mirror of
https://github.com/osm-search/Nominatim.git
synced 2026-03-07 18:44:07 +00:00
add server fronting for search endpoint
This also implements some of the quirks of free-text search of the V1 API, in particular, search for categories and coordinates.
This commit is contained in:
Sarah Hoffmann
@@ -228,6 +228,12 @@ class SearchResults(List[SearchResult]):
|
||||
May be empty when no result was found.
|
||||
"""
|
||||
|
||||
def localize(self, locales: Locales) -> None:
|
||||
""" Apply the given locales to all results.
|
||||
"""
|
||||
for result in self:
|
||||
result.localize(locales)
|
||||
|
||||
|
||||
def _filter_geometries(row: SaRow) -> Dict[str, str]:
|
||||
return {k[9:]: v for k, v in row._mapping.items() # pylint: disable=W0212
|
||||
|
||||
@@ -168,7 +168,7 @@ def _format_details_json(result: napi.DetailedResult, options: Mapping[str, Any]
|
||||
def _format_reverse_xml(results: napi.ReverseResults, options: Mapping[str, Any]) -> str:
|
||||
return format_xml.format_base_xml(results,
|
||||
options, True, 'reversegeocode',
|
||||
{'querystring': 'TODO'})
|
||||
{'querystring': options.get('query', '')})
|
||||
|
||||
|
||||
@dispatch.format_func(napi.ReverseResults, 'geojson')
|
||||
@@ -199,9 +199,13 @@ def _format_reverse_jsonv2(results: napi.ReverseResults,
|
||||
|
||||
@dispatch.format_func(napi.SearchResults, 'xml')
|
||||
def _format_search_xml(results: napi.SearchResults, options: Mapping[str, Any]) -> str:
|
||||
return format_xml.format_base_xml(results,
|
||||
options, False, 'searchresults',
|
||||
{'querystring': 'TODO'})
|
||||
extra = {'querystring': options.get('query', '')}
|
||||
for attr in ('more_url', 'exclude_place_ids', 'viewbox'):
|
||||
if options.get(attr):
|
||||
extra[attr] = options[attr]
|
||||
return format_xml.format_base_xml(results, options, False, 'searchresults',
|
||||
extra)
|
||||
|
||||
|
||||
|
||||
@dispatch.format_func(napi.SearchResults, 'geojson')
|
||||
|
||||
@@ -8,8 +8,12 @@
|
||||
Helper function for parsing parameters and and outputting data
|
||||
specifically for the v1 version of the API.
|
||||
"""
|
||||
from typing import Tuple, Optional, Any, Dict, Iterable
|
||||
from itertools import chain
|
||||
import re
|
||||
|
||||
from nominatim.api.results import SearchResult, SearchResults, SourceTable
|
||||
from nominatim.api.types import SearchDetails, GeometryFormat
|
||||
|
||||
REVERSE_MAX_RANKS = [2, 2, 2, # 0-2 Continent/Sea
|
||||
4, 4, # 3-4 Country
|
||||
@@ -33,6 +37,58 @@ def zoom_to_rank(zoom: int) -> int:
|
||||
return REVERSE_MAX_RANKS[max(0, min(18, zoom))]
|
||||
|
||||
|
||||
FEATURE_TYPE_TO_RANK: Dict[Optional[str], Any] = {
|
||||
'country': (4, 4),
|
||||
'state': (8, 8),
|
||||
'city': (14, 16),
|
||||
'settlement': (8, 20)
|
||||
}
|
||||
|
||||
|
||||
def feature_type_to_rank(feature_type: Optional[str]) -> Tuple[int, int]:
|
||||
""" Convert a feature type parameter to a tuple of
|
||||
feature type name, minimum rank and maximum rank.
|
||||
"""
|
||||
return FEATURE_TYPE_TO_RANK.get(feature_type, (0, 30))
|
||||
|
||||
|
||||
#pylint: disable=too-many-arguments
|
||||
def extend_query_parts(queryparts: dict[str, Any], details: dict[str, Any],
|
||||
feature_type: Optional[str],
|
||||
namedetails: bool, extratags: bool,
|
||||
excluded: Iterable[str]) -> None:
|
||||
""" Add parameters from details dictionary to the query parts
|
||||
dictionary which is suitable as URL parameter dictionary.
|
||||
"""
|
||||
parsed = SearchDetails.from_kwargs(details)
|
||||
if parsed.geometry_output != GeometryFormat.NONE:
|
||||
for flag in parsed.geometry_output:
|
||||
assert flag.name
|
||||
queryparts[f'polygon_{flag.name.lower()}'] = '1'
|
||||
if parsed.address_details:
|
||||
queryparts['addressdetails'] = '1'
|
||||
if namedetails:
|
||||
queryparts['namedetails'] = '1'
|
||||
if extratags:
|
||||
queryparts['extratags'] = '1'
|
||||
if parsed.geometry_simplification > 0.0:
|
||||
queryparts['polygon_threshold'] = f"{parsed.geometry_simplification:.6g}"
|
||||
if parsed.max_results != 10:
|
||||
queryparts['limit'] = str(parsed.max_results)
|
||||
if parsed.countries:
|
||||
queryparts['countrycodes'] = ','.join(parsed.countries)
|
||||
queryparts['exclude_place_ids'] = \
|
||||
','.join(chain(excluded, map(str, parsed.excluded)))
|
||||
if parsed.viewbox:
|
||||
queryparts['viewbox'] = ','.join(f"{c:.7g}" for c in parsed.viewbox.coords)
|
||||
if parsed.bounded_viewbox:
|
||||
queryparts['bounded'] = '1'
|
||||
if not details['dedupe']:
|
||||
queryparts['dedupe'] = '0'
|
||||
if feature_type in FEATURE_TYPE_TO_RANK:
|
||||
queryparts['featureType'] = feature_type
|
||||
|
||||
|
||||
def deduplicate_results(results: SearchResults, max_results: int) -> SearchResults:
|
||||
""" Remove results that look like duplicates.
|
||||
|
||||
@@ -69,3 +125,69 @@ def _is_postcode_relation_for(result: SearchResult, postcode: str) -> bool:
|
||||
and result.category == ('boundary', 'postal_code') \
|
||||
and result.names is not None \
|
||||
and result.names.get('ref') == postcode
|
||||
|
||||
|
||||
def _deg(axis:str) -> str:
|
||||
return f"(?P<{axis}_deg>\\d+\\.\\d+)°?"
|
||||
|
||||
def _deg_min(axis: str) -> str:
|
||||
return f"(?P<{axis}_deg>\\d+)[°\\s]+(?P<{axis}_min>[\\d.]+)?[′']*"
|
||||
|
||||
def _deg_min_sec(axis: str) -> str:
|
||||
return f"(?P<{axis}_deg>\\d+)[°\\s]+(?P<{axis}_min>\\d+)[′'\\s]+(?P<{axis}_sec>[\\d.]+)?[\"″]*"
|
||||
|
||||
COORD_REGEX = [re.compile(r'(?:(?P<pre>.*?)\s+)??' + r + r'(?:\s+(?P<post>.*))?') for r in (
|
||||
r"(?P<ns>[NS])\s*" + _deg('lat') + r"[\s,]+" + r"(?P<ew>[EW])\s*" + _deg('lon'),
|
||||
_deg('lat') + r"\s*(?P<ns>[NS])[\s,]+" + _deg('lon') + r"\s*(?P<ew>[EW])",
|
||||
r"(?P<ns>[NS])\s*" + _deg_min('lat') + r"[\s,]+" + r"(?P<ew>[EW])\s*" + _deg_min('lon'),
|
||||
_deg_min('lat') + r"\s*(?P<ns>[NS])[\s,]+" + _deg_min('lon') + r"\s*(?P<ew>[EW])",
|
||||
r"(?P<ns>[NS])\s*" + _deg_min_sec('lat') + r"[\s,]+" + r"(?P<ew>[EW])\s*" + _deg_min_sec('lon'),
|
||||
_deg_min_sec('lat') + r"\s*(?P<ns>[NS])[\s,]+" + _deg_min_sec('lon') + r"\s*(?P<ew>[EW])",
|
||||
r"\[?(?P<lat_deg>[+-]?\d+\.\d+)[\s,]+(?P<lon_deg>[+-]?\d+\.\d+)\]?"
|
||||
)]
|
||||
|
||||
def extract_coords_from_query(query: str) -> Tuple[str, Optional[float], Optional[float]]:
|
||||
""" Look for something that is formated like a coordinate at the
|
||||
beginning or end of the query. If found, extract the coordinate and
|
||||
return the remaining query (or the empty string if the query
|
||||
consisted of nothing but a coordinate).
|
||||
|
||||
Only the first match will be returned.
|
||||
"""
|
||||
for regex in COORD_REGEX:
|
||||
match = regex.fullmatch(query)
|
||||
if match is None:
|
||||
continue
|
||||
groups = match.groupdict()
|
||||
if not groups['pre'] or not groups['post']:
|
||||
x = float(groups['lon_deg']) \
|
||||
+ float(groups.get('lon_min', 0.0)) / 60.0 \
|
||||
+ float(groups.get('lon_sec', 0.0)) / 3600.0
|
||||
if groups.get('ew') == 'W':
|
||||
x = -x
|
||||
y = float(groups['lat_deg']) \
|
||||
+ float(groups.get('lat_min', 0.0)) / 60.0 \
|
||||
+ float(groups.get('lat_sec', 0.0)) / 3600.0
|
||||
if groups.get('ns') == 'S':
|
||||
y = -y
|
||||
return groups['pre'] or groups['post'] or '', x, y
|
||||
|
||||
return query, None, None
|
||||
|
||||
|
||||
CATEGORY_REGEX = re.compile(r'(?P<pre>.*?)\[(?P<cls>[a-zA-Z_]+)=(?P<typ>[a-zA-Z_]+)\](?P<post>.*)')
|
||||
|
||||
def extract_category_from_query(query: str) -> Tuple[str, Optional[str], Optional[str]]:
|
||||
""" Extract a hidden category specification of the form '[key=value]' from
|
||||
the query. If found, extract key and value and
|
||||
return the remaining query (or the empty string if the query
|
||||
consisted of nothing but a category).
|
||||
|
||||
Only the first match will be returned.
|
||||
"""
|
||||
match = CATEGORY_REGEX.search(query)
|
||||
if match is not None:
|
||||
return (match.group('pre').strip() + ' ' + match.group('post').strip()).strip(), \
|
||||
match.group('cls'), match.group('typ')
|
||||
|
||||
return query, None, None
|
||||
|
||||
@@ -11,8 +11,11 @@ Combine with the scaffolding provided for the various Python ASGI frameworks.
|
||||
from typing import Optional, Any, Type, Callable, NoReturn, Dict, cast
|
||||
from functools import reduce
|
||||
import abc
|
||||
import dataclasses
|
||||
import math
|
||||
from urllib.parse import urlencode
|
||||
|
||||
from nominatim.errors import UsageError
|
||||
from nominatim.config import Configuration
|
||||
import nominatim.api as napi
|
||||
import nominatim.api.logging as loglib
|
||||
@@ -321,7 +324,6 @@ async def reverse_endpoint(api: napi.NominatimAPIAsync, params: ASGIAdaptor) ->
|
||||
fmt = params.parse_format(napi.ReverseResults, 'xml')
|
||||
debug = params.setup_debugging()
|
||||
coord = napi.Point(params.get_float('lon'), params.get_float('lat'))
|
||||
locales = napi.Locales.from_accept_languages(params.get_accepted_languages())
|
||||
|
||||
details = params.parse_geometry_details(fmt)
|
||||
details['max_rank'] = helpers.zoom_to_rank(params.get_int('zoom', 18))
|
||||
@@ -332,12 +334,22 @@ async def reverse_endpoint(api: napi.NominatimAPIAsync, params: ASGIAdaptor) ->
|
||||
if debug:
|
||||
return params.build_response(loglib.get_and_disable())
|
||||
|
||||
fmt_options = {'extratags': params.get_bool('extratags', False),
|
||||
if fmt == 'xml':
|
||||
queryparts = {'lat': str(coord.lat), 'lon': str(coord.lon), 'format': 'xml'}
|
||||
zoom = params.get('zoom', None)
|
||||
if zoom:
|
||||
queryparts['zoom'] = zoom
|
||||
query = urlencode(queryparts)
|
||||
else:
|
||||
query = ''
|
||||
|
||||
fmt_options = {'query': query,
|
||||
'extratags': params.get_bool('extratags', False),
|
||||
'namedetails': params.get_bool('namedetails', False),
|
||||
'addressdetails': params.get_bool('addressdetails', True)}
|
||||
|
||||
if result:
|
||||
result.localize(locales)
|
||||
result.localize(napi.Locales.from_accept_languages(params.get_accepted_languages()))
|
||||
|
||||
output = formatting.format_result(napi.ReverseResults([result] if result else []),
|
||||
fmt, fmt_options)
|
||||
@@ -350,7 +362,6 @@ async def lookup_endpoint(api: napi.NominatimAPIAsync, params: ASGIAdaptor) -> A
|
||||
"""
|
||||
fmt = params.parse_format(napi.SearchResults, 'xml')
|
||||
debug = params.setup_debugging()
|
||||
locales = napi.Locales.from_accept_languages(params.get_accepted_languages())
|
||||
details = params.parse_geometry_details(fmt)
|
||||
|
||||
places = []
|
||||
@@ -371,18 +382,119 @@ async def lookup_endpoint(api: napi.NominatimAPIAsync, params: ASGIAdaptor) -> A
|
||||
'namedetails': params.get_bool('namedetails', False),
|
||||
'addressdetails': params.get_bool('addressdetails', True)}
|
||||
|
||||
for result in results:
|
||||
result.localize(locales)
|
||||
results.localize(napi.Locales.from_accept_languages(params.get_accepted_languages()))
|
||||
|
||||
output = formatting.format_result(results, fmt, fmt_options)
|
||||
|
||||
return params.build_response(output)
|
||||
|
||||
|
||||
async def _unstructured_search(query: str, api: napi.NominatimAPIAsync,
|
||||
details: Dict[str, Any]) -> napi.SearchResults:
|
||||
if not query:
|
||||
return napi.SearchResults()
|
||||
|
||||
# Extract special format for coordinates from query.
|
||||
query, x, y = helpers.extract_coords_from_query(query)
|
||||
if x is not None:
|
||||
assert y is not None
|
||||
details['near'] = napi.Point(x, y)
|
||||
details['near_radius'] = 0.1
|
||||
|
||||
# If no query is left, revert to reverse search.
|
||||
if x is not None and not query:
|
||||
result = await api.reverse(details['near'], **details)
|
||||
if not result:
|
||||
return napi.SearchResults()
|
||||
|
||||
return napi.SearchResults(
|
||||
[napi.SearchResult(**{f.name: getattr(result, f.name)
|
||||
for f in dataclasses.fields(napi.SearchResult)
|
||||
if hasattr(result, f.name)})])
|
||||
|
||||
query, cls, typ = helpers.extract_category_from_query(query)
|
||||
if cls is not None:
|
||||
assert typ is not None
|
||||
return await api.search_category([(cls, typ)], near_query=query, **details)
|
||||
|
||||
return await api.search(query, **details)
|
||||
|
||||
|
||||
async def search_endpoint(api: napi.NominatimAPIAsync, params: ASGIAdaptor) -> Any:
|
||||
""" Server glue for /search endpoint. See API docs for details.
|
||||
"""
|
||||
fmt = params.parse_format(napi.SearchResults, 'jsonv2')
|
||||
debug = params.setup_debugging()
|
||||
details = params.parse_geometry_details(fmt)
|
||||
|
||||
details['countries'] = params.get('countrycodes', None)
|
||||
details['excluded'] = params.get('exclude_place_ids', None)
|
||||
details['viewbox'] = params.get('viewbox', None) or params.get('viewboxlbrt', None)
|
||||
details['bounded_viewbox'] = params.get_bool('bounded', False)
|
||||
details['dedupe'] = params.get_bool('dedupe', True)
|
||||
|
||||
max_results = max(1, min(50, params.get_int('limit', 10)))
|
||||
details['max_results'] = max_results + min(10, max_results) \
|
||||
if details['dedupe'] else max_results
|
||||
|
||||
details['min_rank'], details['max_rank'] = \
|
||||
helpers.feature_type_to_rank(params.get('featureType', ''))
|
||||
|
||||
query = params.get('q', None)
|
||||
queryparts = {}
|
||||
try:
|
||||
if query is not None:
|
||||
queryparts['q'] = query
|
||||
results = await _unstructured_search(query, api, details)
|
||||
else:
|
||||
for key in ('amenity', 'street', 'city', 'county', 'state', 'postalcode', 'country'):
|
||||
details[key] = params.get(key, None)
|
||||
if details[key]:
|
||||
queryparts[key] = details[key]
|
||||
query = ', '.join(queryparts.values())
|
||||
|
||||
results = await api.search_address(**details)
|
||||
except UsageError as err:
|
||||
params.raise_error(str(err))
|
||||
|
||||
results.localize(napi.Locales.from_accept_languages(params.get_accepted_languages()))
|
||||
|
||||
if details['dedupe'] and len(results) > 1:
|
||||
results = helpers.deduplicate_results(results, max_results)
|
||||
|
||||
if debug:
|
||||
return params.build_response(loglib.get_and_disable())
|
||||
|
||||
if fmt == 'xml':
|
||||
helpers.extend_query_parts(queryparts, details,
|
||||
params.get('featureType', ''),
|
||||
params.get_bool('namedetails', False),
|
||||
params.get_bool('extratags', False),
|
||||
(str(r.place_id) for r in results if r.place_id))
|
||||
queryparts['format'] = fmt
|
||||
|
||||
moreurl = urlencode(queryparts)
|
||||
else:
|
||||
moreurl = ''
|
||||
|
||||
fmt_options = {'query': query, 'more_url': moreurl,
|
||||
'exclude_place_ids': queryparts.get('exclude_place_ids'),
|
||||
'viewbox': queryparts.get('viewbox'),
|
||||
'extratags': params.get_bool('extratags', False),
|
||||
'namedetails': params.get_bool('namedetails', False),
|
||||
'addressdetails': params.get_bool('addressdetails', False)}
|
||||
|
||||
output = formatting.format_result(results, fmt, fmt_options)
|
||||
|
||||
return params.build_response(output)
|
||||
|
||||
|
||||
EndpointFunc = Callable[[napi.NominatimAPIAsync, ASGIAdaptor], Any]
|
||||
|
||||
ROUTES = [
|
||||
('status', status_endpoint),
|
||||
('details', details_endpoint),
|
||||
('reverse', reverse_endpoint),
|
||||
('lookup', lookup_endpoint)
|
||||
('lookup', lookup_endpoint),
|
||||
('search', search_endpoint)
|
||||
]
|
||||
|
||||
Reference in New Issue
Block a user