch.c.address_rank < 30,
+                                   sa.func.RegexpWord(hnr_list, t.c.housenumber)))
+
+        inner = sql.subquery()
+
+        # Housenumbers from placex
+        thnr = conn.t.placex.alias('hnr')
+        pid_list = sa.func.ArrayAgg(thnr.c.place_id)
+        place_sql = sa.select(pid_list)\
+                      .where(thnr.c.parent_place_id == inner.c.place_id)\
+                      .where(sa.func.RegexpWord(hnr_list, thnr.c.housenumber))\
+                      .where(thnr.c.linked_place_id == None)\
+                      .where(thnr.c.indexed_status == 0)
+
+        if details.excluded:
+            place_sql = place_sql.where(thnr.c.place_id.not_in(sa.bindparam('excluded')))
+        if self.qualifiers:
+            place_sql = place_sql.where(self.qualifiers.sql_restrict(thnr))
+
+        numerals = [int(n) for n in self.housenumbers.values
+                    if n.isdigit() and len(n) < 8]
+        interpol_sql: SaColumn
+        tiger_sql: SaColumn
+        if numerals and \
+           (not self.qualifiers or ('place', 'house') in self.qualifiers.values):
+            # Housenumbers from interpolations
+            interpol_sql = _make_interpolation_subquery(conn.t.osmline, inner,
+                                                        numerals, details)
+            # Housenumbers from Tiger
+            tiger_sql = sa.case((inner.c.country_code == 'us',
+                                 _make_interpolation_subquery(conn.t.tiger, inner,
+                                                              numerals, details)
+                                 ), else_=None)
+        else:
+            interpol_sql = sa.null()
+            tiger_sql = sa.null()
+
+        unsort = sa.select(inner, place_sql.scalar_subquery().label('placex_hnr'),
+                           interpol_sql.label('interpol_hnr'),
+                           tiger_sql.label('tiger_hnr')).subquery('unsort')
+        sql = sa.select(unsort)\
+                .order_by(unsort.c.accuracy +
+                          sa.case((unsort.c.placex_hnr != None, 0),
+                                  (unsort.c.interpol_hnr != None, 0),
+                                  (unsort.c.tiger_hnr != None, 0),
+                                  else_=1),
+                          sa.case((unsort.c.placex_hnr != None, 1),
+                                  (unsort.c.interpol_hnr != None, 2),
+                                  (unsort.c.tiger_hnr != None, 3),
+                                  else_=4))
+
+        sql = sql.limit(LIMIT_PARAM)
+
+        bind_params = {
+            'limit': details.max_results,
+            'min_rank': details.min_rank,
+            'max_rank': details.max_rank,
+            'viewbox': details.viewbox,
+            'viewbox2': details.viewbox_x2,
+            'near': details.near,
+            'near_radius': details.near_radius,
+            'excluded': details.excluded,
+            'countries': details.countries
+        }
+
+        results = nres.SearchResults()
+        for row in await conn.execute(sql, bind_params):
+            result = nres.create_from_placex_row(row, nres.SearchResult)
+            assert result
+            result.bbox = Bbox.from_wkb(row.bbox)
+            result.accuracy = row.accuracy
+            if row.rank_address < 30:
+                if row.placex_hnr:
+                    subs = _get_placex_housenumbers(conn, row.placex_hnr, details)
+                elif row.interpol_hnr:
+                    subs = _get_osmline(conn, row.interpol_hnr, numerals, details)
+                elif row.tiger_hnr:
+                    subs = _get_tiger(conn, row.tiger_hnr, numerals, row.osm_id, details)
+                else:
+                    subs = None
+
+                if subs is not None:
+                    async for sub in subs:
+                        assert sub.housenumber
+                        sub.accuracy = result.accuracy
+                        if not any(nr in self.housenumbers.values
+                                   for nr in sub.housenumber.split(';')):
+                            sub.accuracy += 0.6
+                        results.append(sub)
+
+                # Only add the street as a result, if it meets all other
+                # filter conditions.
+                if (not details.excluded or result.place_id not in details.excluded)\
+                   and (not self.qualifiers or result.category in self.qualifiers.values)\
+                   and result.rank_address >= details.min_rank:
+                    result.accuracy += 1.0  # penalty for missing housenumber
+                    results.append(result)
+            else:
+                results.append(result)
+
+        return results

src/nominatim_api/search/db_searches/base.py

@@ -0,0 +1,144 @@
+# SPDX-License-Identifier: GPL-3.0-or-later
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2025 by the Nominatim developer community.
+# For a full list of authors see the git log.
+"""
+Interface for classes implementing a database search.
+"""
+from typing import Callable, List
+import abc
+
+import sqlalchemy as sa
+
+from ...typing import SaFromClause, SaSelect, SaColumn, SaExpression, SaLambdaSelect
+from ...sql.sqlalchemy_types import Geometry
+from ...connection import SearchConnection
+from ...types import SearchDetails, DataLayer, GeometryFormat
+from ...results import SearchResults
+
+
+class AbstractSearch(abc.ABC):
+    """ Encapuslation of a single lookup in the database.
+    """
+    SEARCH_PRIO: int = 2
+
+    def __init__(self, penalty: float) -> None:
+        self.penalty = penalty
+
+    @abc.abstractmethod
+    async def lookup(self, conn: SearchConnection, details: SearchDetails) -> SearchResults:
+        """ Find results for the search in the database.
+        """
+
+
+def select_placex(t: SaFromClause) -> SaSelect:
+    """ Return the basic select query for placex which returns all
+        fields necessary to fill a Nominatim result. 't' must either be
+        the placex table or a subquery returning appropriate fields from
+        a placex-related query.
+    """
+    return sa.select(t.c.place_id, t.c.osm_type, t.c.osm_id, t.c.name,
+                     t.c.class_, t.c.type,
+                     t.c.address, t.c.extratags,
+                     t.c.housenumber, t.c.postcode, t.c.country_code,
+                     t.c.wikipedia,
+                     t.c.parent_place_id, t.c.rank_address, t.c.rank_search,
+                     t.c.linked_place_id, t.c.admin_level,
+                     t.c.centroid,
+                     t.c.geometry.ST_Expand(0).label('bbox'))
+
+
+def exclude_places(t: SaFromClause) -> Callable[[], SaExpression]:
+    """ Return an expression to exclude place IDs from the list in the
+        SearchDetails.
+
+        Requires the excluded IDs to be supplied as a bind parameter in SQL.
+    """
+    return lambda: t.c.place_id.not_in(sa.bindparam('excluded'))
+
+
+def filter_by_layer(table: SaFromClause, layers: DataLayer) -> SaColumn:
+    """ Return an expression that filters the given table by layers.
+    """
+    orexpr: List[SaExpression] = []
+    if layers & DataLayer.ADDRESS and layers & DataLayer.POI:
+        orexpr.append(no_index(table.c.rank_address).between(1, 30))
+    elif layers & DataLayer.ADDRESS:
+        orexpr.append(no_index(table.c.rank_address).between(1, 29))
+        orexpr.append(sa.func.IsAddressPoint(table))
+    elif layers & DataLayer.POI:
+        orexpr.append(sa.and_(no_index(table.c.rank_address) == 30,
+                              table.c.class_.not_in(('place', 'building'))))
+
+    if layers & DataLayer.MANMADE:
+        exclude = []
+        if not layers & DataLayer.RAILWAY:
+            exclude.append('railway')
+        if not layers & DataLayer.NATURAL:
+            exclude.extend(('natural', 'water', 'waterway'))
+        orexpr.append(sa.and_(table.c.class_.not_in(tuple(exclude)),
+                              no_index(table.c.rank_address) == 0))
+    else:
+        include = []
+        if layers & DataLayer.RAILWAY:
+            include.append('railway')
+        if layers & DataLayer.NATURAL:
+            include.extend(('natural', 'water', 'waterway'))
+        orexpr.append(sa.and_(table.c.class_.in_(tuple(include)),
+                              no_index(table.c.rank_address) == 0))
+
+    if len(orexpr) == 1:
+        return orexpr[0]
+
+    return sa.or_(*orexpr)
+
+
+def no_index(expr: SaColumn) -> SaColumn:
+    """ Wrap the given expression, so that the query planner will
+        refrain from using the expression for index lookup.
+    """
+    return sa.func.coalesce(sa.null(), expr)
+
+
+def filter_by_area(sql: SaSelect, t: SaFromClause,
+                   details: SearchDetails, avoid_index: bool = False) -> SaSelect:
+    """ Apply SQL statements for filtering by viewbox and near point,
+        if applicable.
+    """
+    if details.near is not None and details.near_radius is not None:
+        if details.near_radius < 0.1 and not avoid_index:
+            sql = sql.where(
+                t.c.geometry.within_distance(sa.bindparam('near', type_=Geometry),
+                                             sa.bindparam('near_radius')))
+        else:
+            sql = sql.where(
+                t.c.geometry.ST_Distance(
+                    sa.bindparam('near', type_=Geometry)) <= sa.bindparam('near_radius'))
+    if details.viewbox is not None and details.bounded_viewbox:
+        sql = sql.where(t.c.geometry.intersects(sa.bindparam('viewbox', type_=Geometry),
+                                                use_index=not avoid_index and
+                                                details.viewbox.area < 0.2))
+
+    return sql
+
+
+def add_geometry_columns(sql: SaLambdaSelect, col: SaColumn, details: SearchDetails) -> SaSelect:
+    """ Add columns for requested geometry formats and return the new query.
+    """
+    out = []
+
+    if details.geometry_simplification > 0.0:
+        col = sa.func.ST_SimplifyPreserveTopology(col, details.geometry_simplification)
+
+    if details.geometry_output & GeometryFormat.GEOJSON:
+        out.append(sa.func.ST_AsGeoJSON(col, 7).label('geometry_geojson'))
+    if details.geometry_output & GeometryFormat.TEXT:
+        out.append(sa.func.ST_AsText(col).label('geometry_text'))
+    if details.geometry_output & GeometryFormat.KML:
+        out.append(sa.func.ST_AsKML(col, 7).label('geometry_kml'))
+    if details.geometry_output & GeometryFormat.SVG:
+        out.append(sa.func.ST_AsSVG(col, 0, 7).label('geometry_svg'))
+
+    return sql.add_columns(*out)

src/nominatim_api/search/db_searches/country_search.py

@@ -0,0 +1,119 @@
+# SPDX-License-Identifier: GPL-3.0-or-later
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2025 by the Nominatim developer community.
+# For a full list of authors see the git log.
+"""
+Implementation of searches for a country.
+"""
+
+import sqlalchemy as sa
+
+from . import base
+from ..db_search_fields import SearchData
+from ... import results as nres
+from ...connection import SearchConnection
+from ...types import SearchDetails, Bbox
+
+
+class CountrySearch(base.AbstractSearch):
+    """ Search for a country name or country code.
+    """
+    SEARCH_PRIO = 0
+
+    def __init__(self, sdata: SearchData) -> None:
+        super().__init__(sdata.penalty)
+        self.countries = sdata.countries
+
+    async def lookup(self, conn: SearchConnection,
+                     details: SearchDetails) -> nres.SearchResults:
+        """ Find results for the search in the database.
+        """
+        t = conn.t.placex
+
+        ccodes = self.countries.values
+        sql = base.select_placex(t)\
+            .add_columns(t.c.importance)\
+            .where(t.c.country_code.in_(ccodes))\
+            .where(t.c.rank_address == 4)
+
+        if details.geometry_output:
+            sql = base.add_geometry_columns(sql, t.c.geometry, details)
+
+        if details.excluded:
+            sql = sql.where(base.exclude_places(t))
+
+        sql = base.filter_by_area(sql, t, details)
+
+        bind_params = {
+            'excluded': details.excluded,
+            'viewbox': details.viewbox,
+            'near': details.near,
+            'near_radius': details.near_radius
+        }
+
+        results = nres.SearchResults()
+        for row in await conn.execute(sql, bind_params):
+            result = nres.create_from_placex_row(row, nres.SearchResult)
+            assert result
+            result.accuracy = self.penalty + self.countries.get_penalty(row.country_code, 5.0)
+            result.bbox = Bbox.from_wkb(row.bbox)
+            results.append(result)
+
+        if not results:
+            results = await self.lookup_in_country_table(conn, details)
+
+        if results:
+            details.min_rank = min(5, details.max_rank)
+            details.max_rank = min(25, details.max_rank)
+
+        return results
+
+    async def lookup_in_country_table(self, conn: SearchConnection,
+                                      details: SearchDetails) -> nres.SearchResults:
+        """ Look up the country in the fallback country tables.
+        """
+        # Avoid the fallback search when this is a more search. Country results
+        # usually are in the first batch of results and it is not possible
+        # to exclude these fallbacks.
+        if details.excluded:
+            return nres.SearchResults()
+
+        t = conn.t.country_name
+        tgrid = conn.t.country_grid
+
+        sql = sa.select(tgrid.c.country_code,
+                        tgrid.c.geometry.ST_Centroid().ST_Collect().ST_Centroid()
+                             .label('centroid'),
+                        tgrid.c.geometry.ST_Collect().ST_Expand(0).label('bbox'))\
+                .where(tgrid.c.country_code.in_(self.countries.values))\
+                .group_by(tgrid.c.country_code)
+
+        sql = base.filter_by_area(sql, tgrid, details, avoid_index=True)
+
+        sub = sql.subquery('grid')
+
+        sql = sa.select(t.c.country_code,
+                        t.c.name.merge(t.c.derived_name).label('name'),
+                        sub.c.centroid, sub.c.bbox)\
+                .join(sub, t.c.country_code == sub.c.country_code)
+
+        if details.geometry_output:
+            sql = base.add_geometry_columns(sql, sub.c.centroid, details)
+
+        bind_params = {
+            'viewbox': details.viewbox,
+            'near': details.near,
+            'near_radius': details.near_radius
+        }
+
+        results = nres.SearchResults()
+        for row in await conn.execute(sql, bind_params):
+            result = nres.create_from_country_row(row, nres.SearchResult)
+            assert result
+            result.bbox = Bbox.from_wkb(row.bbox)
+            result.accuracy = self.penalty + self.countries.get_penalty(row.country_code, 5.0)
+            results.append(result)
+
+        return results

src/nominatim_api/search/db_searches/near_search.py

@@ -0,0 +1,136 @@
+# SPDX-License-Identifier: GPL-3.0-or-later
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2025 by the Nominatim developer community.
+# For a full list of authors see the git log.
+"""
+Implementation of a category search around a place.
+"""
+from typing import List, Tuple
+
+import sqlalchemy as sa
+
+from . import base
+from ...typing import SaBind
+from ...types import SearchDetails, Bbox
+from ...connection import SearchConnection
+from ... import results as nres
+from ..db_search_fields import WeightedCategories
+
+
+LIMIT_PARAM: SaBind = sa.bindparam('limit')
+MIN_RANK_PARAM: SaBind = sa.bindparam('min_rank')
+MAX_RANK_PARAM: SaBind = sa.bindparam('max_rank')
+COUNTRIES_PARAM: SaBind = sa.bindparam('countries')
+
+
+class NearSearch(base.AbstractSearch):
+    """ Category search of a place type near the result of another search.
+    """
+    def __init__(self, penalty: float, categories: WeightedCategories,
+                 search: base.AbstractSearch) -> None:
+        super().__init__(penalty)
+        self.search = search
+        self.categories = categories
+
+    async def lookup(self, conn: SearchConnection,
+                     details: SearchDetails) -> nres.SearchResults:
+        """ Find results for the search in the database.
+        """
+        results = nres.SearchResults()
+        base = await self.search.lookup(conn, details)
+
+        if not base:
+            return results
+
+        base.sort(key=lambda r: (r.accuracy, r.rank_search))
+        max_accuracy = base[0].accuracy + 0.5
+        if base[0].rank_address == 0:
+            min_rank = 0
+            max_rank = 0
+        elif base[0].rank_address < 26:
+            min_rank = 1
+            max_rank = min(25, base[0].rank_address + 4)
+        else:
+            min_rank = 26
+            max_rank = 30
+        base = nres.SearchResults(r for r in base
+                                  if (r.source_table == nres.SourceTable.PLACEX
+                                      and r.accuracy <= max_accuracy
+                                      and r.bbox and r.bbox.area < 20
+                                      and r.rank_address >= min_rank
+                                      and r.rank_address <= max_rank))
+
+        if base:
+            baseids = [b.place_id for b in base[:5] if b.place_id]
+
+            for category, penalty in self.categories:
+                await self.lookup_category(results, conn, baseids, category, penalty, details)
+                if len(results) >= details.max_results:
+                    break
+
+        return results
+
+    async def lookup_category(self, results: nres.SearchResults,
+                              conn: SearchConnection, ids: List[int],
+                              category: Tuple[str, str], penalty: float,
+                              details: SearchDetails) -> None:
+        """ Find places of the given category near the list of
+            place ids and add the results to 'results'.
+        """
+        table = await conn.get_class_table(*category)
+
+        tgeom = conn.t.placex.alias('pgeom')
+
+        if table is None:
+            # No classtype table available, do a simplified lookup in placex.
+            table = conn.t.placex
+            sql = sa.select(table.c.place_id,
+                            sa.func.min(tgeom.c.centroid.ST_Distance(table.c.centroid))
+                              .label('dist'))\
+                    .join(tgeom, table.c.geometry.intersects(tgeom.c.centroid.ST_Expand(0.01)))\
+                    .where(table.c.class_ == category[0])\
+                    .where(table.c.type == category[1])
+        else:
+            # Use classtype table. We can afford to use a larger
+            # radius for the lookup.
+            sql = sa.select(table.c.place_id,
+                            sa.func.min(tgeom.c.centroid.ST_Distance(table.c.centroid))
+                              .label('dist'))\
+                    .join(tgeom,
+                          table.c.centroid.ST_CoveredBy(
+                              sa.case((sa.and_(tgeom.c.rank_address > 9,
+                                               tgeom.c.geometry.is_area()),
+                                       tgeom.c.geometry),
+                                      else_=tgeom.c.centroid.ST_Expand(0.05))))
+
+        inner = sql.where(tgeom.c.place_id.in_(ids))\
+                   .group_by(table.c.place_id).subquery()
+
+        t = conn.t.placex
+        sql = base.select_placex(t).add_columns((-inner.c.dist).label('importance'))\
+                                   .join(inner, inner.c.place_id == t.c.place_id)\
+                                   .order_by(inner.c.dist)
+
+        sql = sql.where(base.no_index(t.c.rank_address).between(MIN_RANK_PARAM, MAX_RANK_PARAM))
+        if details.countries:
+            sql = sql.where(t.c.country_code.in_(COUNTRIES_PARAM))
+        if details.excluded:
+            sql = sql.where(base.exclude_places(t))
+        if details.layers is not None:
+            sql = sql.where(base.filter_by_layer(t, details.layers))
+
+        sql = sql.limit(LIMIT_PARAM)
+
+        bind_params = {'limit': details.max_results,
+                       'min_rank': details.min_rank,
+                       'max_rank': details.max_rank,
+                       'excluded': details.excluded,
+                       'countries': details.countries}
+        for row in await conn.execute(sql, bind_params):
+            result = nres.create_from_placex_row(row, nres.SearchResult)
+            assert result
+            result.accuracy = self.penalty + penalty
+            result.bbox = Bbox.from_wkb(row.bbox)
+            results.append(result)

src/nominatim_api/search/db_searches/place_search.py

@@ -0,0 +1,210 @@
+# SPDX-License-Identifier: GPL-3.0-or-later
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2025 by the Nominatim developer community.
+# For a full list of authors see the git log.
+"""
+Implementation of search for a named place (without housenumber).
+"""
+from typing import cast
+
+import sqlalchemy as sa
+
+from . import base
+from ...typing import SaBind, SaExpression, SaColumn
+from ...types import SearchDetails, Bbox
+from ...sql.sqlalchemy_types import Geometry
+from ...connection import SearchConnection
+from ... import results as nres
+from ..db_search_fields import SearchData
+
+
+LIMIT_PARAM: SaBind = sa.bindparam('limit')
+MIN_RANK_PARAM: SaBind = sa.bindparam('min_rank')
+MAX_RANK_PARAM: SaBind = sa.bindparam('max_rank')
+VIEWBOX_PARAM: SaBind = sa.bindparam('viewbox', type_=Geometry)
+VIEWBOX2_PARAM: SaBind = sa.bindparam('viewbox2', type_=Geometry)
+NEAR_PARAM: SaBind = sa.bindparam('near', type_=Geometry)
+NEAR_RADIUS_PARAM: SaBind = sa.bindparam('near_radius')
+COUNTRIES_PARAM: SaBind = sa.bindparam('countries')
+
+
+class PlaceSearch(base.AbstractSearch):
+    """ Generic search for a named place.
+    """
+    SEARCH_PRIO = 1
+
+    def __init__(self, extra_penalty: float, sdata: SearchData,
+                 expected_count: int, has_address_terms: bool) -> None:
+        assert not sdata.housenumbers
+        super().__init__(sdata.penalty + extra_penalty)
+        self.countries = sdata.countries
+        self.postcodes = sdata.postcodes
+        self.qualifiers = sdata.qualifiers
+        self.lookups = sdata.lookups
+        self.rankings = sdata.rankings
+        self.expected_count = expected_count
+        self.has_address_terms = has_address_terms
+
+    def _inner_search_name_cte(self, conn: SearchConnection,
+                               details: SearchDetails) -> 'sa.CTE':
+        """ Create a subquery that preselects the rows in the search_name
+            table.
+        """
+        t = conn.t.search_name
+
+        penalty: SaExpression = sa.literal(self.penalty)
+        for ranking in self.rankings:
+            penalty += ranking.sql_penalty(t)
+
+        sql = sa.select(t.c.place_id, t.c.search_rank, t.c.address_rank,
+                        t.c.country_code, t.c.centroid,
+                        t.c.name_vector, t.c.nameaddress_vector,
+                        sa.case((t.c.importance > 0, t.c.importance),
+                                else_=0.40001-(sa.cast(t.c.search_rank, sa.Float())/75))
+                          .label('importance'))
+
+        for lookup in self.lookups:
+            sql = sql.where(lookup.sql_condition(t))
+
+        if self.countries:
+            sql = sql.where(t.c.country_code.in_(self.countries.values))
+
+        if self.postcodes:
+            # if a postcode is given, don't search for state or country level objects
+            sql = sql.where(t.c.address_rank > 9)
+            if self.expected_count > 10000:
+                # Many results expected. Restrict by postcode.
+                tpc = conn.t.postcode
+                sql = sql.where(sa.select(tpc.c.postcode)
+                                  .where(tpc.c.postcode.in_(self.postcodes.values))
+                                  .where(t.c.centroid.within_distance(tpc.c.geometry, 0.4))
+                                  .exists())
+
+        if details.viewbox is not None:
+            if details.bounded_viewbox:
+                sql = sql.where(t.c.centroid
+                                   .intersects(VIEWBOX_PARAM,
+                                               use_index=details.viewbox.area < 0.2))
+            else:
+                penalty += sa.case((t.c.centroid.intersects(VIEWBOX_PARAM, use_index=False), 0.0),
+                                   (t.c.centroid.intersects(VIEWBOX2_PARAM, use_index=False), 0.5),
+                                   else_=1.0)
+
+        if details.near is not None and details.near_radius is not None:
+            if details.near_radius < 0.1:
+                sql = sql.where(t.c.centroid.within_distance(NEAR_PARAM,
+                                                             NEAR_RADIUS_PARAM))
+            else:
+                sql = sql.where(t.c.centroid
+                                 .ST_Distance(NEAR_PARAM) < NEAR_RADIUS_PARAM)
+
+        if details.excluded:
+            sql = sql.where(base.exclude_places(t))
+        if details.min_rank > 0:
+            sql = sql.where(sa.or_(t.c.address_rank >= MIN_RANK_PARAM,
+                                   t.c.search_rank >= MIN_RANK_PARAM))
+        if details.max_rank < 30:
+            sql = sql.where(sa.or_(t.c.address_rank <= MAX_RANK_PARAM,
+                                   t.c.search_rank <= MAX_RANK_PARAM))
+
+        sql = sql.add_columns(penalty.label('penalty'))
+
+        inner = sql.limit(5000 if self.qualifiers else 1000)\
+                   .order_by(sa.desc(sa.text('importance')))\
+                   .subquery()
+
+        sql = sa.select(inner.c.place_id, inner.c.search_rank, inner.c.address_rank,
+                        inner.c.country_code, inner.c.centroid, inner.c.importance,
+                        inner.c.penalty)
+
+        # If the query is not an address search or has a geographic preference,
+        # preselect most important items to restrict the number of places
+        # that need to be looked up in placex.
+        if (details.viewbox is None or not details.bounded_viewbox)\
+           and (details.near is None or details.near_radius is None)\
+           and not self.qualifiers:
+            sql = sql.add_columns(sa.func.first_value(inner.c.penalty - inner.c.importance)
+                                    .over(order_by=inner.c.penalty - inner.c.importance)
+                                    .label('min_penalty'))
+
+            inner = sql.subquery()
+
+            sql = sa.select(inner.c.place_id, inner.c.search_rank, inner.c.address_rank,
+                            inner.c.country_code, inner.c.centroid, inner.c.importance,
+                            inner.c.penalty)\
+                    .where(inner.c.penalty - inner.c.importance < inner.c.min_penalty + 0.5)
+
+        return sql.cte('searches')
+
+    async def lookup(self, conn: SearchConnection,
+                     details: SearchDetails) -> nres.SearchResults:
+        """ Find results for the search in the database.
+        """
+        t = conn.t.placex
+        tsearch = self._inner_search_name_cte(conn, details)
+
+        sql = base.select_placex(t).join(tsearch, t.c.place_id == tsearch.c.place_id)
+
+        if details.geometry_output:
+            sql = base.add_geometry_columns(sql, t.c.geometry, details)
+
+        penalty: SaExpression = tsearch.c.penalty
+
+        if self.postcodes:
+            if self.has_address_terms:
+                tpc = conn.t.postcode
+                pcs = self.postcodes.values
+
+                pc_near = sa.select(sa.func.min(tpc.c.geometry.ST_Distance(t.c.centroid)))\
+                            .where(tpc.c.postcode.in_(pcs))\
+                            .scalar_subquery()
+                penalty += sa.case((t.c.postcode.in_(pcs), 0.0),
+                                   else_=sa.func.coalesce(pc_near, cast(SaColumn, 2.0)))
+            else:
+                # High penalty if the postcode is not an exact match.
+                # The postcode search needs to get priority here.
+                penalty += sa.case((t.c.postcode.in_(self.postcodes.values), 0.0), else_=1.0)
+
+        if details.near is not None:
+            sql = sql.add_columns((-tsearch.c.centroid.ST_Distance(NEAR_PARAM))
+                                  .label('importance'))
+            sql = sql.order_by(sa.desc(sa.text('importance')))
+        else:
+            sql = sql.order_by(penalty - tsearch.c.importance)
+            sql = sql.add_columns(tsearch.c.importance)
+
+        sql = sql.add_columns(penalty.label('accuracy'))\
+                 .order_by(sa.text('accuracy'))
+
+        sql = sql.where(t.c.linked_place_id == None)\
+                 .where(t.c.indexed_status == 0)
+        if self.qualifiers:
+            sql = sql.where(self.qualifiers.sql_restrict(t))
+        if details.layers is not None:
+            sql = sql.where(base.filter_by_layer(t, details.layers))
+
+        sql = sql.limit(LIMIT_PARAM)
+
+        bind_params = {
+            'limit': details.max_results,
+            'min_rank': details.min_rank,
+            'max_rank': details.max_rank,
+            'viewbox': details.viewbox,
+            'viewbox2': details.viewbox_x2,
+            'near': details.near,
+            'near_radius': details.near_radius,
+            'excluded': details.excluded,
+            'countries': details.countries
+        }
+
+        results = nres.SearchResults()
+        for row in await conn.execute(sql, bind_params):
+            result = nres.create_from_placex_row(row, nres.SearchResult)
+            assert result
+            result.bbox = Bbox.from_wkb(row.bbox)
+            result.accuracy = row.accuracy
+            results.append(result)
+
+        return results

src/nominatim_api/search/db_searches/poi_search.py

@@ -0,0 +1,114 @@
+# SPDX-License-Identifier: GPL-3.0-or-later
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2025 by the Nominatim developer community.
+# For a full list of authors see the git log.
+"""
+Implementation of category search.
+"""
+from typing import List
+
+import sqlalchemy as sa
+
+from . import base
+from ..db_search_fields import SearchData
+from ... import results as nres
+from ...typing import SaBind, SaRow, SaSelect, SaLambdaSelect
+from ...sql.sqlalchemy_types import Geometry
+from ...connection import SearchConnection
+from ...types import SearchDetails, Bbox
+
+
+LIMIT_PARAM: SaBind = sa.bindparam('limit')
+VIEWBOX_PARAM: SaBind = sa.bindparam('viewbox', type_=Geometry)
+NEAR_PARAM: SaBind = sa.bindparam('near', type_=Geometry)
+NEAR_RADIUS_PARAM: SaBind = sa.bindparam('near_radius')
+
+
+class PoiSearch(base.AbstractSearch):
+    """ Category search in a geographic area.
+    """
+    def __init__(self, sdata: SearchData) -> None:
+        super().__init__(sdata.penalty)
+        self.qualifiers = sdata.qualifiers
+        self.countries = sdata.countries
+
+    async def lookup(self, conn: SearchConnection,
+                     details: SearchDetails) -> nres.SearchResults:
+        """ Find results for the search in the database.
+        """
+        bind_params = {
+            'limit': details.max_results,
+            'viewbox': details.viewbox,
+            'near': details.near,
+            'near_radius': details.near_radius,
+            'excluded': details.excluded
+        }
+
+        t = conn.t.placex
+
+        rows: List[SaRow] = []
+
+        if details.near and details.near_radius is not None and details.near_radius < 0.2:
+            # simply search in placex table
+            def _base_query() -> SaSelect:
+                return base.select_placex(t) \
+                           .add_columns((-t.c.centroid.ST_Distance(NEAR_PARAM))
+                                        .label('importance'))\
+                           .where(t.c.linked_place_id == None) \
+                           .where(t.c.geometry.within_distance(NEAR_PARAM, NEAR_RADIUS_PARAM)) \
+                           .order_by(t.c.centroid.ST_Distance(NEAR_PARAM)) \
+                           .limit(LIMIT_PARAM)
+
+            classtype = self.qualifiers.values
+            if len(classtype) == 1:
+                cclass, ctype = classtype[0]
+                sql: SaLambdaSelect = sa.lambda_stmt(
+                    lambda: _base_query().where(t.c.class_ == cclass)
+                                         .where(t.c.type == ctype))
+            else:
+                sql = _base_query().where(sa.or_(*(sa.and_(t.c.class_ == cls, t.c.type == typ)
+                                                   for cls, typ in classtype)))
+
+            if self.countries:
+                sql = sql.where(t.c.country_code.in_(self.countries.values))
+
+            if details.viewbox is not None and details.bounded_viewbox:
+                sql = sql.where(t.c.geometry.intersects(VIEWBOX_PARAM))
+
+            rows.extend(await conn.execute(sql, bind_params))
+        else:
+            # use the class type tables
+            for category in self.qualifiers.values:
+                table = await conn.get_class_table(*category)
+                if table is not None:
+                    sql = base.select_placex(t)\
+                               .add_columns(t.c.importance)\
+                               .join(table, t.c.place_id == table.c.place_id)\
+                               .where(t.c.class_ == category[0])\
+                               .where(t.c.type == category[1])
+
+                    if details.viewbox is not None and details.bounded_viewbox:
+                        sql = sql.where(table.c.centroid.intersects(VIEWBOX_PARAM))
+
+                    if details.near and details.near_radius is not None:
+                        sql = sql.order_by(table.c.centroid.ST_Distance(NEAR_PARAM))\
+                                 .where(table.c.centroid.within_distance(NEAR_PARAM,
+                                                                         NEAR_RADIUS_PARAM))
+
+                    if self.countries:
+                        sql = sql.where(t.c.country_code.in_(self.countries.values))
+
+                    sql = sql.limit(LIMIT_PARAM)
+                    rows.extend(await conn.execute(sql, bind_params))
+
+        results = nres.SearchResults()
+        for row in rows:
+            result = nres.create_from_placex_row(row, nres.SearchResult)
+            assert result
+            result.accuracy = self.penalty + self.qualifiers.get_penalty((row.class_, row.type))
+            result.bbox = Bbox.from_wkb(row.bbox)
+            results.append(result)
+
+        return results

src/nominatim_api/search/db_searches/postcode_search.py

@@ -0,0 +1,129 @@
+# SPDX-License-Identifier: GPL-3.0-or-later
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2025 by the Nominatim developer community.
+# For a full list of authors see the git log.
+"""
+Implementation of search for a postcode.
+"""
+
+import sqlalchemy as sa
+
+from . import base
+from ...typing import SaBind, SaExpression
+from ...sql.sqlalchemy_types import Geometry, IntArray
+from ...connection import SearchConnection
+from ...types import SearchDetails, Bbox
+from ... import results as nres
+from ..db_search_fields import SearchData
+
+
+LIMIT_PARAM: SaBind = sa.bindparam('limit')
+VIEWBOX_PARAM: SaBind = sa.bindparam('viewbox', type_=Geometry)
+VIEWBOX2_PARAM: SaBind = sa.bindparam('viewbox2', type_=Geometry)
+NEAR_PARAM: SaBind = sa.bindparam('near', type_=Geometry)
+
+
+class PostcodeSearch(base.AbstractSearch):
+    """ Search for a postcode.
+    """
+    def __init__(self, extra_penalty: float, sdata: SearchData) -> None:
+        super().__init__(sdata.penalty + extra_penalty)
+        self.countries = sdata.countries
+        self.postcodes = sdata.postcodes
+        self.lookups = sdata.lookups
+        self.rankings = sdata.rankings
+
+    async def lookup(self, conn: SearchConnection,
+                     details: SearchDetails) -> nres.SearchResults:
+        """ Find results for the search in the database.
+        """
+        t = conn.t.postcode
+        pcs = self.postcodes.values
+
+        sql = sa.select(t.c.place_id, t.c.parent_place_id,
+                        t.c.rank_search, t.c.rank_address,
+                        t.c.postcode, t.c.country_code,
+                        t.c.geometry.label('centroid'))\
+                .where(t.c.postcode.in_(pcs))
+
+        if details.geometry_output:
+            sql = base.add_geometry_columns(sql, t.c.geometry, details)
+
+        penalty: SaExpression = sa.literal(self.penalty)
+
+        if details.viewbox is not None and not details.bounded_viewbox:
+            penalty += sa.case((t.c.geometry.intersects(VIEWBOX_PARAM), 0.0),
+                               (t.c.geometry.intersects(VIEWBOX2_PARAM), 0.5),
+                               else_=1.0)
+
+        if details.near is not None:
+            sql = sql.order_by(t.c.geometry.ST_Distance(NEAR_PARAM))
+
+        sql = base.filter_by_area(sql, t, details)
+
+        if self.countries:
+            sql = sql.where(t.c.country_code.in_(self.countries.values))
+
+        if details.excluded:
+            sql = sql.where(base.exclude_places(t))
+
+        if self.lookups:
+            assert len(self.lookups) == 1
+            tsearch = conn.t.search_name
+            sql = sql.where(tsearch.c.place_id == t.c.parent_place_id)\
+                     .where((tsearch.c.name_vector + tsearch.c.nameaddress_vector)
+                            .contains(sa.type_coerce(self.lookups[0].tokens,
+                                                     IntArray)))
+            # Do NOT add rerank penalties based on the address terms.
+            # The standard rerank penalty only checks the address vector
+            # while terms may appear in name and address vector. This would
+            # lead to overly high penalties.
+            # We assume that a postcode is precise enough to not require
+            # additional full name matches.
+
+        penalty += sa.case(*((t.c.postcode == v, p) for v, p in self.postcodes),
+                           else_=1.0)
+
+        sql = sql.add_columns(penalty.label('accuracy'))
+        sql = sql.order_by('accuracy').limit(LIMIT_PARAM)
+
+        bind_params = {
+            'limit': details.max_results,
+            'viewbox': details.viewbox,
+            'viewbox2': details.viewbox_x2,
+            'near': details.near,
+            'near_radius': details.near_radius,
+            'excluded': details.excluded
+        }
+
+        results = nres.SearchResults()
+        for row in await conn.execute(sql, bind_params):
+            p = conn.t.placex
+            placex_sql = base.select_placex(p)\
+                .add_columns(p.c.importance)\
+                .where(sa.text("""class = 'boundary'
+                                  AND type = 'postal_code'
+                                  AND osm_type = 'R'"""))\
+                .where(p.c.country_code == row.country_code)\
+                .where(p.c.postcode == row.postcode)\
+                .limit(1)
+
+            if details.geometry_output:
+                placex_sql = base.add_geometry_columns(placex_sql, p.c.geometry, details)
+
+            for prow in await conn.execute(placex_sql, bind_params):
+                result = nres.create_from_placex_row(prow, nres.SearchResult)
+                if result is not None:
+                    result.bbox = Bbox.from_wkb(prow.bbox)
+                break
+            else:
+                result = nres.create_from_postcode_row(row, nres.SearchResult)
+
+            assert result
+            if result.place_id not in details.excluded:
+                result.accuracy = row.accuracy
+                results.append(result)
+
+        return results

src/nominatim_api/search/geocoder.py

@@ -2,7 +2,7 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Public interface to the search code.
@@ -50,6 +50,9 @@ class ForwardGeocoder:
             self.query_analyzer = await make_query_analyzer(self.conn)
 
         query = await self.query_analyzer.analyze_query(phrases)
+        query.compute_direction_penalty()
+        log().var_dump('Query direction penalty',
+                       lambda: f"[{'LR' if query.dir_penalty < 0 else 'RL'}] {query.dir_penalty}")
 
         searches: List[AbstractSearch] = []
         if query.num_token_slots() > 0:
@@ -80,7 +83,7 @@ class ForwardGeocoder:
         min_ranking = searches[0].penalty + 2.0
         prev_penalty = 0.0
         for i, search in enumerate(searches):
-            if search.penalty > prev_penalty and (search.penalty > min_ranking or i > 20):
+            if search.penalty > prev_penalty and (search.penalty > min_ranking or i > 15):
                 break
             log().table_dump(f"{i + 1}. Search", _dump_searches([search], query))
             log().var_dump('Params', self.params)
@@ -115,17 +118,20 @@ class ForwardGeocoder:
         """ Remove badly matching results, sort by ranking and
             limit to the configured number of results.
         """
-        if results:
-            results.sort(key=lambda r: (r.ranking, 0 if r.bbox is None else -r.bbox.area))
-            min_rank = results[0].rank_search
-            min_ranking = results[0].ranking
-            results = SearchResults(r for r in results
-                                    if (r.ranking + 0.03 * (r.rank_search - min_rank)
-                                        < min_ranking + 0.5))
+        results.sort(key=lambda r: (r.ranking, 0 if r.bbox is None else -r.bbox.area))
 
-            results = SearchResults(results[:self.limit])
+        final = SearchResults()
+        min_rank = results[0].rank_search
+        min_ranking = results[0].ranking
 
-        return results
+        for r in results:
+            if r.ranking + 0.03 * (r.rank_search - min_rank) < min_ranking + 0.5:
+                final.append(r)
+                min_rank = min(r.rank_search, min_rank)
+            if len(final) == self.limit:
+                break
+
+        return final
 
     def rerank_by_query(self, query: QueryStruct, results: SearchResults) -> None:
         """ Adjust the accuracy of the localized result according to how well
@@ -150,17 +156,16 @@ class ForwardGeocoder:
             if not words:
                 continue
             for qword in qwords:
-                wdist = max(difflib.SequenceMatcher(a=qword, b=w).quick_ratio() for w in words)
-                if wdist < 0.5:
-                    distance += len(qword)
-                else:
-                    distance += (1.0 - wdist) * len(qword)
+                # only add distance penalty if there is no perfect match
+                if qword not in words:
+                    wdist = max(difflib.SequenceMatcher(a=qword, b=w).quick_ratio() for w in words)
+                    distance += len(qword) if wdist < 0.4 else 1
             # Compensate for the fact that country names do not get a
             # match penalty yet by the tokenizer.
             # Temporary hack that needs to be removed!
             if result.rank_address == 4:
                 distance *= 2
-            result.accuracy += distance * 0.4 / sum(len(w) for w in qwords)
+            result.accuracy += distance * 0.3 / sum(len(w) for w in qwords)
 
     async def lookup_pois(self, categories: List[Tuple[str, str]],
                           phrases: List[Phrase]) -> SearchResults:
@@ -208,9 +213,10 @@ class ForwardGeocoder:
             results = self.pre_filter_results(results)
             await add_result_details(self.conn, results, self.params)
             log().result_dump('Preliminary Results', ((r.accuracy, r) for r in results))
-            self.rerank_by_query(query, results)
-            log().result_dump('Results after reranking', ((r.accuracy, r) for r in results))
-            results = self.sort_and_cut_results(results)
+            if len(results) > 1:
+                self.rerank_by_query(query, results)
+                log().result_dump('Results after reranking', ((r.accuracy, r) for r in results))
+                results = self.sort_and_cut_results(results)
             log().result_dump('Final Results', ((r.accuracy, r) for r in results))
 
         return results
@@ -238,7 +244,7 @@ def _dump_searches(searches: List[AbstractSearch], query: QueryStruct,
         if not lk:
             return ''
 
-        return f"{lk.lookup_type}({lk.column}{tk(lk.tokens)})"
+        return f"{lk.lookup_type.__name__}({lk.column}{tk(lk.tokens)})"
 
     def fmt_cstr(c: Any) -> str:
         if not c:

src/nominatim_api/search/icu_tokenizer.py

@@ -2,13 +2,12 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Implementation of query analysis for the ICU tokenizer.
 """
 from typing import Tuple, Dict, List, Optional, Iterator, Any, cast
-from collections import defaultdict
 import dataclasses
 import difflib
 import re
@@ -25,62 +24,30 @@ from ..connection import SearchConnection
 from ..logging import log
 from . import query as qmod
 from ..query_preprocessing.config import QueryConfig
+from ..query_preprocessing.base import QueryProcessingFunc
 from .query_analyzer_factory import AbstractQueryAnalyzer
+from .postcode_parser import PostcodeParser
 
 
 DB_TO_TOKEN_TYPE = {
-    'W': qmod.TokenType.WORD,
-    'w': qmod.TokenType.PARTIAL,
-    'H': qmod.TokenType.HOUSENUMBER,
-    'P': qmod.TokenType.POSTCODE,
-    'C': qmod.TokenType.COUNTRY
+    'W': qmod.TOKEN_WORD,
+    'w': qmod.TOKEN_PARTIAL,
+    'H': qmod.TOKEN_HOUSENUMBER,
+    'P': qmod.TOKEN_POSTCODE,
+    'C': qmod.TOKEN_COUNTRY
 }
 
-PENALTY_IN_TOKEN_BREAK = {
-     qmod.BreakType.START: 0.5,
-     qmod.BreakType.END: 0.5,
-     qmod.BreakType.PHRASE: 0.5,
-     qmod.BreakType.SOFT_PHRASE: 0.5,
-     qmod.BreakType.WORD: 0.1,
-     qmod.BreakType.PART: 0.0,
-     qmod.BreakType.TOKEN: 0.0
+PENALTY_BREAK = {
+     qmod.BREAK_START: -0.5,
+     qmod.BREAK_END: -0.5,
+     qmod.BREAK_PHRASE: -0.5,
+     qmod.BREAK_SOFT_PHRASE: -0.5,
+     qmod.BREAK_WORD: 0.1,
+     qmod.BREAK_PART: 0.2,
+     qmod.BREAK_TOKEN: 0.4
 }
 
 
-@dataclasses.dataclass
-class QueryPart:
-    """ Normalized and transliterated form of a single term in the query.
-        When the term came out of a split during the transliteration,
-        the normalized string is the full word before transliteration.
-        The word number keeps track of the word before transliteration
-        and can be used to identify partial transliterated terms.
-        Penalty is the break penalty for the break following the token.
-    """
-    token: str
-    normalized: str
-    word_number: int
-    penalty: float
-
-
-QueryParts = List[QueryPart]
-WordDict = Dict[str, List[qmod.TokenRange]]
-
-
-def yield_words(terms: List[QueryPart], start: int) -> Iterator[Tuple[str, qmod.TokenRange]]:
-    """ Return all combinations of words in the terms list after the
-        given position.
-    """
-    total = len(terms)
-    for first in range(start, total):
-        word = terms[first].token
-        penalty = PENALTY_IN_TOKEN_BREAK[qmod.BreakType.WORD]
-        yield word, qmod.TokenRange(first, first + 1, penalty=penalty)
-        for last in range(first + 1, min(first + 20, total)):
-            word = ' '.join((word, terms[last].token))
-            penalty += terms[last - 1].penalty
-            yield word, qmod.TokenRange(first, last + 1, penalty=penalty)
-
-
 @dataclasses.dataclass
 class ICUToken(qmod.Token):
     """ Specialised token for ICU tokenizer.
@@ -111,13 +78,13 @@ class ICUToken(qmod.Token):
         self.penalty += (distance/len(self.lookup_word))
 
     @staticmethod
-    def from_db_row(row: SaRow, base_penalty: float = 0.0) -> 'ICUToken':
+    def from_db_row(row: SaRow) -> 'ICUToken':
         """ Create a ICUToken from the row of the word table.
         """
         count = 1 if row.info is None else row.info.get('count', 1)
         addr_count = 1 if row.info is None else row.info.get('addr_count', 1)
 
-        penalty = base_penalty
+        penalty = 0.0
         if row.type == 'w':
             penalty += 0.3
         elif row.type == 'W':
@@ -146,60 +113,51 @@ class ICUToken(qmod.Token):
                         addr_count=max(1, addr_count))
 
 
-class ICUQueryAnalyzer(AbstractQueryAnalyzer):
-    """ Converter for query strings into a tokenized query
-        using the tokens created by a ICU tokenizer.
-    """
-    def __init__(self, conn: SearchConnection) -> None:
-        self.conn = conn
+@dataclasses.dataclass
+class ICUAnalyzerConfig:
+    postcode_parser: PostcodeParser
+    normalizer: Transliterator
+    transliterator: Transliterator
+    preprocessors: List[QueryProcessingFunc]
 
-    async def setup(self) -> None:
-        """ Set up static data structures needed for the analysis.
-        """
-        async def _make_normalizer() -> Any:
-            rules = await self.conn.get_property('tokenizer_import_normalisation')
-            return Transliterator.createFromRules("normalization", rules)
+    @staticmethod
+    async def create(conn: SearchConnection) -> 'ICUAnalyzerConfig':
+        rules = await conn.get_property('tokenizer_import_normalisation')
+        normalizer = Transliterator.createFromRules("normalization", rules)
 
-        self.normalizer = await self.conn.get_cached_value('ICUTOK', 'normalizer',
-                                                           _make_normalizer)
+        rules = await conn.get_property('tokenizer_import_transliteration')
+        transliterator = Transliterator.createFromRules("transliteration", rules)
 
-        async def _make_transliterator() -> Any:
-            rules = await self.conn.get_property('tokenizer_import_transliteration')
-            return Transliterator.createFromRules("transliteration", rules)
-
-        self.transliterator = await self.conn.get_cached_value('ICUTOK', 'transliterator',
-                                                               _make_transliterator)
-
-        await self._setup_preprocessing()
-
-        if 'word' not in self.conn.t.meta.tables:
-            sa.Table('word', self.conn.t.meta,
-                     sa.Column('word_id', sa.Integer),
-                     sa.Column('word_token', sa.Text, nullable=False),
-                     sa.Column('type', sa.Text, nullable=False),
-                     sa.Column('word', sa.Text),
-                     sa.Column('info', Json))
-
-    async def _setup_preprocessing(self) -> None:
-        """ Load the rules for preprocessing and set up the handlers.
-        """
-
-        rules = self.conn.config.load_sub_configuration('icu_tokenizer.yaml',
-                                                        config='TOKENIZER_CONFIG')
-        preprocessing_rules = rules.get('query-preprocessing', [])
-
-        self.preprocessors = []
+        preprocessing_rules = conn.config.load_sub_configuration('icu_tokenizer.yaml',
+                                                                 config='TOKENIZER_CONFIG')\
+                                         .get('query-preprocessing', [])
 
+        preprocessors: List[QueryProcessingFunc] = []
         for func in preprocessing_rules:
             if 'step' not in func:
                 raise UsageError("Preprocessing rule is missing the 'step' attribute.")
             if not isinstance(func['step'], str):
                 raise UsageError("'step' attribute must be a simple string.")
 
-            module = self.conn.config.load_plugin_module(
+            module = conn.config.load_plugin_module(
                         func['step'], 'nominatim_api.query_preprocessing')
-            self.preprocessors.append(
-                module.create(QueryConfig(func).set_normalizer(self.normalizer)))
+            preprocessors.append(
+                module.create(QueryConfig(func).set_normalizer(normalizer)))
+
+        return ICUAnalyzerConfig(PostcodeParser(conn.config),
+                                 normalizer, transliterator, preprocessors)
+
+
+class ICUQueryAnalyzer(AbstractQueryAnalyzer):
+    """ Converter for query strings into a tokenized query
+        using the tokens created by a ICU tokenizer.
+    """
+    def __init__(self, conn: SearchConnection, config: ICUAnalyzerConfig) -> None:
+        self.conn = conn
+        self.postcode_parser = config.postcode_parser
+        self.normalizer = config.normalizer
+        self.transliterator = config.transliterator
+        self.preprocessors = config.preprocessors
 
     async def analyze_query(self, phrases: List[qmod.Phrase]) -> qmod.QueryStruct:
         """ Analyze the given list of phrases and return the
@@ -214,26 +172,38 @@ class ICUQueryAnalyzer(AbstractQueryAnalyzer):
         if not query.source:
             return query
 
-        parts, words = self.split_query(query)
-        log().var_dump('Transliterated query', lambda: _dump_transliterated(query, parts))
+        self.split_query(query)
+        log().var_dump('Transliterated query', lambda: query.get_transliterated_query())
+        words = query.extract_words()
 
         for row in await self.lookup_in_db(list(words.keys())):
             for trange in words[row.word_token]:
-                token = ICUToken.from_db_row(row, trange.penalty or 0.0)
+                # Create a new token for each position because the token
+                # penalty can vary depending on the position in the query.
+                # (See rerank_tokens() below.)
+                token = ICUToken.from_db_row(row)
                 if row.type == 'S':
                     if row.info['op'] in ('in', 'near'):
                         if trange.start == 0:
-                            query.add_token(trange, qmod.TokenType.NEAR_ITEM, token)
+                            query.add_token(trange, qmod.TOKEN_NEAR_ITEM, token)
                     else:
                         if trange.start == 0 and trange.end == query.num_token_slots():
-                            query.add_token(trange, qmod.TokenType.NEAR_ITEM, token)
+                            query.add_token(trange, qmod.TOKEN_NEAR_ITEM, token)
                         else:
-                            query.add_token(trange, qmod.TokenType.QUALIFIER, token)
+                            query.add_token(trange, qmod.TOKEN_QUALIFIER, token)
                 else:
                     query.add_token(trange, DB_TO_TOKEN_TYPE[row.type], token)
 
-        self.add_extra_tokens(query, parts)
-        self.rerank_tokens(query, parts)
+        self.add_extra_tokens(query)
+        for start, end, pc in self.postcode_parser.parse(query):
+            term = ' '.join(n.term_lookup for n in query.nodes[start + 1:end + 1])
+            query.add_token(qmod.TokenRange(start, end),
+                            qmod.TOKEN_POSTCODE,
+                            ICUToken(penalty=0.1, token=0, count=1, addr_count=1,
+                                     lookup_word=pc, word_token=term,
+                                     info=None))
+        self.rerank_tokens(query)
+        self.compute_break_penalties(query)
 
         log().table_dump('Word tokens', _dump_word_tokens(query))
 
@@ -244,19 +214,11 @@ class ICUQueryAnalyzer(AbstractQueryAnalyzer):
             standardized form search will work with. All information removed
             at this stage is inevitably lost.
         """
-        return cast(str, self.normalizer.transliterate(text))
+        return cast(str, self.normalizer.transliterate(text)).strip('-: ')
 
-    def split_query(self, query: qmod.QueryStruct) -> Tuple[QueryParts, WordDict]:
+    def split_query(self, query: qmod.QueryStruct) -> None:
         """ Transliterate the phrases and split them into tokens.
-
-            Returns the list of transliterated tokens together with their
-            normalized form and a dictionary of words for lookup together
-            with their position.
         """
-        parts: QueryParts = []
-        phrase_start = 0
-        words = defaultdict(list)
-        wordnr = 0
         for phrase in query.source:
             query.nodes[-1].ptype = phrase.ptype
             phrase_split = re.split('([ :-])', phrase.text)
@@ -271,78 +233,92 @@ class ICUQueryAnalyzer(AbstractQueryAnalyzer):
                 if trans:
                     for term in trans.split(' '):
                         if term:
-                            parts.append(QueryPart(term, word, wordnr,
-                                                   PENALTY_IN_TOKEN_BREAK[qmod.BreakType.TOKEN]))
-                            query.add_node(qmod.BreakType.TOKEN, phrase.ptype)
-                    query.nodes[-1].btype = qmod.BreakType(breakchar)
-                    parts[-1].penalty = PENALTY_IN_TOKEN_BREAK[qmod.BreakType(breakchar)]
-                wordnr += 1
+                            query.add_node(qmod.BREAK_TOKEN, phrase.ptype, term, word)
+                    query.nodes[-1].btype = breakchar
 
-            for word, wrange in yield_words(parts, phrase_start):
-                words[word].append(wrange)
-
-            phrase_start = len(parts)
-        query.nodes[-1].btype = qmod.BreakType.END
-
-        return parts, words
+        query.nodes[-1].btype = qmod.BREAK_END
 
     async def lookup_in_db(self, words: List[str]) -> 'sa.Result[Any]':
         """ Return the token information from the database for the
             given word tokens.
+
+            This function excludes postcode tokens
         """
         t = self.conn.t.meta.tables['word']
-        return await self.conn.execute(t.select().where(t.c.word_token.in_(words)))
+        return await self.conn.execute(t.select()
+                                        .where(t.c.word_token.in_(words))
+                                        .where(t.c.type != 'P'))
 
-    def add_extra_tokens(self, query: qmod.QueryStruct, parts: QueryParts) -> None:
+    def add_extra_tokens(self, query: qmod.QueryStruct) -> None:
         """ Add tokens to query that are not saved in the database.
         """
-        for part, node, i in zip(parts, query.nodes, range(1000)):
-            if len(part.token) <= 4 and part.token.isdigit()\
-               and not node.has_tokens(i+1, qmod.TokenType.HOUSENUMBER):
-                query.add_token(qmod.TokenRange(i, i+1), qmod.TokenType.HOUSENUMBER,
+        need_hnr = False
+        for i, node in enumerate(query.nodes):
+            is_full_token = node.btype not in (qmod.BREAK_TOKEN, qmod.BREAK_PART)
+            if need_hnr and is_full_token \
+                    and len(node.term_normalized) <= 4 and node.term_normalized.isdigit():
+                query.add_token(qmod.TokenRange(i-1, i), qmod.TOKEN_HOUSENUMBER,
                                 ICUToken(penalty=0.5, token=0,
-                                         count=1, addr_count=1, lookup_word=part.token,
-                                         word_token=part.token, info=None))
+                                         count=1, addr_count=1,
+                                         lookup_word=node.term_lookup,
+                                         word_token=node.term_lookup, info=None))
 
-    def rerank_tokens(self, query: qmod.QueryStruct, parts: QueryParts) -> None:
+            need_hnr = is_full_token and not node.has_tokens(i+1, qmod.TOKEN_HOUSENUMBER)
+
+    def rerank_tokens(self, query: qmod.QueryStruct) -> None:
         """ Add penalties to tokens that depend on presence of other token.
         """
-        for i, node, tlist in query.iter_token_lists():
-            if tlist.ttype == qmod.TokenType.POSTCODE:
-                for repl in node.starting:
-                    if repl.end == tlist.end and repl.ttype != qmod.TokenType.POSTCODE \
-                       and (repl.ttype != qmod.TokenType.HOUSENUMBER
-                            or len(tlist.tokens[0].lookup_word) > 4):
-                        repl.add_penalty(0.39)
-            elif (tlist.ttype == qmod.TokenType.HOUSENUMBER
-                  and len(tlist.tokens[0].lookup_word) <= 3):
-                if any(c.isdigit() for c in tlist.tokens[0].lookup_word):
-                    for repl in node.starting:
-                        if repl.end == tlist.end and repl.ttype != qmod.TokenType.HOUSENUMBER:
-                            repl.add_penalty(0.5 - tlist.tokens[0].penalty)
-            elif tlist.ttype not in (qmod.TokenType.COUNTRY, qmod.TokenType.PARTIAL):
-                norm = parts[i].normalized
-                for j in range(i + 1, tlist.end):
-                    if parts[j - 1].word_number != parts[j].word_number:
-                        norm += '  ' + parts[j].normalized
-                for token in tlist.tokens:
-                    cast(ICUToken, token).rematch(norm)
+        for start, end, tlist in query.iter_tokens_by_edge():
+            if len(tlist) > 1:
+                # If it looks like a Postcode, give preference.
+                if qmod.TOKEN_POSTCODE in tlist:
+                    for ttype, tokens in tlist.items():
+                        if ttype != qmod.TOKEN_POSTCODE and \
+                               (ttype != qmod.TOKEN_HOUSENUMBER or
+                                start + 1 > end or
+                                len(query.nodes[end].term_lookup) > 4):
+                            for token in tokens:
+                                token.penalty += 0.39
 
+                # If it looks like a simple housenumber, prefer that.
+                if qmod.TOKEN_HOUSENUMBER in tlist:
+                    hnr_lookup = tlist[qmod.TOKEN_HOUSENUMBER][0].lookup_word
+                    if len(hnr_lookup) <= 3 and any(c.isdigit() for c in hnr_lookup):
+                        penalty = 0.5 - tlist[qmod.TOKEN_HOUSENUMBER][0].penalty
+                        for ttype, tokens in tlist.items():
+                            if ttype != qmod.TOKEN_HOUSENUMBER:
+                                for token in tokens:
+                                    token.penalty += penalty
 
-def _dump_transliterated(query: qmod.QueryStruct, parts: QueryParts) -> str:
-    out = query.nodes[0].btype.value
-    for node, part in zip(query.nodes[1:], parts):
-        out += part.token + node.btype.value
-    return out
+            # rerank tokens against the normalized form
+            norm = ' '.join(n.term_normalized for n in query.nodes[start + 1:end + 1]
+                            if n.btype != qmod.BREAK_TOKEN)
+            if not norm:
+                # Can happen when the token only covers a partial term
+                norm = query.nodes[start + 1].term_normalized
+            for ttype, tokens in tlist.items():
+                if ttype != qmod.TOKEN_COUNTRY:
+                    for token in tokens:
+                        cast(ICUToken, token).rematch(norm)
+
+    def compute_break_penalties(self, query: qmod.QueryStruct) -> None:
+        """ Set the break penalties for the nodes in the query.
+        """
+        for node in query.nodes:
+            node.penalty = PENALTY_BREAK[node.btype]
 
 
 def _dump_word_tokens(query: qmod.QueryStruct) -> Iterator[List[Any]]:
-    yield ['type', 'token', 'word_token', 'lookup_word', 'penalty', 'count', 'info']
-    for node in query.nodes:
+    yield ['type', 'from', 'to', 'token', 'word_token', 'lookup_word', 'penalty', 'count', 'info']
+    for i, node in enumerate(query.nodes):
+        if node.partial is not None:
+            t = cast(ICUToken, node.partial)
+            yield [qmod.TOKEN_PARTIAL, str(i), str(i + 1), t.token,
+                   t.word_token, t.lookup_word, t.penalty, t.count, t.info]
         for tlist in node.starting:
             for token in tlist.tokens:
                 t = cast(ICUToken, token)
-                yield [tlist.ttype.name, t.token, t.word_token or '',
+                yield [tlist.ttype, str(i), str(tlist.end), t.token, t.word_token or '',
                        t.lookup_word or '', t.penalty, t.count, t.info]
 
 
@@ -350,7 +326,17 @@ async def create_query_analyzer(conn: SearchConnection) -> AbstractQueryAnalyzer
     """ Create and set up a new query analyzer for a database based
         on the ICU tokenizer.
     """
-    out = ICUQueryAnalyzer(conn)
-    await out.setup()
+    async def _get_config() -> ICUAnalyzerConfig:
+        if 'word' not in conn.t.meta.tables:
+            sa.Table('word', conn.t.meta,
+                     sa.Column('word_id', sa.Integer),
+                     sa.Column('word_token', sa.Text, nullable=False),
+                     sa.Column('type', sa.Text, nullable=False),
+                     sa.Column('word', sa.Text),
+                     sa.Column('info', Json))
 
-    return out
+        return await ICUAnalyzerConfig.create(conn)
+
+    config = await conn.get_cached_value('ICUTOK', 'config', _get_config)
+
+    return ICUQueryAnalyzer(conn, config)

src/nominatim_api/search/postcode_parser.py

@@ -0,0 +1,104 @@
+# SPDX-License-Identifier: GPL-3.0-or-later
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2025 by the Nominatim developer community.
+# For a full list of authors see the git log.
+"""
+Handling of arbitrary postcode tokens in tokenized query string.
+"""
+from typing import Tuple, Set, Dict, List
+import re
+from collections import defaultdict
+
+import yaml
+
+from ..config import Configuration
+from . import query as qmod
+
+
+class PostcodeParser:
+    """ Pattern-based parser for postcodes in tokenized queries.
+
+        The postcode patterns are read from the country configuration.
+        The parser does currently not return country restrictions.
+    """
+
+    def __init__(self, config: Configuration) -> None:
+        # skip over includes here to avoid loading the complete country name data
+        yaml.add_constructor('!include', lambda loader, node: [],
+                             Loader=yaml.SafeLoader)
+        cdata = yaml.safe_load(config.find_config_file('country_settings.yaml')
+                                     .read_text(encoding='utf-8'))
+
+        unique_patterns: Dict[str, Dict[str, List[str]]] = {}
+        for cc, data in cdata.items():
+            if data.get('postcode'):
+                pat = data['postcode']['pattern'].replace('d', '[0-9]').replace('l', '[A-Z]')
+                out = data['postcode'].get('output')
+                if pat not in unique_patterns:
+                    unique_patterns[pat] = defaultdict(list)
+                unique_patterns[pat][out].append(cc.upper())
+
+        self.global_pattern = re.compile(
+                '(?:(?P<cc>[A-Z][A-Z])(?P<space>[ -]?))?(?P<pc>(?:(?:'
+                + ')|(?:'.join(unique_patterns) + '))[:, >].*)')
+
+        self.local_patterns = [(re.compile(f"{pat}[:, >]"), list(info.items()))
+                               for pat, info in unique_patterns.items()]
+
+    def parse(self, query: qmod.QueryStruct) -> Set[Tuple[int, int, str]]:
+        """ Parse postcodes in the given list of query tokens taking into
+            account the list of breaks from the nodes.
+
+            The result is a sequence of tuples with
+            [start node id, end node id, postcode token]
+        """
+        nodes = query.nodes
+        outcodes: Set[Tuple[int, int, str]] = set()
+
+        terms = [n.term_normalized.upper() + n.btype for n in nodes]
+        for i in range(query.num_token_slots()):
+            if nodes[i].btype in '<,: ' and nodes[i + 1].btype != '`' \
+                    and (i == 0 or nodes[i - 1].ptype != qmod.PHRASE_POSTCODE):
+                if nodes[i].ptype == qmod.PHRASE_ANY:
+                    word = terms[i + 1]
+                    if word[-1] in ' -' and nodes[i + 2].btype != '`' \
+                            and nodes[i + 1].ptype == qmod.PHRASE_ANY:
+                        word += terms[i + 2]
+                        if word[-1] in ' -' and nodes[i + 3].btype != '`' \
+                                and nodes[i + 2].ptype == qmod.PHRASE_ANY:
+                            word += terms[i + 3]
+
+                    self._match_word(word, i, False, outcodes)
+                elif nodes[i].ptype == qmod.PHRASE_POSTCODE:
+                    word = terms[i + 1]
+                    for j in range(i + 1, query.num_token_slots()):
+                        if nodes[j].ptype != qmod.PHRASE_POSTCODE:
+                            break
+                        word += terms[j + 1]
+
+                    self._match_word(word, i, True, outcodes)
+
+        return outcodes
+
+    def _match_word(self, word: str, pos: int, fullmatch: bool,
+                    outcodes: Set[Tuple[int, int, str]]) -> None:
+        # Use global pattern to check for presence of any postcode.
+        m = self.global_pattern.fullmatch(word)
+        if m:
+            # If there was a match, check against each pattern separately
+            # because multiple patterns might be machting at the end.
+            cc = m.group('cc')
+            pc_word = m.group('pc')
+            cc_spaces = len(m.group('space') or '')
+            for pattern, info in self.local_patterns:
+                lm = pattern.fullmatch(pc_word) if fullmatch else pattern.match(pc_word)
+                if lm:
+                    trange = (pos, pos + cc_spaces + sum(c in ' ,-:>' for c in lm.group(0)))
+                    for out, out_ccs in info:
+                        if cc is None or cc in out_ccs:
+                            if out:
+                                outcodes.add((*trange, lm.expand(out)))
+                            else:
+                                outcodes.add((*trange, lm.group(0)[:-1]))

src/nominatim_api/search/query.py

@@ -2,99 +2,111 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Datastructures for a tokenized query.
 """
-from typing import List, Tuple, Optional, Iterator
+from typing import Dict, List, Tuple, Optional, Iterator
 from abc import ABC, abstractmethod
+from collections import defaultdict
 import dataclasses
-import enum
+
+# Precomputed denominator for the computation of the linear regression slope
+# used to determine the query direction.
+# The x value for the regression computation will be the position of the
+# token in the query. Thus we know the x values will be [0, query length).
+# As the denominator only depends on the x values, we can pre-compute here
+# the denominatior to use for a given query length.
+# Note that query length of two or less is special cased and will not use
+# the values from this array. Thus it is not a problem that they are 0.
+LINFAC = [i * (sum(si * si for si in range(i)) - (i - 1) * i * (i - 1) / 4)
+          for i in range(50)]
 
 
-class BreakType(enum.Enum):
-    """ Type of break between tokens.
-    """
-    START = '<'
-    """ Begin of the query. """
-    END = '>'
-    """ End of the query. """
-    PHRASE = ','
-    """ Hard break between two phrases. Address parts cannot cross hard
-        phrase boundaries."""
-    SOFT_PHRASE = ':'
-    """ Likely break between two phrases. Address parts should not cross soft
-        phrase boundaries. Soft breaks can be inserted by a preprocessor
-        that is analysing the input string.
-    """
-    WORD = ' '
-    """ Break between words. """
-    PART = '-'
-    """ Break inside a word, for example a hyphen or apostrophe. """
-    TOKEN = '`'
-    """ Break created as a result of tokenization.
-        This may happen in languages without spaces between words.
+BreakType = str
+""" Type of break between tokens.
+"""
+BREAK_START = '<'
+""" Begin of the query. """
+BREAK_END = '>'
+""" End of the query. """
+BREAK_PHRASE = ','
+""" Hard break between two phrases. Address parts cannot cross hard
+    phrase boundaries."""
+BREAK_SOFT_PHRASE = ':'
+""" Likely break between two phrases. Address parts should not cross soft
+    phrase boundaries. Soft breaks can be inserted by a preprocessor
+    that is analysing the input string.
+"""
+BREAK_WORD = ' '
+""" Break between words. """
+BREAK_PART = '-'
+""" Break inside a word, for example a hyphen or apostrophe. """
+BREAK_TOKEN = '`'
+""" Break created as a result of tokenization.
+    This may happen in languages without spaces between words.
+"""
+
+
+TokenType = str
+""" Type of token.
+"""
+TOKEN_WORD = 'W'
+""" Full name of a place. """
+TOKEN_PARTIAL = 'w'
+""" Word term without breaks, does not necessarily represent a full name. """
+TOKEN_HOUSENUMBER = 'H'
+""" Housenumber term. """
+TOKEN_POSTCODE = 'P'
+""" Postal code term. """
+TOKEN_COUNTRY = 'C'
+""" Country name or reference. """
+TOKEN_QUALIFIER = 'Q'
+""" Special term used together with name (e.g. _Hotel_ Bellevue). """
+TOKEN_NEAR_ITEM = 'N'
+""" Special term used as searchable object(e.g. supermarket in ...). """
+
+
+PhraseType = int
+""" Designation of a phrase.
+"""
+PHRASE_ANY = 0
+""" No specific designation (i.e. source is free-form query). """
+PHRASE_AMENITY = 1
+""" Contains name or type of a POI. """
+PHRASE_STREET = 2
+""" Contains a street name optionally with a housenumber. """
+PHRASE_CITY = 3
+""" Contains the postal city. """
+PHRASE_COUNTY = 4
+""" Contains the equivalent of a county. """
+PHRASE_STATE = 5
+""" Contains a state or province. """
+PHRASE_POSTCODE = 6
+""" Contains a postal code. """
+PHRASE_COUNTRY = 7
+""" Contains the country name or code. """
+
+
+def _phrase_compatible_with(ptype: PhraseType, ttype: TokenType,
+                            is_full_phrase: bool) -> bool:
+    """ Check if the given token type can be used with the phrase type.
     """
+    if ptype == PHRASE_ANY:
+        return not is_full_phrase or ttype != TOKEN_QUALIFIER
+    if ptype == PHRASE_AMENITY:
+        return ttype in (TOKEN_WORD, TOKEN_PARTIAL)\
+               or (is_full_phrase and ttype == TOKEN_NEAR_ITEM)\
+               or (not is_full_phrase and ttype == TOKEN_QUALIFIER)
+    if ptype == PHRASE_STREET:
+        return ttype in (TOKEN_WORD, TOKEN_PARTIAL, TOKEN_HOUSENUMBER)
+    if ptype == PHRASE_POSTCODE:
+        return ttype == TOKEN_POSTCODE
+    if ptype == PHRASE_COUNTRY:
+        return ttype == TOKEN_COUNTRY
 
-
-class TokenType(enum.Enum):
-    """ Type of token.
-    """
-    WORD = enum.auto()
-    """ Full name of a place. """
-    PARTIAL = enum.auto()
-    """ Word term without breaks, does not necessarily represent a full name. """
-    HOUSENUMBER = enum.auto()
-    """ Housenumber term. """
-    POSTCODE = enum.auto()
-    """ Postal code term. """
-    COUNTRY = enum.auto()
-    """ Country name or reference. """
-    QUALIFIER = enum.auto()
-    """ Special term used together with name (e.g. _Hotel_ Bellevue). """
-    NEAR_ITEM = enum.auto()
-    """ Special term used as searchable object(e.g. supermarket in ...). """
-
-
-class PhraseType(enum.Enum):
-    """ Designation of a phrase.
-    """
-    NONE = 0
-    """ No specific designation (i.e. source is free-form query). """
-    AMENITY = enum.auto()
-    """ Contains name or type of a POI. """
-    STREET = enum.auto()
-    """ Contains a street name optionally with a housenumber. """
-    CITY = enum.auto()
-    """ Contains the postal city. """
-    COUNTY = enum.auto()
-    """ Contains the equivalent of a county. """
-    STATE = enum.auto()
-    """ Contains a state or province. """
-    POSTCODE = enum.auto()
-    """ Contains a postal code. """
-    COUNTRY = enum.auto()
-    """ Contains the country name or code. """
-
-    def compatible_with(self, ttype: TokenType,
-                        is_full_phrase: bool) -> bool:
-        """ Check if the given token type can be used with the phrase type.
-        """
-        if self == PhraseType.NONE:
-            return not is_full_phrase or ttype != TokenType.QUALIFIER
-        if self == PhraseType.AMENITY:
-            return ttype in (TokenType.WORD, TokenType.PARTIAL)\
-                   or (is_full_phrase and ttype == TokenType.NEAR_ITEM)\
-                   or (not is_full_phrase and ttype == TokenType.QUALIFIER)
-        if self == PhraseType.STREET:
-            return ttype in (TokenType.WORD, TokenType.PARTIAL, TokenType.HOUSENUMBER)
-        if self == PhraseType.POSTCODE:
-            return ttype == TokenType.POSTCODE
-        if self == PhraseType.COUNTRY:
-            return ttype == TokenType.COUNTRY
-
-        return ttype in (TokenType.WORD, TokenType.PARTIAL)
+    return ttype in (TOKEN_WORD, TOKEN_PARTIAL)
 
 
 @dataclasses.dataclass
@@ -122,7 +134,6 @@ class TokenRange:
     """
     start: int
     end: int
-    penalty: Optional[float] = None
 
     def __lt__(self, other: 'TokenRange') -> bool:
         return self.end <= other.start
@@ -171,10 +182,58 @@ class TokenList:
 @dataclasses.dataclass
 class QueryNode:
     """ A node of the query representing a break between terms.
+
+        The node also contains information on the source term
+        ending at the node. The tokens are created from this information.
     """
     btype: BreakType
     ptype: PhraseType
+
+    penalty: float
+    """ Penalty for having a word break at this position. The penalty
+        may be negative, when a word break is more likely than continuing
+        the word after the node.
+    """
+    term_lookup: str
+    """ Transliterated term ending at this node.
+    """
+    term_normalized: str
+    """ Normalised form of term ending at this node.
+        When the token resulted from a split during transliteration,
+        then this string contains the complete source term.
+    """
+
     starting: List[TokenList] = dataclasses.field(default_factory=list)
+    """ List of all full tokens starting at this node.
+    """
+    partial: Optional[Token] = None
+    """ Base token going to the next node.
+        May be None when the query has parts for which no words are known.
+        Note that the query may still be parsable when there are other
+        types of tokens spanning over the gap.
+    """
+
+    @property
+    def word_break_penalty(self) -> float:
+        """ Penalty to apply when a words ends at this node.
+        """
+        return max(0, self.penalty)
+
+    @property
+    def word_continuation_penalty(self) -> float:
+        """ Penalty to apply when a word continues over this node
+            (i.e. is a multi-term word).
+        """
+        return max(0, -self.penalty)
+
+    def name_address_ratio(self) -> float:
+        """ Return the propability that the partial token belonging to
+            this node forms part of a name (as opposed of part of the address).
+        """
+        if self.partial is None:
+            return 0.5
+
+        return self.partial.count / (self.partial.count + self.partial.addr_count)
 
     def has_tokens(self, end: int, *ttypes: TokenType) -> bool:
         """ Check if there are tokens of the given types ending at the
@@ -211,26 +270,36 @@ class QueryStruct:
         need to be direct neighbours. Thus the query is represented as a
         directed acyclic graph.
 
+        A query also has a direction penalty 'dir_penalty'. This describes
+        the likelyhood if the query should be read from left-to-right or
+        vice versa. A negative 'dir_penalty' should be read as a penalty on
+        right-to-left reading, while a positive value represents a penalty
+        for left-to-right reading. The default value is 0, which is equivalent
+        to having no information about the reading.
+
         When created, a query contains a single node: the start of the
         query. Further nodes can be added by appending to 'nodes'.
     """
 
     def __init__(self, source: List[Phrase]) -> None:
         self.source = source
+        self.dir_penalty = 0.0
         self.nodes: List[QueryNode] = \
-            [QueryNode(BreakType.START, source[0].ptype if source else PhraseType.NONE)]
+            [QueryNode(BREAK_START, source[0].ptype if source else PHRASE_ANY,
+                       0.0, '', '')]
 
     def num_token_slots(self) -> int:
         """ Return the length of the query in vertice steps.
         """
         return len(self.nodes) - 1
 
-    def add_node(self, btype: BreakType, ptype: PhraseType) -> None:
+    def add_node(self, btype: BreakType, ptype: PhraseType,
+                 term_lookup: str = '', term_normalized: str = '') -> None:
         """ Append a new break node with the given break type.
             The phrase type denotes the type for any tokens starting
             at the node.
         """
-        self.nodes.append(QueryNode(btype, ptype))
+        self.nodes.append(QueryNode(btype, ptype, 0.0, term_lookup, term_normalized))
 
     def add_token(self, trange: TokenRange, ttype: TokenType, token: Token) -> None:
         """ Add a token to the query. 'start' and 'end' are the indexes of the
@@ -243,37 +312,70 @@ class QueryStruct:
             be added to, then the token is silently dropped.
         """
         snode = self.nodes[trange.start]
-        full_phrase = snode.btype in (BreakType.START, BreakType.PHRASE)\
-            and self.nodes[trange.end].btype in (BreakType.PHRASE, BreakType.END)
-        if snode.ptype.compatible_with(ttype, full_phrase):
-            tlist = snode.get_tokens(trange.end, ttype)
-            if tlist is None:
-                snode.starting.append(TokenList(trange.end, ttype, [token]))
-            else:
-                tlist.append(token)
+        if ttype == TOKEN_PARTIAL:
+            assert snode.partial is None
+            if _phrase_compatible_with(snode.ptype, TOKEN_PARTIAL, False):
+                snode.partial = token
+        else:
+            full_phrase = snode.btype in (BREAK_START, BREAK_PHRASE)\
+                and self.nodes[trange.end].btype in (BREAK_PHRASE, BREAK_END)
+            if _phrase_compatible_with(snode.ptype, ttype, full_phrase):
+                tlist = snode.get_tokens(trange.end, ttype)
+                if tlist is None:
+                    snode.starting.append(TokenList(trange.end, ttype, [token]))
+                else:
+                    tlist.append(token)
+
+    def compute_direction_penalty(self) -> None:
+        """ Recompute the direction probability from the partial tokens
+            of each node.
+        """
+        n = len(self.nodes) - 1
+        if n <= 1 or n >= 50:
+            self.dir_penalty = 0
+        elif n == 2:
+            self.dir_penalty = (self.nodes[1].name_address_ratio()
+                                - self.nodes[0].name_address_ratio()) / 3
+        else:
+            ratios = [n.name_address_ratio() for n in self.nodes[:-1]]
+            self.dir_penalty = (n * sum(i * r for i, r in enumerate(ratios))
+                                - sum(ratios) * n * (n - 1) / 2) / LINFAC[n]
 
     def get_tokens(self, trange: TokenRange, ttype: TokenType) -> List[Token]:
         """ Get the list of tokens of a given type, spanning the given
             nodes. The nodes must exist. If no tokens exist, an
             empty list is returned.
+
+            Cannot be used to get the partial token.
         """
+        assert ttype != TOKEN_PARTIAL
         return self.nodes[trange.start].get_tokens(trange.end, ttype) or []
 
-    def get_partials_list(self, trange: TokenRange) -> List[Token]:
-        """ Create a list of partial tokens between the given nodes.
-            The list is composed of the first token of type PARTIAL
-            going to the subsequent node. Such PARTIAL tokens are
-            assumed to exist.
+    def get_in_word_penalty(self, trange: TokenRange) -> float:
+        """ Gets the sum of penalties for all token transitions
+            within the given range.
         """
-        return [next(iter(self.get_tokens(TokenRange(i, i+1), TokenType.PARTIAL)))
-                for i in range(trange.start, trange.end)]
+        return sum(n.word_continuation_penalty
+                   for n in self.nodes[trange.start + 1:trange.end])
 
-    def iter_token_lists(self) -> Iterator[Tuple[int, QueryNode, TokenList]]:
-        """ Iterator over all token lists in the query.
+    def iter_partials(self, trange: TokenRange) -> Iterator[Token]:
+        """ Iterate over the partial tokens between the given nodes.
+            Missing partials are ignored.
+        """
+        return (n.partial for n in self.nodes[trange.start:trange.end] if n.partial is not None)
+
+    def iter_tokens_by_edge(self) -> Iterator[Tuple[int, int, Dict[TokenType, List[Token]]]]:
+        """ Iterator over all tokens except partial ones grouped by edge.
+
+            Returns the start and end node indexes and a dictionary
+            of list of tokens by token type.
         """
         for i, node in enumerate(self.nodes):
+            by_end: Dict[int, Dict[TokenType, List[Token]]] = defaultdict(dict)
             for tlist in node.starting:
-                yield i, node, tlist
+                by_end[tlist.end][tlist.ttype] = tlist.tokens
+            for end, endlist in by_end.items():
+                yield i, end, endlist
 
     def find_lookup_word_by_id(self, token: int) -> str:
         """ Find the first token with the given token ID and return
@@ -282,8 +384,45 @@ class QueryStruct:
             debugging.
         """
         for node in self.nodes:
+            if node.partial is not None and node.partial.token == token:
+                return f"[P]{node.partial.lookup_word}"
             for tlist in node.starting:
                 for t in tlist.tokens:
                     if t.token == token:
-                        return f"[{tlist.ttype.name[0]}]{t.lookup_word}"
+                        return f"[{tlist.ttype}]{t.lookup_word}"
         return 'None'
+
+    def get_transliterated_query(self) -> str:
+        """ Return a string representation of the transliterated query
+            with the character representation of the different break types.
+
+            For debugging purposes only.
+        """
+        return ''.join(''.join((n.term_lookup, n.btype)) for n in self.nodes)
+
+    def extract_words(self, start: int = 0,
+                      endpos: Optional[int] = None) -> Dict[str, List[TokenRange]]:
+        """ Add all combinations of words that can be formed from the terms
+            between the given start and endnode. The terms are joined with
+            spaces for each break. Words can never go across a BREAK_PHRASE.
+
+            The functions returns a dictionary of possible words with their
+            position within the query.
+        """
+        if endpos is None:
+            endpos = len(self.nodes)
+
+        words: Dict[str, List[TokenRange]] = defaultdict(list)
+
+        for first, first_node in enumerate(self.nodes[start + 1:endpos], start):
+            word = first_node.term_lookup
+            words[word].append(TokenRange(first, first + 1))
+            if first_node.btype != BREAK_PHRASE:
+                max_last = min(first + 20, endpos)
+                for last, last_node in enumerate(self.nodes[first + 2:max_last], first + 2):
+                    word = ' '.join((word, last_node.term_lookup))
+                    words[word].append(TokenRange(first, last))
+                    if last_node.btype == BREAK_PHRASE:
+                        break
+
+        return words

src/nominatim_api/search/token_assignment.py

@@ -23,16 +23,6 @@ class TypedRange:
     trange: qmod.TokenRange
 
 
-PENALTY_TOKENCHANGE = {
-    qmod.BreakType.START: 0.0,
-    qmod.BreakType.END: 0.0,
-    qmod.BreakType.PHRASE: 0.0,
-    qmod.BreakType.SOFT_PHRASE: 0.0,
-    qmod.BreakType.WORD: 0.1,
-    qmod.BreakType.PART: 0.2,
-    qmod.BreakType.TOKEN: 0.4
-}
-
 TypedRangeSeq = List[TypedRange]
 
 
@@ -56,17 +46,17 @@ class TokenAssignment:
         """
         out = TokenAssignment()
         for token in ranges:
-            if token.ttype == qmod.TokenType.PARTIAL:
+            if token.ttype == qmod.TOKEN_PARTIAL:
                 out.address.append(token.trange)
-            elif token.ttype == qmod.TokenType.HOUSENUMBER:
+            elif token.ttype == qmod.TOKEN_HOUSENUMBER:
                 out.housenumber = token.trange
-            elif token.ttype == qmod.TokenType.POSTCODE:
+            elif token.ttype == qmod.TOKEN_POSTCODE:
                 out.postcode = token.trange
-            elif token.ttype == qmod.TokenType.COUNTRY:
+            elif token.ttype == qmod.TOKEN_COUNTRY:
                 out.country = token.trange
-            elif token.ttype == qmod.TokenType.NEAR_ITEM:
+            elif token.ttype == qmod.TOKEN_NEAR_ITEM:
                 out.near_item = token.trange
-            elif token.ttype == qmod.TokenType.QUALIFIER:
+            elif token.ttype == qmod.TOKEN_QUALIFIER:
                 out.qualifier = token.trange
         return out
 
@@ -84,7 +74,7 @@ class _TokenSequence:
         self.penalty = penalty
 
     def __str__(self) -> str:
-        seq = ''.join(f'[{r.trange.start} - {r.trange.end}: {r.ttype.name}]' for r in self.seq)
+        seq = ''.join(f'[{r.trange.start} - {r.trange.end}: {r.ttype}]' for r in self.seq)
         return f'{seq} (dir: {self.direction}, penalty: {self.penalty})'
 
     @property
@@ -105,7 +95,7 @@ class _TokenSequence:
         """
         # Country and category must be the final term for left-to-right
         return len(self.seq) > 1 and \
-            self.seq[-1].ttype in (qmod.TokenType.COUNTRY, qmod.TokenType.NEAR_ITEM)
+            self.seq[-1].ttype in (qmod.TOKEN_COUNTRY, qmod.TOKEN_NEAR_ITEM)
 
     def appendable(self, ttype: qmod.TokenType) -> Optional[int]:
         """ Check if the give token type is appendable to the existing sequence.
@@ -114,23 +104,23 @@ class _TokenSequence:
             new direction of the sequence after adding such a type. The
             token is not added.
         """
-        if ttype == qmod.TokenType.WORD:
+        if ttype == qmod.TOKEN_WORD:
             return None
 
         if not self.seq:
             # Append unconditionally to the empty list
-            if ttype == qmod.TokenType.COUNTRY:
+            if ttype == qmod.TOKEN_COUNTRY:
                 return -1
-            if ttype in (qmod.TokenType.HOUSENUMBER, qmod.TokenType.QUALIFIER):
+            if ttype in (qmod.TOKEN_HOUSENUMBER, qmod.TOKEN_QUALIFIER):
                 return 1
             return self.direction
 
         # Name tokens are always acceptable and don't change direction
-        if ttype == qmod.TokenType.PARTIAL:
+        if ttype == qmod.TOKEN_PARTIAL:
             # qualifiers cannot appear in the middle of the query. They need
             # to be near the next phrase.
             if self.direction == -1 \
-               and any(t.ttype == qmod.TokenType.QUALIFIER for t in self.seq[:-1]):
+               and any(t.ttype == qmod.TOKEN_QUALIFIER for t in self.seq[:-1]):
                 return None
             return self.direction
 
@@ -138,61 +128,61 @@ class _TokenSequence:
         if self.has_types(ttype):
             return None
 
-        if ttype == qmod.TokenType.HOUSENUMBER:
+        if ttype == qmod.TOKEN_HOUSENUMBER:
             if self.direction == 1:
-                if len(self.seq) == 1 and self.seq[0].ttype == qmod.TokenType.QUALIFIER:
+                if len(self.seq) == 1 and self.seq[0].ttype == qmod.TOKEN_QUALIFIER:
                     return None
                 if len(self.seq) > 2 \
-                   or self.has_types(qmod.TokenType.POSTCODE, qmod.TokenType.COUNTRY):
+                   or self.has_types(qmod.TOKEN_POSTCODE, qmod.TOKEN_COUNTRY):
                     return None  # direction left-to-right: housenumber must come before anything
             elif (self.direction == -1
-                  or self.has_types(qmod.TokenType.POSTCODE, qmod.TokenType.COUNTRY)):
+                  or self.has_types(qmod.TOKEN_POSTCODE, qmod.TOKEN_COUNTRY)):
                 return -1  # force direction right-to-left if after other terms
 
             return self.direction
 
-        if ttype == qmod.TokenType.POSTCODE:
+        if ttype == qmod.TOKEN_POSTCODE:
             if self.direction == -1:
-                if self.has_types(qmod.TokenType.HOUSENUMBER, qmod.TokenType.QUALIFIER):
+                if self.has_types(qmod.TOKEN_HOUSENUMBER, qmod.TOKEN_QUALIFIER):
                     return None
                 return -1
             if self.direction == 1:
-                return None if self.has_types(qmod.TokenType.COUNTRY) else 1
-            if self.has_types(qmod.TokenType.HOUSENUMBER, qmod.TokenType.QUALIFIER):
+                return None if self.has_types(qmod.TOKEN_COUNTRY) else 1
+            if self.has_types(qmod.TOKEN_HOUSENUMBER, qmod.TOKEN_QUALIFIER):
                 return 1
             return self.direction
 
-        if ttype == qmod.TokenType.COUNTRY:
+        if ttype == qmod.TOKEN_COUNTRY:
             return None if self.direction == -1 else 1
 
-        if ttype == qmod.TokenType.NEAR_ITEM:
+        if ttype == qmod.TOKEN_NEAR_ITEM:
             return self.direction
 
-        if ttype == qmod.TokenType.QUALIFIER:
+        if ttype == qmod.TOKEN_QUALIFIER:
             if self.direction == 1:
                 if (len(self.seq) == 1
-                    and self.seq[0].ttype in (qmod.TokenType.PARTIAL, qmod.TokenType.NEAR_ITEM)) \
+                    and self.seq[0].ttype in (qmod.TOKEN_PARTIAL, qmod.TOKEN_NEAR_ITEM)) \
                    or (len(self.seq) == 2
-                       and self.seq[0].ttype == qmod.TokenType.NEAR_ITEM
-                       and self.seq[1].ttype == qmod.TokenType.PARTIAL):
+                       and self.seq[0].ttype == qmod.TOKEN_NEAR_ITEM
+                       and self.seq[1].ttype == qmod.TOKEN_PARTIAL):
                     return 1
                 return None
             if self.direction == -1:
                 return -1
 
-            tempseq = self.seq[1:] if self.seq[0].ttype == qmod.TokenType.NEAR_ITEM else self.seq
+            tempseq = self.seq[1:] if self.seq[0].ttype == qmod.TOKEN_NEAR_ITEM else self.seq
             if len(tempseq) == 0:
                 return 1
-            if len(tempseq) == 1 and self.seq[0].ttype == qmod.TokenType.HOUSENUMBER:
+            if len(tempseq) == 1 and self.seq[0].ttype == qmod.TOKEN_HOUSENUMBER:
                 return None
-            if len(tempseq) > 1 or self.has_types(qmod.TokenType.POSTCODE, qmod.TokenType.COUNTRY):
+            if len(tempseq) > 1 or self.has_types(qmod.TOKEN_POSTCODE, qmod.TOKEN_COUNTRY):
                 return -1
             return 0
 
         return None
 
     def advance(self, ttype: qmod.TokenType, end_pos: int,
-                btype: qmod.BreakType) -> Optional['_TokenSequence']:
+                force_break: bool, break_penalty: float) -> Optional['_TokenSequence']:
         """ Return a new token sequence state with the given token type
             extended.
         """
@@ -205,7 +195,7 @@ class _TokenSequence:
             new_penalty = 0.0
         else:
             last = self.seq[-1]
-            if btype != qmod.BreakType.PHRASE and last.ttype == ttype:
+            if not force_break and last.ttype == ttype:
                 # extend the existing range
                 newseq = self.seq[:-1] + [TypedRange(ttype, last.trange.replace_end(end_pos))]
                 new_penalty = 0.0
@@ -213,7 +203,7 @@ class _TokenSequence:
                 # start a new range
                 newseq = list(self.seq) + [TypedRange(ttype,
                                                       qmod.TokenRange(last.trange.end, end_pos))]
-                new_penalty = PENALTY_TOKENCHANGE[btype]
+                new_penalty = break_penalty
 
         return _TokenSequence(newseq, newdir, self.penalty + new_penalty)
 
@@ -240,18 +230,18 @@ class _TokenSequence:
         # housenumbers may not be further than 2 words from the beginning.
         # If there are two words in front, give it a penalty.
         hnrpos = next((i for i, tr in enumerate(self.seq)
-                       if tr.ttype == qmod.TokenType.HOUSENUMBER),
+                       if tr.ttype == qmod.TOKEN_HOUSENUMBER),
                       None)
         if hnrpos is not None:
             if self.direction != -1:
-                priors = sum(1 for t in self.seq[:hnrpos] if t.ttype == qmod.TokenType.PARTIAL)
+                priors = sum(1 for t in self.seq[:hnrpos] if t.ttype == qmod.TOKEN_PARTIAL)
                 if not self._adapt_penalty_from_priors(priors, -1):
                     return False
             if self.direction != 1:
-                priors = sum(1 for t in self.seq[hnrpos+1:] if t.ttype == qmod.TokenType.PARTIAL)
+                priors = sum(1 for t in self.seq[hnrpos+1:] if t.ttype == qmod.TOKEN_PARTIAL)
                 if not self._adapt_penalty_from_priors(priors, 1):
                     return False
-            if any(t.ttype == qmod.TokenType.NEAR_ITEM for t in self.seq):
+            if any(t.ttype == qmod.TOKEN_NEAR_ITEM for t in self.seq):
                 self.penalty += 1.0
 
         return True
@@ -269,10 +259,9 @@ class _TokenSequence:
             # <address>,<postcode> should give preference to address search
             if base.postcode.start == 0:
                 penalty = self.penalty
-                self.direction = -1  # name searches are only possible backwards
             else:
                 penalty = self.penalty + 0.1
-                self.direction = 1  # name searches are only possible forwards
+            penalty += 0.1 * max(0, len(base.address) - 1)
             yield dataclasses.replace(base, penalty=penalty)
 
     def _get_assignments_address_forward(self, base: TokenAssignment,
@@ -282,8 +271,17 @@ class _TokenSequence:
         """
         first = base.address[0]
 
+        # The postcode must come after the name.
+        if base.postcode and base.postcode < first:
+            log().var_dump('skip forward', (base.postcode, first))
+            return
+
+        penalty = self.penalty
+        if not base.country and self.direction == 1 and query.dir_penalty > 0:
+            penalty += query.dir_penalty
+
         log().comment('first word = name')
-        yield dataclasses.replace(base, penalty=self.penalty,
+        yield dataclasses.replace(base, penalty=penalty,
                                   name=first, address=base.address[1:])
 
         # To paraphrase:
@@ -293,22 +291,23 @@ class _TokenSequence:
         #  * the containing phrase is strictly typed
         if (base.housenumber and first.end < base.housenumber.start)\
            or (base.qualifier and base.qualifier > first)\
-           or (query.nodes[first.start].ptype != qmod.PhraseType.NONE):
+           or (query.nodes[first.start].ptype != qmod.PHRASE_ANY):
             return
 
-        penalty = self.penalty
-
         # Penalty for:
         #  * <name>, <street>, <housenumber> , ...
         #  * queries that are comma-separated
         if (base.housenumber and base.housenumber > first) or len(query.source) > 1:
             penalty += 0.25
 
+        if self.direction == 0 and query.dir_penalty > 0:
+            penalty += query.dir_penalty
+
         for i in range(first.start + 1, first.end):
             name, addr = first.split(i)
             log().comment(f'split first word = name ({i - first.start})')
             yield dataclasses.replace(base, name=name, address=[addr] + base.address[1:],
-                                      penalty=penalty + PENALTY_TOKENCHANGE[query.nodes[i].btype])
+                                      penalty=penalty + query.nodes[i].word_break_penalty)
 
     def _get_assignments_address_backward(self, base: TokenAssignment,
                                           query: qmod.QueryStruct) -> Iterator[TokenAssignment]:
@@ -317,9 +316,18 @@ class _TokenSequence:
         """
         last = base.address[-1]
 
-        if self.direction == -1 or len(base.address) > 1:
+        # The postcode must come before the name for backward direction.
+        if base.postcode and base.postcode > last:
+            log().var_dump('skip backward', (base.postcode, last))
+            return
+
+        penalty = self.penalty
+        if not base.country and self.direction == -1 and query.dir_penalty < 0:
+            penalty -= query.dir_penalty
+
+        if self.direction == -1 or len(base.address) > 1 or base.postcode:
             log().comment('last word = name')
-            yield dataclasses.replace(base, penalty=self.penalty,
+            yield dataclasses.replace(base, penalty=penalty,
                                       name=last, address=base.address[:-1])
 
         # To paraphrase:
@@ -329,20 +337,22 @@ class _TokenSequence:
         #  * the containing phrase is strictly typed
         if (base.housenumber and last.start > base.housenumber.end)\
            or (base.qualifier and base.qualifier < last)\
-           or (query.nodes[last.start].ptype != qmod.PhraseType.NONE):
+           or (query.nodes[last.start].ptype != qmod.PHRASE_ANY):
             return
 
-        penalty = self.penalty
         if base.housenumber and base.housenumber < last:
             penalty += 0.4
         if len(query.source) > 1:
             penalty += 0.25
 
+        if self.direction == 0 and query.dir_penalty < 0:
+            penalty -= query.dir_penalty
+
         for i in range(last.start + 1, last.end):
             addr, name = last.split(i)
             log().comment(f'split last word = name ({i - last.start})')
             yield dataclasses.replace(base, name=name, address=base.address[:-1] + [addr],
-                                      penalty=penalty + PENALTY_TOKENCHANGE[query.nodes[i].btype])
+                                      penalty=penalty + query.nodes[i].word_break_penalty)
 
     def get_assignments(self, query: qmod.QueryStruct) -> Iterator[TokenAssignment]:
         """ Yield possible assignments for the current sequence.
@@ -370,11 +380,11 @@ class _TokenSequence:
             if base.postcode and base.postcode.start == 0:
                 self.penalty += 0.1
 
-            # Right-to-left reading of the address
+            # Left-to-right reading of the address
             if self.direction != -1:
                 yield from self._get_assignments_address_forward(base, query)
 
-            # Left-to-right reading of the address
+            # Right-to-left reading of the address
             if self.direction != 1:
                 yield from self._get_assignments_address_backward(base, query)
 
@@ -393,18 +403,32 @@ def yield_token_assignments(query: qmod.QueryStruct) -> Iterator[TokenAssignment
         another. It does not include penalties for transitions within a
         type.
     """
-    todo = [_TokenSequence([], direction=0 if query.source[0].ptype == qmod.PhraseType.NONE else 1)]
+    todo = [_TokenSequence([], direction=0 if query.source[0].ptype == qmod.PHRASE_ANY else 1)]
 
     while todo:
         state = todo.pop()
         node = query.nodes[state.end_pos]
 
         for tlist in node.starting:
-            newstate = state.advance(tlist.ttype, tlist.end, node.btype)
-            if newstate is not None:
-                if newstate.end_pos == query.num_token_slots():
-                    if newstate.recheck_sequence():
-                        log().var_dump('Assignment', newstate)
-                        yield from newstate.get_assignments(query)
-                elif not newstate.is_final():
-                    todo.append(newstate)
+            yield from _append_state_to_todo(
+                query, todo,
+                state.advance(tlist.ttype, tlist.end,
+                              True, node.word_break_penalty))
+
+        if node.partial is not None:
+            yield from _append_state_to_todo(
+                query, todo,
+                state.advance(qmod.TOKEN_PARTIAL, state.end_pos + 1,
+                              node.btype == qmod.BREAK_PHRASE,
+                              node.word_break_penalty))
+
+
+def _append_state_to_todo(query: qmod.QueryStruct, todo: List[_TokenSequence],
+                          newstate: Optional[_TokenSequence]) -> Iterator[TokenAssignment]:
+    if newstate is not None:
+        if newstate.end_pos == query.num_token_slots():
+            if newstate.recheck_sequence():
+                log().var_dump('Assignment', newstate)
+                yield from newstate.get_assignments(query)
+        elif not newstate.is_final():
+            todo.append(newstate)

src/nominatim_api/server/falcon/server.py

@@ -190,7 +190,7 @@ def get_application(project_dir: Path,
     """
     apimw = APIMiddleware(project_dir, environ)
 
-    middleware: List[object] = [apimw]
+    middleware: List[Any] = [apimw]
     log_file = apimw.config.LOG_FILE
     if log_file:
         middleware.append(FileLoggingMiddleware(log_file))

src/nominatim_api/server/starlette/server.py

@@ -143,7 +143,7 @@ def get_application(project_dir: Path,
 
     log_file = config.LOG_FILE
     if log_file:
-        middleware.append(Middleware(FileLoggingMiddleware, file_name=log_file))
+        middleware.append(Middleware(FileLoggingMiddleware, file_name=log_file))  # type: ignore
 
     exceptions: Dict[Any, Callable[[Request, Exception], Awaitable[Response]]] = {
         TimeoutError: timeout_error,

src/nominatim_api/sql/sqlalchemy_functions.py

@@ -122,15 +122,18 @@ class IsAddressPoint(sa.sql.functions.GenericFunction[Any]):
 
     def __init__(self, table: sa.Table) -> None:
         super().__init__(table.c.rank_address,
-                         table.c.housenumber, table.c.name)
+                         table.c.housenumber, table.c.name, table.c.address)
 
 
 @compiles(IsAddressPoint)
 def default_is_address_point(element: IsAddressPoint,
                              compiler: 'sa.Compiled', **kw: Any) -> str:
-    rank, hnr, name = list(element.clauses)
-    return "(%s = 30 AND (%s IS NOT NULL OR %s ? 'addr:housename'))" % (
+    rank, hnr, name, address = list(element.clauses)
+    return "(%s = 30 AND (%s IS NULL OR NOT %s ? '_inherited')" \
+           " AND (%s IS NOT NULL OR %s ? 'addr:housename'))" % (
                 compiler.process(rank, **kw),
+                compiler.process(address, **kw),
+                compiler.process(address, **kw),
                 compiler.process(hnr, **kw),
                 compiler.process(name, **kw))
 
@@ -138,9 +141,11 @@ def default_is_address_point(element: IsAddressPoint,
 @compiles(IsAddressPoint, 'sqlite')
 def sqlite_is_address_point(element: IsAddressPoint,
                             compiler: 'sa.Compiled', **kw: Any) -> str:
-    rank, hnr, name = list(element.clauses)
-    return "(%s = 30 AND coalesce(%s, json_extract(%s, '$.addr:housename')) IS NOT NULL)" % (
+    rank, hnr, name, address = list(element.clauses)
+    return "(%s = 30 AND json_extract(%s, '$._inherited') IS NULL" \
+           " AND coalesce(%s, json_extract(%s, '$.addr:housename')) IS NOT NULL)" % (
                 compiler.process(rank, **kw),
+                compiler.process(address, **kw),
                 compiler.process(hnr, **kw),
                 compiler.process(name, **kw))
 

src/nominatim_api/sql/sqlalchemy_types/geometry.py

@@ -173,7 +173,7 @@ class Geometry(types.UserDefinedType):  # type: ignore[type-arg]
     def __init__(self, subtype: str = 'Geometry'):
         self.subtype = subtype
 
-    def get_col_spec(self) -> str:
+    def get_col_spec(self, **_: Any) -> str:
         return f'GEOMETRY({self.subtype}, 4326)'
 
     def bind_processor(self, dialect: 'sa.Dialect') -> Callable[[Any], str]:

src/nominatim_api/sql/sqlalchemy_types/json.py

@@ -2,7 +2,7 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Common json type for different dialects.
@@ -24,6 +24,6 @@ class Json(sa.types.TypeDecorator[Any]):
 
     def load_dialect_impl(self, dialect: SaDialect) -> sa.types.TypeEngine[Any]:
         if dialect.name == 'postgresql':
-            return JSONB(none_as_null=True)  # type: ignore[no-untyped-call]
+            return JSONB(none_as_null=True)
 
         return sqlite_json(none_as_null=True)

src/nominatim_api/types.py

@@ -144,7 +144,7 @@ class Point(NamedTuple):
         except ValueError as exc:
             raise UsageError('Point parameter needs to be numbers.') from exc
 
-        if x < -180.0 or x > 180.0 or y < -90.0 or y > 90.0:
+        if not -180 <= x <= 180 or not -90 <= y <= 90.0:
             raise UsageError('Point coordinates invalid.')
 
         return Point(x, y)

src/nominatim_api/v1/classtypes.py

@@ -25,8 +25,8 @@ def get_label_tag(category: Tuple[str, str], extratags: Optional[Mapping[str, st
     elif rank < 26 and extratags and 'linked_place' in extratags:
         label = extratags['linked_place']
     elif category == ('boundary', 'administrative'):
-        label = ADMIN_LABELS.get((country or '', int(rank/2)))\
-                or ADMIN_LABELS.get(('', int(rank/2)))\
+        label = ADMIN_LABELS.get((country or '', rank // 2))\
+                or ADMIN_LABELS.get(('', rank // 2))\
                 or 'Administrative'
     elif category[1] == 'postal_code':
         label = 'postcode'

src/nominatim_api/v1/format_json.py

@@ -84,8 +84,9 @@ def format_base_json(results: Union[ReverseResults, SearchResults],
 
         _write_osm_id(out, result.osm_object)
 
-        out.keyval('lat', f"{result.centroid.lat}")\
-           .keyval('lon', f"{result.centroid.lon}")\
+        # lat and lon must be string values
+        out.keyval('lat', f"{result.centroid.lat:0.7f}")\
+           .keyval('lon', f"{result.centroid.lon:0.7f}")\
            .keyval(class_label, result.category[0])\
            .keyval('type', result.category[1])\
            .keyval('place_rank', result.rank_search)\
@@ -112,6 +113,7 @@ def format_base_json(results: Union[ReverseResults, SearchResults],
         if options.get('namedetails', False):
             out.keyval('namedetails', result.names)
 
+        # must be string values
         bbox = cl.bbox_from_result(result)
         out.key('boundingbox').start_array()\
            .value(f"{bbox.minlat:0.7f}").next()\
@@ -249,6 +251,9 @@ def format_base_geocodejson(results: Union[ReverseResults, SearchResults],
                         out.keyval(f"level{line.admin_level}", line.local_name)
             out.end_object().next()
 
+        if options.get('extratags', False):
+            out.keyval('extra', result.extratags)
+
         out.end_object().next().end_object().next()
 
         out.key('geometry').raw(result.geometry.get('geojson')

src/nominatim_api/v1/format_xml.py

@@ -90,7 +90,7 @@ def format_base_xml(results: Union[ReverseResults, SearchResults],
         result will be output, otherwise a list.
     """
     root = ET.Element(xml_root_tag)
-    root.set('timestamp', dt.datetime.utcnow().strftime('%a, %d %b %Y %H:%M:%S +00:00'))
+    root.set('timestamp', dt.datetime.now(dt.timezone.utc).strftime('%a, %d %b %Y %H:%M:%S +00:00'))
     root.set('attribution', cl.OSM_ATTRIBUTION)
     for k, v in xml_extra_info.items():
         root.set(k, v)

src/nominatim_api/v1/server_glue.py

@@ -374,14 +374,17 @@ async def deletable_endpoint(api: NominatimAPIAsync, params: ASGIAdaptor) -> Any
     """
     fmt = parse_format(params, RawDataList, 'json')
 
+    results = RawDataList()
     async with api.begin() as conn:
-        sql = sa.text(""" SELECT p.place_id, country_code,
-                                 name->'name' as name, i.*
-                          FROM placex p, import_polygon_delete i
-                          WHERE p.osm_id = i.osm_id AND p.osm_type = i.osm_type
-                                AND p.class = i.class AND p.type = i.type
-                      """)
-        results = RawDataList(r._asdict() for r in await conn.execute(sql))
+        for osm_type in ('N', 'W', 'R'):
+            sql = sa.text(""" SELECT p.place_id, country_code,
+                                     name->'name' as name, i.*
+                              FROM placex p, import_polygon_delete i
+                              WHERE i.osm_type = :osm_type
+                                    AND p.osm_id = i.osm_id AND p.osm_type = :osm_type
+                                    AND p.class = i.class AND p.type = i.type
+                          """)
+            results.extend(r._asdict() for r in await conn.execute(sql, {'osm_type': osm_type}))
 
     return build_response(params, params.formatting().format_result(results, fmt, {}))
 

src/nominatim_api/version.py

@@ -8,4 +8,4 @@
 Version information for the Nominatim API.
 """
 
-NOMINATIM_API_VERSION = '5.0.0'
+NOMINATIM_API_VERSION = '5.1.0'

src/nominatim_db/cli.py

@@ -2,16 +2,15 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Command-line interface to the Nominatim functions for import, update,
 database administration and querying.
 """
-from typing import Optional, Any
+from typing import Optional, List, Mapping
 import importlib
 import logging
-import os
 import sys
 import argparse
 import asyncio
@@ -81,13 +80,14 @@ class CommandlineParser:
         parser.set_defaults(command=cmd)
         cmd.add_args(parser)
 
-    def run(self, **kwargs: Any) -> int:
+    def run(self, cli_args: Optional[List[str]],
+            environ: Optional[Mapping[str, str]]) -> int:
         """ Parse the command line arguments of the program and execute the
             appropriate subcommand.
         """
         args = NominatimArgs()
         try:
-            self.parser.parse_args(args=kwargs.get('cli_args'), namespace=args)
+            self.parser.parse_args(args=cli_args, namespace=args)
         except SystemExit:
             return 1
 
@@ -101,23 +101,19 @@ class CommandlineParser:
 
         args.project_dir = Path(args.project_dir).resolve()
 
-        if 'cli_args' not in kwargs:
+        if cli_args is None:
             logging.basicConfig(stream=sys.stderr,
                                 format='%(asctime)s: %(message)s',
                                 datefmt='%Y-%m-%d %H:%M:%S',
                                 level=max(4 - args.verbose, 1) * 10)
 
-        args.config = Configuration(args.project_dir,
-                                    environ=kwargs.get('environ', os.environ))
-        args.config.set_libdirs(osm2pgsql=kwargs['osm2pgsql_path'])
+        args.config = Configuration(args.project_dir, environ=environ)
 
         log = logging.getLogger()
         log.warning('Using project directory: %s', str(args.project_dir))
 
         try:
-            ret = args.command.run(args)
-
-            return ret
+            return args.command.run(args)
         except UsageError as exception:
             if log.isEnabledFor(logging.DEBUG):
                 raise  # use Python's exception printing
@@ -233,9 +229,16 @@ def get_set_parser() -> CommandlineParser:
     return parser
 
 
-def nominatim(**kwargs: Any) -> int:
+def nominatim(cli_args: Optional[List[str]] = None,
+              environ: Optional[Mapping[str, str]] = None) -> int:
     """\
     Command-line tools for importing, updating, administrating and
     querying the Nominatim database.
+
+    'cli_args' is a list of parameters for the command to run. If not given,
+    sys.args will be used.
+
+    'environ' is the dictionary of environment variables containing the
+    Nominatim configuration. When None, the os.environ is inherited.
     """
-    return get_set_parser().run(**kwargs)
+    return get_set_parser().run(cli_args=cli_args, environ=environ)

src/nominatim_db/clicmd/args.py

@@ -2,7 +2,7 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Provides custom functions over command-line arguments.
@@ -136,6 +136,7 @@ class NominatimArgs:
     import_from_wiki: bool
     import_from_csv: Optional[str]
     no_replace: bool
+    min: int
 
     # Arguments to all query functions
     format: str
@@ -186,7 +187,7 @@ class NominatimArgs:
             from the command line arguments. The resulting dict can be
             further customized and then used in `run_osm2pgsql()`.
         """
-        return dict(osm2pgsql=self.config.OSM2PGSQL_BINARY or self.config.lib_dir.osm2pgsql,
+        return dict(osm2pgsql=self.config.OSM2PGSQL_BINARY,
                     osm2pgsql_cache=self.osm2pgsql_cache or default_cache,
                     osm2pgsql_style=self.config.get_import_style_file(),
                     osm2pgsql_style_path=self.config.lib_dir.lua,

src/nominatim_db/clicmd/special_phrases.py

@@ -58,6 +58,8 @@ class ImportSpecialPhrases:
                            help='Import special phrases from a CSV file')
         group.add_argument('--no-replace', action='store_true',
                            help='Keep the old phrases and only add the new ones')
+        group.add_argument('--min', type=int, default=0,
+                           help='Restrict special phrases by minimum occurance')
 
     def run(self, args: NominatimArgs) -> int:
 
@@ -82,7 +84,9 @@ class ImportSpecialPhrases:
 
         tokenizer = tokenizer_factory.get_tokenizer_for_db(args.config)
         should_replace = not args.no_replace
+        min = args.min
+
         with connect(args.config.get_libpq_dsn()) as db_connection:
             SPImporter(
                 args.config, db_connection, loader
-            ).import_phrases(tokenizer, should_replace)
+            ).import_phrases(tokenizer, should_replace, min)

src/nominatim_db/config.py

@@ -2,7 +2,7 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Nominatim configuration accessor.
@@ -73,7 +73,6 @@ class Configuration:
             self.project_dir = None
 
         class _LibDirs:
-            osm2pgsql: Path
             sql = paths.SQLLIB_DIR
             lua = paths.LUALIB_DIR
             data = paths.DATA_DIR

src/nominatim_db/db/connection.py

@@ -102,10 +102,10 @@ def server_version_tuple(conn: Connection) -> Tuple[int, int]:
         Converts correctly for pre-10 and post-10 PostgreSQL versions.
     """
     version = conn.info.server_version
-    if version < 100000:
-        return (int(version / 10000), int((version % 10000) / 100))
-
-    return (int(version / 10000), version % 10000)
+    major, minor = divmod(version, 10000)
+    if major < 10:
+        minor //= 100
+    return major, minor
 
 
 def postgis_version_tuple(conn: Connection) -> Tuple[int, int]:

src/nominatim_db/indexer/progress.py

@@ -50,8 +50,8 @@ class ProgressLogger:
         places_per_sec = self.done_places / done_time
         eta = (self.total_places - self.done_places) / places_per_sec
 
-        LOG.warning("Done %d in %d @ %.3f per second - %s ETA (seconds): %.2f",
-                    self.done_places, int(done_time),
+        LOG.warning("Done %d in %.0f @ %.3f per second - %s ETA (seconds): %.2f",
+                    self.done_places, done_time,
                     places_per_sec, self.name, eta)
 
         self.next_info += int(places_per_sec) * self.log_interval
@@ -68,8 +68,8 @@ class ProgressLogger:
             diff_seconds = (rank_end_time - self.rank_start_time).total_seconds()
             places_per_sec = self.done_places / diff_seconds
 
-        LOG.warning("Done %d/%d in %d @ %.3f per second - FINISHED %s\n",
-                    self.done_places, self.total_places, int(diff_seconds),
+        LOG.warning("Done %d/%d in %.0f @ %.3f per second - FINISHED %s\n",
+                    self.done_places, self.total_places, diff_seconds,
                     places_per_sec, self.name)
 
         return self.done_places

src/nominatim_db/tokenizer/base.py

@@ -2,7 +2,7 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Abstract class definitions for tokenizers. These base classes are here
@@ -10,7 +10,6 @@ mainly for documentation purposes.
 """
 from abc import ABC, abstractmethod
 from typing import List, Tuple, Dict, Any, Optional, Iterable
-from pathlib import Path
 
 from ..typing import Protocol
 from ..config import Configuration
@@ -38,7 +37,7 @@ class AbstractAnalyzer(ABC):
         """
 
     @abstractmethod
-    def get_word_token_info(self, words: List[str]) -> List[Tuple[str, str, int]]:
+    def get_word_token_info(self, words: List[str]) -> List[Tuple[str, str, Optional[int]]]:
         """ Return token information for the given list of words.
 
             The function is used for testing and debugging only
@@ -232,6 +231,6 @@ class TokenizerModule(Protocol):
         own tokenizer.
     """
 
-    def create(self, dsn: str, data_dir: Path) -> AbstractTokenizer:
+    def create(self, dsn: str) -> AbstractTokenizer:
         """ Factory for new tokenizers.
         """

src/nominatim_db/tokenizer/factory.py

@@ -2,7 +2,7 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Functions for creating a tokenizer or initialising the right one for an
@@ -52,19 +52,10 @@ def create_tokenizer(config: Configuration, init_db: bool = True,
     if module_name is None:
         module_name = config.TOKENIZER
 
-    # Create the directory for the tokenizer data
-    assert config.project_dir is not None
-    basedir = config.project_dir / 'tokenizer'
-    if not basedir.exists():
-        basedir.mkdir()
-    elif not basedir.is_dir():
-        LOG.fatal("Tokenizer directory '%s' cannot be created.", basedir)
-        raise UsageError("Tokenizer setup failed.")
-
     # Import and initialize the tokenizer.
     tokenizer_module = _import_tokenizer(module_name)
 
-    tokenizer = tokenizer_module.create(config.get_libpq_dsn(), basedir)
+    tokenizer = tokenizer_module.create(config.get_libpq_dsn())
     tokenizer.init_new_db(config, init_db=init_db)
 
     with connect(config.get_libpq_dsn()) as conn:
@@ -79,12 +70,6 @@ def get_tokenizer_for_db(config: Configuration) -> AbstractTokenizer:
         The function looks up the appropriate tokenizer in the database
         and initialises it.
     """
-    assert config.project_dir is not None
-    basedir = config.project_dir / 'tokenizer'
-    if not basedir.is_dir():
-        # Directory will be repopulated by tokenizer below.
-        basedir.mkdir()
-
     with connect(config.get_libpq_dsn()) as conn:
         name = properties.get_property(conn, 'tokenizer')
 
@@ -94,7 +79,7 @@ def get_tokenizer_for_db(config: Configuration) -> AbstractTokenizer:
 
     tokenizer_module = _import_tokenizer(name)
 
-    tokenizer = tokenizer_module.create(config.get_libpq_dsn(), basedir)
+    tokenizer = tokenizer_module.create(config.get_libpq_dsn())
     tokenizer.init_from_project(config)
 
     return tokenizer

src/nominatim_db/tokenizer/icu_tokenizer.py

@@ -2,7 +2,7 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Tokenizer implementing normalisation as used before Nominatim 4 but using
@@ -12,7 +12,6 @@ from typing import Optional, Sequence, List, Tuple, Mapping, Any, cast, \
                    Dict, Set, Iterable
 import itertools
 import logging
-from pathlib import Path
 
 from psycopg.types.json import Jsonb
 from psycopg import sql as pysql
@@ -38,10 +37,10 @@ WORD_TYPES = (('country_names', 'C'),
               ('housenumbers', 'H'))
 
 
-def create(dsn: str, data_dir: Path) -> 'ICUTokenizer':
+def create(dsn: str) -> 'ICUTokenizer':
     """ Create a new instance of the tokenizer provided by this module.
     """
-    return ICUTokenizer(dsn, data_dir)
+    return ICUTokenizer(dsn)
 
 
 class ICUTokenizer(AbstractTokenizer):
@@ -50,9 +49,8 @@ class ICUTokenizer(AbstractTokenizer):
         normalization routines in Nominatim 3.
     """
 
-    def __init__(self, dsn: str, data_dir: Path) -> None:
+    def __init__(self, dsn: str) -> None:
         self.dsn = dsn
-        self.data_dir = data_dir
         self.loader: Optional[ICURuleLoader] = None
 
     def init_new_db(self, config: Configuration, init_db: bool = True) -> None:
@@ -121,10 +119,10 @@ class ICUTokenizer(AbstractTokenizer):
                            SELECT unnest(nameaddress_vector) as id, count(*)
                                  FROM search_name GROUP BY id)
                   SELECT coalesce(a.id, w.id) as id,
-                         (CASE WHEN w.count is null THEN '{}'::JSONB
+                         (CASE WHEN w.count is null or w.count <= 1 THEN '{}'::JSONB
                               ELSE jsonb_build_object('count', w.count) END
                           ||
-                          CASE WHEN a.count is null THEN '{}'::JSONB
+                          CASE WHEN a.count is null or a.count <= 1 THEN '{}'::JSONB
                               ELSE jsonb_build_object('addr_count', a.count) END) as info
                   FROM word_freq w FULL JOIN addr_freq a ON a.id = w.id;
                   """)
@@ -134,9 +132,10 @@ class ICUTokenizer(AbstractTokenizer):
                 drop_tables(conn, 'tmp_word')
                 cur.execute("""CREATE TABLE tmp_word AS
                                 SELECT word_id, word_token, type, word,
-                                       (CASE WHEN wf.info is null THEN word.info
-                                        ELSE coalesce(word.info, '{}'::jsonb) || wf.info
-                                        END) as info
+                                       coalesce(word.info, '{}'::jsonb)
+                                       - 'count' - 'addr_count' ||
+                                       coalesce(wf.info, '{}'::jsonb)
+                                       as info
                                 FROM word LEFT JOIN word_frequencies wf
                                      ON word.word_id = wf.id
                             """)
@@ -339,7 +338,7 @@ class ICUNameAnalyzer(AbstractAnalyzer):
         """
         return cast(str, self.token_analysis.normalizer.transliterate(name)).strip()
 
-    def get_word_token_info(self, words: Sequence[str]) -> List[Tuple[str, str, int]]:
+    def get_word_token_info(self, words: Sequence[str]) -> List[Tuple[str, str, Optional[int]]]:
         """ Return token information for the given list of words.
             If a word starts with # it is assumed to be a full name
             otherwise is a partial name.
@@ -363,11 +362,11 @@ class ICUNameAnalyzer(AbstractAnalyzer):
             cur.execute("""SELECT word_token, word_id
                             FROM word WHERE word_token = ANY(%s) and type = 'W'
                         """, (list(full_tokens.values()),))
-            full_ids = {r[0]: r[1] for r in cur}
+            full_ids = {r[0]: cast(int, r[1]) for r in cur}
             cur.execute("""SELECT word_token, word_id
                             FROM word WHERE word_token = ANY(%s) and type = 'w'""",
                         (list(partial_tokens.values()),))
-            part_ids = {r[0]: r[1] for r in cur}
+            part_ids = {r[0]: cast(int, r[1]) for r in cur}
 
         return [(k, v, full_ids.get(v, None)) for k, v in full_tokens.items()] \
             + [(k, v, part_ids.get(v, None)) for k, v in partial_tokens.items()]
@@ -381,76 +380,15 @@ class ICUNameAnalyzer(AbstractAnalyzer):
         return postcode.strip().upper()
 
     def update_postcodes_from_db(self) -> None:
-        """ Update postcode tokens in the word table from the location_postcode
-            table.
+        """ Postcode update.
+
+            Removes all postcodes from the word table because they are not
+            needed. Postcodes are recognised by pattern.
         """
         assert self.conn is not None
-        analyzer = self.token_analysis.analysis.get('@postcode')
 
         with self.conn.cursor() as cur:
-            # First get all postcode names currently in the word table.
-            cur.execute("SELECT DISTINCT word FROM word WHERE type = 'P'")
-            word_entries = set((entry[0] for entry in cur))
-
-            # Then compute the required postcode names from the postcode table.
-            needed_entries = set()
-            cur.execute("SELECT country_code, postcode FROM location_postcode")
-            for cc, postcode in cur:
-                info = PlaceInfo({'country_code': cc,
-                                  'class': 'place', 'type': 'postcode',
-                                  'address': {'postcode': postcode}})
-                address = self.sanitizer.process_names(info)[1]
-                for place in address:
-                    if place.kind == 'postcode':
-                        if analyzer is None:
-                            postcode_name = place.name.strip().upper()
-                            variant_base = None
-                        else:
-                            postcode_name = analyzer.get_canonical_id(place)
-                            variant_base = place.get_attr("variant")
-
-                        if variant_base:
-                            needed_entries.add(f'{postcode_name}@{variant_base}')
-                        else:
-                            needed_entries.add(postcode_name)
-                        break
-
-        # Now update the word table.
-        self._delete_unused_postcode_words(word_entries - needed_entries)
-        self._add_missing_postcode_words(needed_entries - word_entries)
-
-    def _delete_unused_postcode_words(self, tokens: Iterable[str]) -> None:
-        assert self.conn is not None
-        if tokens:
-            with self.conn.cursor() as cur:
-                cur.execute("DELETE FROM word WHERE type = 'P' and word = any(%s)",
-                            (list(tokens), ))
-
-    def _add_missing_postcode_words(self, tokens: Iterable[str]) -> None:
-        assert self.conn is not None
-        if not tokens:
-            return
-
-        analyzer = self.token_analysis.analysis.get('@postcode')
-        terms = []
-
-        for postcode_name in tokens:
-            if '@' in postcode_name:
-                term, variant = postcode_name.split('@', 2)
-                term = self._search_normalized(term)
-                if analyzer is None:
-                    variants = [term]
-                else:
-                    variants = analyzer.compute_variants(variant)
-                    if term not in variants:
-                        variants.append(term)
-            else:
-                variants = [self._search_normalized(postcode_name)]
-            terms.append((postcode_name, variants))
-
-        if terms:
-            with self.conn.cursor() as cur:
-                cur.executemany("""SELECT create_postcode_word(%s, %s)""", terms)
+            cur.execute("DELETE FROM word WHERE type = 'P'")
 
     def update_special_phrases(self, phrases: Iterable[Tuple[str, str, str, str]],
                                should_replace: bool) -> None:
@@ -645,10 +583,14 @@ class ICUNameAnalyzer(AbstractAnalyzer):
             if word_id:
                 result = self._cache.housenumbers.get(word_id, result)
                 if result[0] is None:
-                    variants = analyzer.compute_variants(word_id)
+                    varout = analyzer.compute_variants(word_id)
+                    if isinstance(varout, tuple):
+                        variants = varout[0]
+                    else:
+                        variants = varout
                     if variants:
                         hid = execute_scalar(self.conn, "SELECT create_analyzed_hnr_id(%s, %s)",
-                                             (word_id, list(variants)))
+                                             (word_id, variants))
                         result = hid, variants[0]
                         self._cache.housenumbers[word_id] = result
 
@@ -693,13 +635,17 @@ class ICUNameAnalyzer(AbstractAnalyzer):
 
             full, part = self._cache.names.get(token_id, (None, None))
             if full is None:
-                variants = analyzer.compute_variants(word_id)
+                varset = analyzer.compute_variants(word_id)
+                if isinstance(varset, tuple):
+                    variants, lookups = varset
+                else:
+                    variants, lookups = varset, None
                 if not variants:
                     continue
 
                 with self.conn.cursor() as cur:
-                    cur.execute("SELECT * FROM getorcreate_full_word(%s, %s)",
-                                (token_id, variants))
+                    cur.execute("SELECT * FROM getorcreate_full_word(%s, %s, %s)",
+                                (token_id, variants, lookups))
                     full, part = cast(Tuple[int, List[int]], cur.fetchone())
 
                 self._cache.names[token_id] = (full, part)
@@ -718,32 +664,9 @@ class ICUNameAnalyzer(AbstractAnalyzer):
         analyzer = self.token_analysis.analysis.get('@postcode')
 
         if analyzer is None:
-            postcode_name = item.name.strip().upper()
-            variant_base = None
+            return item.name.strip().upper()
         else:
-            postcode_name = analyzer.get_canonical_id(item)
-            variant_base = item.get_attr("variant")
-
-        if variant_base:
-            postcode = f'{postcode_name}@{variant_base}'
-        else:
-            postcode = postcode_name
-
-        if postcode not in self._cache.postcodes:
-            term = self._search_normalized(postcode_name)
-            if not term:
-                return None
-
-            variants = {term}
-            if analyzer is not None and variant_base:
-                variants.update(analyzer.compute_variants(variant_base))
-
-            with self.conn.cursor() as cur:
-                cur.execute("SELECT create_postcode_word(%s, %s)",
-                            (postcode, list(variants)))
-            self._cache.postcodes.add(postcode)
-
-        return postcode_name
+            return analyzer.get_canonical_id(item)
 
 
 class _TokenInfo:
@@ -836,5 +759,4 @@ class _TokenCache:
         self.names: Dict[str, Tuple[int, List[int]]] = {}
         self.partials: Dict[str, int] = {}
         self.fulls: Dict[str, List[int]] = {}
-        self.postcodes: Set[str] = set()
         self.housenumbers: Dict[str, Tuple[Optional[int], Optional[str]]] = {}

src/nominatim_db/tokenizer/token_analysis/base.py

@@ -7,7 +7,7 @@
 """
 Common data types and protocols for analysers.
 """
-from typing import Mapping, List, Any
+from typing import Mapping, List, Any, Union, Tuple
 
 from ...typing import Protocol
 from ...data.place_name import PlaceName
@@ -33,7 +33,7 @@ class Analyzer(Protocol):
                     for example because the character set in use does not match.
         """
 
-    def compute_variants(self, canonical_id: str) -> List[str]:
+    def compute_variants(self, canonical_id: str) -> Union[List[str], Tuple[List[str], List[str]]]:
         """ Compute the transliterated spelling variants for the given
             canonical ID.
 

src/nominatim_db/tokenizer/token_analysis/generic.py

@@ -2,20 +2,19 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Generic processor for names that creates abbreviation variants.
 """
-from typing import Mapping, Dict, Any, Iterable, Iterator, Optional, List, cast
+from typing import Mapping, Dict, Any, Iterable, Optional, List, cast, Tuple
 import itertools
 
-import datrie
-
 from ...errors import UsageError
 from ...data.place_name import PlaceName
 from .config_variants import get_variant_config
 from .generic_mutation import MutationVariantGenerator
+from .simple_trie import SimpleTrie
 
 # Configuration section
 
@@ -25,8 +24,7 @@ def configure(rules: Mapping[str, Any], normalizer: Any, _: Any) -> Dict[str, An
     """
     config: Dict[str, Any] = {}
 
-    config['replacements'], config['chars'] = get_variant_config(rules.get('variants'),
-                                                                 normalizer)
+    config['replacements'], _ = get_variant_config(rules.get('variants'), normalizer)
     config['variant_only'] = rules.get('mode', '') == 'variant-only'
 
     # parse mutation rules
@@ -68,12 +66,8 @@ class GenericTokenAnalysis:
         self.variant_only = config['variant_only']
 
         # Set up datrie
-        if config['replacements']:
-            self.replacements = datrie.Trie(config['chars'])
-            for src, repllist in config['replacements']:
-                self.replacements[src] = repllist
-        else:
-            self.replacements = None
+        self.replacements: Optional[SimpleTrie[List[str]]] = \
+            SimpleTrie(config['replacements']) if config['replacements'] else None
 
         # set up mutation rules
         self.mutations = [MutationVariantGenerator(*cfg) for cfg in config['mutations']]
@@ -84,7 +78,7 @@ class GenericTokenAnalysis:
         """
         return cast(str, self.norm.transliterate(name.name)).strip()
 
-    def compute_variants(self, norm_name: str) -> List[str]:
+    def compute_variants(self, norm_name: str) -> Tuple[List[str], List[str]]:
         """ Compute the spelling variants for the given normalized name
             and transliterate the result.
         """
@@ -93,18 +87,20 @@ class GenericTokenAnalysis:
         for mutation in self.mutations:
             variants = mutation.generate(variants)
 
-        return [name for name in self._transliterate_unique_list(norm_name, variants) if name]
-
-    def _transliterate_unique_list(self, norm_name: str,
-                                   iterable: Iterable[str]) -> Iterator[Optional[str]]:
-        seen = set()
+        varset = set(map(str.strip, variants))
         if self.variant_only:
-            seen.add(norm_name)
+            varset.discard(norm_name)
 
-        for variant in map(str.strip, iterable):
-            if variant not in seen:
-                seen.add(variant)
-                yield self.to_ascii.transliterate(variant).strip()
+        trans = []
+        norm = []
+
+        for var in varset:
+            t = self.to_ascii.transliterate(var).strip()
+            if t:
+                trans.append(t)
+                norm.append(var)
+
+        return trans, norm
 
     def _generate_word_variants(self, norm_name: str) -> Iterable[str]:
         baseform = '^ ' + norm_name + ' ^'
@@ -116,10 +112,10 @@ class GenericTokenAnalysis:
             pos = 0
             force_space = False
             while pos < baselen:
-                full, repl = self.replacements.longest_prefix_item(baseform[pos:],
-                                                                   (None, None))
-                if full is not None:
-                    done = baseform[startpos:pos]
+                frm = pos
+                repl, pos = self.replacements.longest_prefix(baseform, pos)
+                if repl is not None:
+                    done = baseform[startpos:frm]
                     partials = [v + done + r
                                 for v, r in itertools.product(partials, repl)
                                 if not force_space or r.startswith(' ')]
@@ -128,11 +124,10 @@ class GenericTokenAnalysis:
                         # to be helpful. Only use the original term.
                         startpos = 0
                         break
-                    startpos = pos + len(full)
-                    if full[-1] == ' ':
-                        startpos -= 1
+                    if baseform[pos - 1] == ' ':
+                        pos -= 1
                         force_space = True
-                    pos = startpos
+                    startpos = pos
                 else:
                     pos += 1
                     force_space = False

src/nominatim_db/tokenizer/token_analysis/simple_trie.py

@@ -0,0 +1,84 @@
+# SPDX-License-Identifier: GPL-3.0-or-later
+#
+# This file is part of Nominatim. (https://nominatim.org)
+#
+# Copyright (C) 2025 by the Nominatim developer community.
+# For a full list of authors see the git log.
+"""
+Simple dict-based implementation of a trie structure.
+"""
+from typing import TypeVar, Generic, Tuple, Optional, List, Dict
+from collections import defaultdict
+
+T = TypeVar('T')
+
+
+class SimpleTrie(Generic[T]):
+    """ A simple read-only trie structure.
+        This structure supports examply one lookup operation,
+        which is longest-prefix lookup.
+    """
+
+    def __init__(self, data: Optional[List[Tuple[str, T]]] = None) -> None:
+        self._tree: Dict[str, 'SimpleTrie[T]'] = defaultdict(SimpleTrie[T])
+        self._value: Optional[T] = None
+        self._prefix = ''
+
+        if data:
+            for key, value in data:
+                self._add(key, 0, value)
+
+            self._make_compact()
+
+    def _add(self, word: str, pos: int, value: T) -> None:
+        """ (Internal) Add a sub-word to the trie.
+            The word is added from index 'pos'. If the sub-word to add
+            is empty, then the trie saves the given value.
+        """
+        if pos < len(word):
+            self._tree[word[pos]]._add(word, pos + 1, value)
+        else:
+            self._value = value
+
+    def _make_compact(self) -> None:
+        """ (Internal) Compress tree where there is exactly one subtree
+            and no value.
+
+            Compression works recursively starting at the leaf.
+        """
+        for t in self._tree.values():
+            t._make_compact()
+
+        if len(self._tree) == 1 and self._value is None:
+            assert not self._prefix
+            for k, v in self._tree.items():
+                self._prefix = k + v._prefix
+                self._tree = v._tree
+                self._value = v._value
+
+    def longest_prefix(self, word: str, start: int = 0) -> Tuple[Optional[T], int]:
+        """ Return the longest prefix match for the given word starting at
+            the position 'start'.
+
+            The function returns a tuple with the value for the longest match and
+            the position of the word after the match. If no match was found at
+            all, the function returns (None, start).
+        """
+        cur = self
+        pos = start
+        result: Tuple[Optional[T], int] = None, start
+
+        while True:
+            if cur._prefix:
+                if not word.startswith(cur._prefix, pos):
+                    return result
+                pos += len(cur._prefix)
+
+            if cur._value:
+                result = cur._value, pos
+
+            if pos >= len(word) or word[pos] not in cur._tree:
+                return result
+
+            cur = cur._tree[word[pos]]
+            pos += 1

src/nominatim_db/tools/check_database.py

@@ -2,7 +2,7 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Collection of functions that check if the database is complete and functional.
@@ -163,12 +163,8 @@ def check_connection(conn: Any, config: Configuration) -> CheckResult:
              Database version ({db_version}) doesn't match Nominatim version ({nom_version})
 
              Hints:
-             * Are you connecting to the correct database?
-
              {instruction}
 
-             Check the Migration chapter of the Administration Guide.
-
              Project directory: {config.project_dir}
              Current setting of NOMINATIM_DATABASE_DSN: {config.DATABASE_DSN}
              """)
@@ -176,24 +172,25 @@ def check_database_version(conn: Connection, config: Configuration) -> CheckResu
     """ Checking database_version matches Nominatim software version
     """
 
-    if table_exists(conn, 'nominatim_properties'):
+    db_version_str = None
+    if not table_exists(conn, 'nominatim_properties'):
+        instruction = 'Are you connecting to the correct database?'
+    else:
         db_version_str = properties.get_property(conn, 'database_version')
-    else:
-        db_version_str = None
 
-    if db_version_str is not None:
-        db_version = parse_version(db_version_str)
+        if db_version_str is None:
+            instruction = 'Database version not found. Did the import finish?'
+        else:
+            db_version = parse_version(db_version_str)
 
-        if db_version == NOMINATIM_VERSION:
-            return CheckState.OK
+            if db_version == NOMINATIM_VERSION:
+                return CheckState.OK
 
-        instruction = (
-            'Run migrations: nominatim admin --migrate'
-            if db_version < NOMINATIM_VERSION
-            else 'You need to upgrade the Nominatim software.'
-        )
-    else:
-        instruction = ''
+            instruction = (
+                "Run migrations: 'nominatim admin --migrate'"
+                if db_version < NOMINATIM_VERSION
+                else 'You need to upgrade the Nominatim software.'
+            ) + ' Check the Migration chapter of the Administration Guide.'
 
     return CheckState.FATAL, dict(db_version=db_version_str,
                                   nom_version=NOMINATIM_VERSION,

src/nominatim_db/tools/database_import.py

@@ -127,7 +127,7 @@ def import_osm_data(osm_files: Union[Path, Sequence[Path]],
                 fsize += os.stat(str(fname)).st_size
         else:
             fsize = os.stat(str(osm_files)).st_size
-        options['osm2pgsql_cache'] = int(min((mem.available + mem.cached) * 0.75,
+        options['osm2pgsql_cache'] = int(min((mem.available + getattr(mem, 'cached', 0)) * 0.75,
                                              fsize * 2) / 1024 / 1024) + 1
 
     run_osm2pgsql(options)

src/nominatim_db/tools/exec_utils.py

@@ -2,7 +2,7 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Helper functions for executing external programs.
@@ -37,21 +37,17 @@ def run_osm2pgsql(options: Mapping[str, Any]) -> None:
            '--style', str(options['osm2pgsql_style'])
            ]
 
-    if str(options['osm2pgsql_style']).endswith('.lua'):
-        env['LUA_PATH'] = ';'.join((str(options['osm2pgsql_style_path'] / '?.lua'),
-                                    os.environ.get('LUA_PATH', ';')))
-        env['THEMEPARK_PATH'] = str(options['osm2pgsql_style_path'] / 'themes')
-        if 'THEMEPARK_PATH' in os.environ:
-            env['THEMEPARK_PATH'] += ':' + os.environ['THEMEPARK_PATH']
-        cmd.extend(('--output', 'flex'))
+    env['LUA_PATH'] = ';'.join((str(options['osm2pgsql_style_path'] / '?.lua'),
+                                os.environ.get('LUA_PATH', ';')))
+    env['THEMEPARK_PATH'] = str(options['osm2pgsql_style_path'] / 'themes')
+    if 'THEMEPARK_PATH' in os.environ:
+        env['THEMEPARK_PATH'] += ':' + os.environ['THEMEPARK_PATH']
+    cmd.extend(('--output', 'flex'))
 
-        for flavour in ('data', 'index'):
-            if options['tablespaces'][f"main_{flavour}"]:
-                env[f"NOMINATIM_TABLESPACE_PLACE_{flavour.upper()}"] = \
-                    options['tablespaces'][f"main_{flavour}"]
-    else:
-        cmd.extend(('--output', 'gazetteer', '--hstore', '--latlon'))
-        cmd.extend(_mk_tablespace_options('main', options))
+    for flavour in ('data', 'index'):
+        if options['tablespaces'][f"main_{flavour}"]:
+            env[f"NOMINATIM_TABLESPACE_PLACE_{flavour.upper()}"] = \
+                options['tablespaces'][f"main_{flavour}"]
 
     if options['flatnode_file']:
         cmd.extend(('--flat-nodes', options['flatnode_file']))
@@ -85,7 +81,7 @@ def _mk_tablespace_options(ttype: str, options: Mapping[str, Any]) -> List[str]:
 
 
 def _find_osm2pgsql_cmd(cmdline: Optional[str]) -> str:
-    if cmdline is not None:
+    if cmdline:
         return cmdline
 
     in_path = shutil.which('osm2pgsql')

src/nominatim_db/tools/postcodes.py

@@ -2,7 +2,7 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Functions for importing, updating and otherwise maintaining the table
@@ -64,11 +64,15 @@ class _PostcodeCollector:
             if normalized:
                 self.collected[normalized] += (x, y)
 
-    def commit(self, conn: Connection, analyzer: AbstractAnalyzer, project_dir: Path) -> None:
-        """ Update postcodes for the country from the postcodes selected so far
-            as well as any externally supplied postcodes.
+    def commit(self, conn: Connection, analyzer: AbstractAnalyzer,
+               project_dir: Optional[Path]) -> None:
+        """ Update postcodes for the country from the postcodes selected so far.
+
+            When 'project_dir' is set, then any postcode files found in this
+            directory are taken into account as well.
         """
-        self._update_from_external(analyzer, project_dir)
+        if project_dir is not None:
+            self._update_from_external(analyzer, project_dir)
         to_add, to_delete, to_update = self._compute_changes(conn)
 
         LOG.info("Processing country '%s' (%s added, %s deleted, %s updated).",
@@ -170,7 +174,7 @@ class _PostcodeCollector:
         return None
 
 
-def update_postcodes(dsn: str, project_dir: Path, tokenizer: AbstractTokenizer) -> None:
+def update_postcodes(dsn: str, project_dir: Optional[Path], tokenizer: AbstractTokenizer) -> None:
     """ Update the table of artificial postcodes.
 
         Computes artificial postcode centroids from the placex table,

src/nominatim_db/tools/refresh.py

@@ -2,12 +2,12 @@
 #
 # This file is part of Nominatim. (https://nominatim.org)
 #
-# Copyright (C) 2024 by the Nominatim developer community.
+# Copyright (C) 2025 by the Nominatim developer community.
 # For a full list of authors see the git log.
 """
 Functions for bringing auxiliary data in the database up-to-date.
 """
-from typing import MutableSequence, Tuple, Any, Type, Mapping, Sequence, List, cast
+from typing import MutableSequence, Tuple, Any, Mapping, Sequence, List
 import csv
 import gzip
 import logging
@@ -212,31 +212,16 @@ def recompute_importance(conn: Connection) -> None:
              WHERE s.place_id = d.linked_place_id and d.wikipedia is not null
                    and (s.wikipedia is null or s.importance < d.importance);
             """)
+        cur.execute("""
+            UPDATE search_name s SET importance = p.importance
+             FROM placex p
+             WHERE s.place_id = p.place_id AND s.importance != p.importance
+            """)
 
         cur.execute('ALTER TABLE placex ENABLE TRIGGER ALL')
     conn.commit()
 
 
-def _quote_php_variable(var_type: Type[Any], config: Configuration,
-                        conf_name: str) -> str:
-    if var_type == bool:
-        return 'true' if config.get_bool(conf_name) else 'false'
-
-    if var_type == int:
-        return cast(str, getattr(config, conf_name))
-
-    if not getattr(config, conf_name):
-        return 'false'
-
-    if var_type == Path:
-        value = str(config.get_path(conf_name) or '')
-    else:
-        value = getattr(config, conf_name)
-
-    quoted = value.replace("'", "\\'")
-    return f"'{quoted}'"
-
-
 def invalidate_osm_object(osm_type: str, osm_id: int, conn: Connection,
                           recursive: bool = True) -> None:
     """ Mark the given OSM object for reindexing. When 'recursive' is set

src/nominatim_db/tools/special_phrases/sp_importer.py

@@ -16,7 +16,6 @@
 from typing import Iterable, Tuple, Mapping, Sequence, Optional, Set
 import logging
 import re
-
 from psycopg.sql import Identifier, SQL
 
 from ...typing import Protocol
@@ -65,7 +64,32 @@ class SPImporter():
         # special phrases class/type on the wiki.
         self.table_phrases_to_delete: Set[str] = set()
 
-    def import_phrases(self, tokenizer: AbstractTokenizer, should_replace: bool) -> None:
+    def get_classtype_pairs(self, min: int = 0) -> Set[Tuple[str, str]]:
+        """
+            Returns list of allowed special phrases from the database,
+            restricting to a list of combinations of classes and types
+            which occur equal to or more than a specified amount of times.
+
+            Default value for this is 0, which allows everything in database.
+        """
+        db_combinations = set()
+
+        query = f"""
+        SELECT class AS CLS, type AS typ
+        FROM placex
+        GROUP BY class, type
+        HAVING COUNT(*) >= {min}
+        """
+
+        with self.db_connection.cursor() as db_cursor:
+            db_cursor.execute(SQL(query))
+            for row in db_cursor:
+                db_combinations.add((row[0], row[1]))
+
+        return db_combinations
+
+    def import_phrases(self, tokenizer: AbstractTokenizer, should_replace: bool,
+                       min: int = 0) -> None:
         """
             Iterate through all SpecialPhrases extracted from the
             loader and import them into the database.
@@ -85,9 +109,10 @@ class SPImporter():
             if result:
                 class_type_pairs.add(result)
 
-        self._create_classtype_table_and_indexes(class_type_pairs)
+        self._create_classtype_table_and_indexes(class_type_pairs, min)
         if should_replace:
             self._remove_non_existent_tables_from_db()
+
         self.db_connection.commit()
 
         with tokenizer.name_analyzer() as analyzer:
@@ -163,7 +188,8 @@ class SPImporter():
         return (phrase.p_class, phrase.p_type)
 
     def _create_classtype_table_and_indexes(self,
-                                            class_type_pairs: Iterable[Tuple[str, str]]) -> None:
+                                            class_type_pairs: Iterable[Tuple[str, str]],
+                                            min: int = 0) -> None:
         """
             Create table place_classtype for each given pair.
             Also create indexes on place_id and centroid.
@@ -177,10 +203,19 @@ class SPImporter():
         with self.db_connection.cursor() as db_cursor:
             db_cursor.execute("CREATE INDEX idx_placex_classtype ON placex (class, type)")
 
+        if min:
+            allowed_special_phrases = self.get_classtype_pairs(min)
+
         for pair in class_type_pairs:
             phrase_class = pair[0]
             phrase_type = pair[1]
 
+            # Will only filter if min is not 0
+            if min and (phrase_class, phrase_type) not in allowed_special_phrases:
+                LOG.warning("Skipping phrase %s=%s: not in allowed special phrases",
+                            phrase_class, phrase_type)
+                continue
+
             table_name = _classtype_table(phrase_class, phrase_type)
 
             if table_name in self.table_phrases_to_delete: