Commit Graph

3458 Commits

Author SHA1 Message Date
Sarah Hoffmann
50fccb52be remove unused test files 2021-11-30 11:44:10 +01:00
Sarah Hoffmann
b90e719da5 organise python tests in subdirectories
The directories follow the same structure as the modules in
nominatim/.
2021-11-30 11:22:26 +01:00
Sarah Hoffmann
97f1723181 Merge pull request #2530 from lonvia/declassify-highway
Change default rank for highway objects to 30
2021-11-25 08:41:25 +01:00
Sarah Hoffmann
80e0a3cce4 change default rank for highway objects to 30
The highway key is being used more and more for non-ways these
days. This clashes with Nominatim's assumption that essentially
everything that has a highway tag can be used as the street part
of the address.

Change the default rank of highway objects to 30 to avoid this.
Only the known values for streets keep the rank 26 and are now
listed explicitly.
2021-11-24 22:10:40 +01:00
Sarah Hoffmann
79effae933 Merge pull request #2529 from lonvia/sort-street-results-by-tiger-housenumber
Take tiger housenumber into account when ranking street results
2021-11-24 16:23:41 +01:00
Sarah Hoffmann
810056349f add migration for inclusive housenumber Tiger index 2021-11-24 12:03:20 +01:00
Sarah Hoffmann
b1d490ea53 add index for Tiger housenumber queries 2021-11-24 11:10:20 +01:00
Sarah Hoffmann
345637290b take Tiger housenumbers into account when ranking street results
Queries with a housenumber need to rank streets higher that
have the requested housenumber attached. We already do that for
ordinary housenumber objects and for interpolations. This
adds support for Tiger housenumbers as well.

Fixes #2501.
2021-11-24 11:10:20 +01:00
Sarah Hoffmann
9628df3031 Merge pull request #2528 from lonvia/allow-french-extra-housenumbers
Don't penalize French 'bis' housenumbers
2021-11-21 10:53:20 +01:00
Sarah Hoffmann
423f338d04 Merge pull request #2526 from lonvia/docs-moving-database
Add a section about moving the database to another machine
2021-11-19 21:14:53 +01:00
Sarah Hoffmann
3a2597e5c4 don't penalize French 'bis' housenumbers
House numbers of the form '9 bis' are usual in France. So
be a bit more lenient before adding penalties to house numbers
with letters in them.

Fixes #2527.
2021-11-19 21:12:17 +01:00
Sarah Hoffmann
641f261495 Merge pull request #2525 from lonvia/fix-replication-indexer
Fix instantiation of indexer for replication
2021-11-19 16:16:30 +01:00
Sarah Hoffmann
5884a6e7a6 add a section about moving the database to another machine 2021-11-19 16:11:32 +01:00
Sarah Hoffmann
10e979e841 only instantiate indexer once for replication
Also makes sure that indexer object exists everywhere were needed.

See #2518.
2021-11-19 14:48:58 +01:00
Sarah Hoffmann
8dc1441635 Merge pull request #2517 from lonvia/transliteration-special-chars
ICU: avoid non-alphanumerical characters in transliteration
2021-11-11 07:42:42 +01:00
Sarah Hoffmann
c79dcfad9a make sure housenumbers are properly quoted 2021-11-10 20:44:28 +01:00
Sarah Hoffmann
1886952666 avoid special characters in word tokens
Transliteration should only consist of ASCII letters
and numbers. Avoid any other characters.
2021-11-10 17:14:13 +01:00
Sarah Hoffmann
7326b246b7 Merge pull request #2516 from lonvia/test-for-website-dir
Better error reporting when API script does not exist
2021-11-10 13:27:09 +01:00
Sarah Hoffmann
345c812e43 better error reporting when API script does not exist
Check if the API script exists on the expected location before
running php-cli. This way we can add a useful hint about the
project directory.

Fixes #2513.
2021-11-10 11:58:20 +01:00
Sarah Hoffmann
fd4ba3989e Merge pull request #2511 from lonvia/fix-combination-error-needs-address
Fix boolean combination of NeedsAddress flag
2021-11-06 12:11:55 +01:00
Sarah Hoffmann
e2d2571ad0 fix combination of NeedsAddress flag
When dealing with multiple partial terms, only keep the
flag, when all partial terms are so frequent as to need
an address.

Fixes #2510.
2021-11-05 22:18:37 +01:00
Sarah Hoffmann
d479a0585d prepare release 4.0.0 v4.0.0 2021-11-02 20:27:55 +01:00
Sarah Hoffmann
addfae31b6 fix typo 2021-11-02 11:09:17 +01:00
Sarah Hoffmann
ccf61db726 Merge pull request #2502 from lonvia/improve-development-documentation
Extend developer's documentation
2021-11-01 16:12:23 +01:00
Sarah Hoffmann
5b86b2078a docs: add overview over indexing 2021-11-01 11:04:03 +01:00
Sarah Hoffmann
a069479340 docs: section about database layout
Replaces the import description which basically was
table layout only now.
2021-10-29 12:03:22 +02:00
Sarah Hoffmann
d11bf9288e Merge pull request #2498 from lonvia/ordering-for-unlisted-place-results
Include unlisted places in ordering by housenumber
2021-10-28 15:28:47 +02:00
Sarah Hoffmann
86eeb4d2ed Merge pull request #2497 from lonvia/docs-maintenance
docs: add new maintenance section
2021-10-28 11:33:34 +02:00
Sarah Hoffmann
2275fe59ab include unlisted places in ordering by housenumber
When ordering results by the fact that they have a housenumber,
also take cases into account where the housenumber is on the
place itself. This may happen when the search includes the name
of the place and the housenumber or for addr:place addresses
where the place is unlisted.
2021-10-28 11:27:31 +02:00
Sarah Hoffmann
48be8c33ba docs: add new maintenance section
currently used for postcode updates, word count updates and
deleted relations.
2021-10-28 09:22:37 +02:00
Sarah Hoffmann
d3d07128b2 Merge pull request #2495 from lonvia/fix-normalization-in-php
ICU: use correct normalization during search
2021-10-27 14:40:42 +02:00
Sarah Hoffmann
37eeccbf4c ICU: use normalization from config in PHP
The TERM_NORMALIZATION config option is no longer applicable.
That was already documented but not yet implemented.
2021-10-27 11:32:44 +02:00
Sarah Hoffmann
1722fc537f bdd: add tests for non-latin scripts 2021-10-26 17:29:03 +02:00
Sarah Hoffmann
b240b182cb Merge pull request #2493 from lonvia/handle-frequent-partials
Tune search queries with frequent partial words
2021-10-26 17:00:43 +02:00
Sarah Hoffmann
c0f347fc8c adapt BDD tests to stricter partial search 2021-10-26 15:52:57 +02:00
Sarah Hoffmann
53dbe58ada do not count words when in reverse-only mode 2021-10-26 12:00:13 +02:00
Sarah Hoffmann
2c4b798f9b further refactor setup to keep function small 2021-10-26 12:00:13 +02:00
Sarah Hoffmann
1cf14a8e94 searches for house numbers must have an address 2021-10-26 12:00:13 +02:00
Sarah Hoffmann
4864bf1509 disallow search for partials without address
Very frequent partial terms take too long to look up and
do not return any valuable results unless the search is
further narrowed down by an address.
2021-10-26 12:00:13 +02:00
Sarah Hoffmann
9934421442 make word count computation part of the import
Accurate word counts are now essential when using
the ICU tokenizer and don't hurt for the legacy one.

Adds about an hour import time.
2021-10-26 12:00:13 +02:00
Sarah Hoffmann
d7267c1603 actions: move ICU tests into its own run 2021-10-26 11:59:13 +02:00
Sarah Hoffmann
5c778c6d32 Merge pull request #2486 from lonvia/fix-special-phrases
Fix parsing of operator in special phrases
2021-10-25 21:45:08 +02:00
Sarah Hoffmann
85797acf1e ICU: add an index over word_ids
Needed for keyword lookup in the details response.
2021-10-25 21:33:27 +02:00
Sarah Hoffmann
c4f5c11a4e be case-insensitve about special phrase operator 2021-10-25 19:51:20 +02:00
Sarah Hoffmann
5a1c3dbea3 fix parsing of operator in special phrases
Because of unstripped input, the operators wouldn't match.
2021-10-25 19:46:30 +02:00
Sarah Hoffmann
8e439d3dd9 Merge pull request #2484 from lonvia/fix-index-use
Reverse: add index hints
2021-10-25 17:20:42 +02:00
Sarah Hoffmann
9ebf921c53 Merge pull request #2483 from lonvia/fix-warming
Fix warming for ICU tokenizer
2021-10-25 16:21:36 +02:00
Sarah Hoffmann
7bd9094aaa reverse: add index hints
The fairly complex where condition of idx_placex_geometry_placenode
won't always be matched by the query planner if the condition
part doesn't appear verbatim in the query.

Fixes #2480.
2021-10-25 15:01:03 +02:00
Sarah Hoffmann
16cc395f78 fix warming for ICU tokenizer
Running the warm-up search requests requires querying
the most frequent words. This must be done via the tokenizer
to honor the different formats of the word table.
2021-10-25 13:08:16 +02:00
Sarah Hoffmann
13e7398566 allow relative paths for log files 2021-10-25 10:26:05 +02:00