Merge remote-tracking branch 'upstream/master'

[nominatim.git] / nominatim / api / core.py

diff --git a/nominatim/api/core.py b/nominatim/api/core.py
index 116a2625c76028a4e4d3de1d9358b82b509b3653..0f1dd7153af8f582410f398ab984d57ff8a18fb2 100644 (file)
--- a/nominatim/api/core.py
+++ b/nominatim/api/core.py
@@ -7,34 +7,43 @@
 """
 Implementation of classes for API access via libraries.
 """
 """
 Implementation of classes for API access via libraries.
 """

-from typing import Mapping, Optional, Any, AsyncIterator, Dict
+from typing import Mapping, Optional, Any, AsyncIterator, Dict, Sequence, List, Tuple

 import asyncio
 import asyncio

+import sys

 import contextlib
 from pathlib import Path
 
 import sqlalchemy as sa
 import sqlalchemy.ext.asyncio as sa_asyncio
 import contextlib
 from pathlib import Path
 
 import sqlalchemy as sa
 import sqlalchemy.ext.asyncio as sa_asyncio

-import asyncpg

 
 

+from nominatim.errors import UsageError

 from nominatim.db.sqlalchemy_schema import SearchTables
 from nominatim.db.sqlalchemy_schema import SearchTables

+from nominatim.db.async_core_library import PGCORE_LIB, PGCORE_ERROR

 from nominatim.config import Configuration
 from nominatim.api.connection import SearchConnection
 from nominatim.api.status import get_status, StatusResult
 from nominatim.config import Configuration
 from nominatim.api.connection import SearchConnection
 from nominatim.api.status import get_status, StatusResult

-from nominatim.api.lookup import get_place_by_id
+from nominatim.api.lookup import get_detailed_place, get_simple_place

 from nominatim.api.reverse import ReverseGeocoder
 from nominatim.api.reverse import ReverseGeocoder

-from nominatim.api.types import PlaceRef, LookupDetails, AnyPoint, DataLayer
-from nominatim.api.results import DetailedResult, ReverseResult
+from nominatim.api.search import ForwardGeocoder, Phrase, PhraseType, make_query_analyzer
+import nominatim.api.types as ntyp
+from nominatim.api.results import DetailedResult, ReverseResult, SearchResults

 
 
 class NominatimAPIAsync:
     """ API loader asynchornous version.
     """
     def __init__(self, project_dir: Path,
 
 
 class NominatimAPIAsync:
     """ API loader asynchornous version.
     """
     def __init__(self, project_dir: Path,

-                 environ: Optional[Mapping[str, str]] = None) -> None:
+                 environ: Optional[Mapping[str, str]] = None,
+                 loop: Optional[asyncio.AbstractEventLoop] = None) -> None:

         self.config = Configuration(project_dir, environ)
         self.config = Configuration(project_dir, environ)

+        self.query_timeout = self.config.get_int('QUERY_TIMEOUT') \
+                             if self.config.QUERY_TIMEOUT else None

         self.server_version = 0
 
         self.server_version = 0
 

-        self._engine_lock = asyncio.Lock()
+        if sys.version_info >= (3, 10):
+            self._engine_lock = asyncio.Lock()
+        else:
+            self._engine_lock = asyncio.Lock(loop=loop) # pylint: disable=unexpected-keyword-arg

         self._engine: Optional[sa_asyncio.AsyncEngine] = None
         self._tables: Optional[SearchTables] = None
         self._property_cache: Dict[str, Any] = {'DB:server_version': 0}
         self._engine: Optional[sa_asyncio.AsyncEngine] = None
         self._tables: Optional[SearchTables] = None
         self._property_cache: Dict[str, Any] = {'DB:server_version': 0}


@@ -52,26 +61,26 @@ class NominatimAPIAsync:
                 return
 
             dsn = self.config.get_database_params()
                 return
 
             dsn = self.config.get_database_params()

+            pool_size = self.config.get_int('API_POOL_SIZE')
+
+            query = {k: v for k, v in dsn.items()
+                      if k not in ('user', 'password', 'dbname', 'host', 'port')}

 
             dburl = sa.engine.URL.create(
 
             dburl = sa.engine.URL.create(

-                       'postgresql+asyncpg',
+                       f'postgresql+{PGCORE_LIB}',

                        database=dsn.get('dbname'),
                        username=dsn.get('user'), password=dsn.get('password'),
                        host=dsn.get('host'), port=int(dsn['port']) if 'port' in dsn else None,
                        database=dsn.get('dbname'),
                        username=dsn.get('user'), password=dsn.get('password'),
                        host=dsn.get('host'), port=int(dsn['port']) if 'port' in dsn else None,

-                       query={k: v for k, v in dsn.items()
-                              if k not in ('user', 'password', 'dbname', 'host', 'port')})
-            engine = sa_asyncio.create_async_engine(
-                             dburl, future=True,
-                             connect_args={'server_settings': {
-                                'DateStyle': 'sql,european',
-                                'max_parallel_workers_per_gather': '0'
-                             }})
+                       query=query)
+            engine = sa_asyncio.create_async_engine(dburl, future=True,
+                                                    max_overflow=0, pool_size=pool_size,
+                                                    echo=self.config.get_bool('DEBUG_SQL'))

 
             try:
                 async with engine.begin() as conn:
                     result = await conn.scalar(sa.text('SHOW server_version_num'))
                     server_version = int(result)
 
             try:
                 async with engine.begin() as conn:
                     result = await conn.scalar(sa.text('SHOW server_version_num'))
                     server_version = int(result)

-            except asyncpg.PostgresError:
+            except (PGCORE_ERROR, sa.exc.OperationalError):

                 server_version = 0
 
             if server_version >= 110000:
                 server_version = 0
 
             if server_version >= 110000:


@@ -79,6 +88,7 @@ class NominatimAPIAsync:
                 def _on_connect(dbapi_con: Any, _: Any) -> None:
                     cursor = dbapi_con.cursor()
                     cursor.execute("SET jit_above_cost TO '-1'")
                 def _on_connect(dbapi_con: Any, _: Any) -> None:
                     cursor = dbapi_con.cursor()
                     cursor.execute("SET jit_above_cost TO '-1'")

+                    cursor.execute("SET max_parallel_workers_per_gather TO '0'")

                 # Make sure that all connections get the new settings
                 await self.close()
 
                 # Make sure that all connections get the new settings
                 await self.close()
 


@@ -120,26 +130,42 @@ class NominatimAPIAsync:
         """
         try:
             async with self.begin() as conn:
         """
         try:
             async with self.begin() as conn:

+                conn.set_query_timeout(self.query_timeout)

                 status = await get_status(conn)
                 status = await get_status(conn)

-        except asyncpg.PostgresError:
+        except (PGCORE_ERROR, sa.exc.OperationalError):

             return StatusResult(700, 'Database connection failed')
 
         return status
 
 
             return StatusResult(700, 'Database connection failed')
 
         return status
 
 

-    async def lookup(self, place: PlaceRef,
-                     details: Optional[LookupDetails] = None) -> Optional[DetailedResult]:
+    async def details(self, place: ntyp.PlaceRef, **params: Any) -> Optional[DetailedResult]:

         """ Get detailed information about a place in the database.
 
             Returns None if there is no entry under the given ID.
         """
         """ Get detailed information about a place in the database.
 
             Returns None if there is no entry under the given ID.
         """

+        details = ntyp.LookupDetails.from_kwargs(params)
+        async with self.begin() as conn:
+            conn.set_query_timeout(self.query_timeout)
+            if details.keywords:
+                await make_query_analyzer(conn)
+            return await get_detailed_place(conn, place, details)
+
+
+    async def lookup(self, places: Sequence[ntyp.PlaceRef], **params: Any) -> SearchResults:
+        """ Get simple information about a list of places.
+
+            Returns a list of place information for all IDs that were found.
+        """
+        details = ntyp.LookupDetails.from_kwargs(params)

         async with self.begin() as conn:
         async with self.begin() as conn:

-            return await get_place_by_id(conn, place, details or LookupDetails())
+            conn.set_query_timeout(self.query_timeout)
+            if details.keywords:
+                await make_query_analyzer(conn)
+            return SearchResults(filter(None,
+                                        [await get_simple_place(conn, p, details) for p in places]))

 
 
 
 

-    async def reverse(self, coord: AnyPoint, max_rank: Optional[int] = None,
-                      layer: Optional[DataLayer] = None,
-                      details: Optional[LookupDetails] = None) -> Optional[ReverseResult]:
+    async def reverse(self, coord: ntyp.AnyPoint, **params: Any) -> Optional[ReverseResult]:

         """ Find a place by its coordinates. Also known as reverse geocoding.
 
             Returns the closest result that can be found or None if
         """ Find a place by its coordinates. Also known as reverse geocoding.
 
             Returns the closest result that can be found or None if


@@ -150,15 +176,115 @@ class NominatimAPIAsync:
             # There are no results to be expected outside valid coordinates.
             return None
 
             # There are no results to be expected outside valid coordinates.
             return None
 

-        if layer is None:
-            layer = DataLayer.ADDRESS | DataLayer.POI
+        details = ntyp.ReverseDetails.from_kwargs(params)
+        async with self.begin() as conn:
+            conn.set_query_timeout(self.query_timeout)
+            if details.keywords:
+                await make_query_analyzer(conn)
+            geocoder = ReverseGeocoder(conn, details)
+            return await geocoder.lookup(coord)
+

 
 

-        max_rank = max(0, min(max_rank or 30, 30))
+    async def search(self, query: str, **params: Any) -> SearchResults:
+        """ Find a place by free-text search. Also known as forward geocoding.
+        """
+        query = query.strip()
+        if not query:
+            raise UsageError('Nothing to search for.')

 
         async with self.begin() as conn:
 
         async with self.begin() as conn:

-            geocoder = ReverseGeocoder(conn, max_rank, layer,
-                                       details or LookupDetails())
-            return await geocoder.lookup(coord)
+            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(',')]
+            return await geocoder.lookup(phrases)
+
+
+    # pylint: disable=too-many-arguments,too-many-branches
+    async def search_address(self, amenity: Optional[str] = None,
+                             street: Optional[str] = None,
+                             city: Optional[str] = None,
+                             county: Optional[str] = None,
+                             state: Optional[str] = None,
+                             country: Optional[str] = None,
+                             postalcode: Optional[str] = None,
+                             **params: Any) -> SearchResults:
+        """ Find an address using structured search.
+        """
+        async with self.begin() as conn:
+            conn.set_query_timeout(self.query_timeout)
+            details = ntyp.SearchDetails.from_kwargs(params)
+
+            phrases: List[Phrase] = []
+
+            if amenity:
+                phrases.append(Phrase(PhraseType.AMENITY, amenity))
+            if street:
+                phrases.append(Phrase(PhraseType.STREET, street))
+            if city:
+                phrases.append(Phrase(PhraseType.CITY, city))
+            if county:
+                phrases.append(Phrase(PhraseType.COUNTY, county))
+            if state:
+                phrases.append(Phrase(PhraseType.STATE, state))
+            if postalcode:
+                phrases.append(Phrase(PhraseType.POSTCODE, postalcode))
+            if country:
+                phrases.append(Phrase(PhraseType.COUNTRY, country))
+
+            if not phrases:
+                raise UsageError('Nothing to search for.')
+
+            if amenity or street:
+                details.restrict_min_max_rank(26, 30)
+            elif city:
+                details.restrict_min_max_rank(13, 25)
+            elif county:
+                details.restrict_min_max_rank(10, 12)
+            elif state:
+                details.restrict_min_max_rank(5, 9)
+            elif postalcode:
+                details.restrict_min_max_rank(5, 11)
+            else:
+                details.restrict_min_max_rank(4, 4)
+
+            if 'layers' not in params:
+                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)
+            return await geocoder.lookup(phrases)
+
+
+    async def search_category(self, categories: List[Tuple[str, str]],
+                              near_query: Optional[str] = None,
+                              **params: Any) -> SearchResults:
+        """ Find an object of a certain category near another place.
+            The near place may either be given as an unstructured search
+            query in itself or as coordinates.
+        """
+        if not categories:
+            return SearchResults()
+
+        details = ntyp.SearchDetails.from_kwargs(params)
+        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(',')]
+            else:
+                phrases = []
+                if details.keywords:
+                    await make_query_analyzer(conn)
+
+            geocoder = ForwardGeocoder(conn, details,
+                                       self.config.get_int('REQUEST_TIMEOUT') \
+                                         if self.config.REQUEST_TIMEOUT else None)
+            return await geocoder.lookup_pois(categories, phrases)
+

 
 
 class NominatimAPI:
 
 
 class NominatimAPI:


@@ -168,7 +294,7 @@ class NominatimAPI:
     def __init__(self, project_dir: Path,
                  environ: Optional[Mapping[str, str]] = None) -> None:
         self._loop = asyncio.new_event_loop()
     def __init__(self, project_dir: Path,
                  environ: Optional[Mapping[str, str]] = None) -> None:
         self._loop = asyncio.new_event_loop()

-        self._async_api = NominatimAPIAsync(project_dir, environ)
+        self._async_api = NominatimAPIAsync(project_dir, environ, loop=self._loop)

 
 
     def close(self) -> None:
 
 
     def close(self) -> None:


@@ -192,20 +318,59 @@ class NominatimAPI:
         return self._loop.run_until_complete(self._async_api.status())
 
 
         return self._loop.run_until_complete(self._async_api.status())
 
 

-    def lookup(self, place: PlaceRef,
-               details: Optional[LookupDetails] = None) -> Optional[DetailedResult]:
+    def details(self, place: ntyp.PlaceRef, **params: Any) -> Optional[DetailedResult]:

         """ Get detailed information about a place in the database.
         """
         """ Get detailed information about a place in the database.
         """

-        return self._loop.run_until_complete(self._async_api.lookup(place, details))
+        return self._loop.run_until_complete(self._async_api.details(place, **params))
+
+
+    def lookup(self, places: Sequence[ntyp.PlaceRef], **params: Any) -> SearchResults:
+        """ Get simple information about a list of places.
+
+            Returns a list of place information for all IDs that were found.
+        """
+        return self._loop.run_until_complete(self._async_api.lookup(places, **params))

 
 
 
 

-    def reverse(self, coord: AnyPoint, max_rank: Optional[int] = None,
-                layer: Optional[DataLayer] = None,
-                details: Optional[LookupDetails] = None) -> Optional[ReverseResult]:
+    def reverse(self, coord: ntyp.AnyPoint, **params: Any) -> Optional[ReverseResult]:

         """ Find a place by its coordinates. Also known as reverse geocoding.
 
             Returns the closest result that can be found or None if
             no place matches the given criteria.
         """
         """ Find a place by its coordinates. Also known as reverse geocoding.
 
             Returns the closest result that can be found or None if
             no place matches the given criteria.
         """

+        return self._loop.run_until_complete(self._async_api.reverse(coord, **params))
+
+
+    def search(self, query: str, **params: Any) -> SearchResults:
+        """ Find a place by free-text search. Also known as forward geocoding.
+        """
+        return self._loop.run_until_complete(
+                   self._async_api.search(query, **params))
+
+
+    # pylint: disable=too-many-arguments
+    def search_address(self, amenity: Optional[str] = None,
+                       street: Optional[str] = None,
+                       city: Optional[str] = None,
+                       county: Optional[str] = None,
+                       state: Optional[str] = None,
+                       country: Optional[str] = None,
+                       postalcode: Optional[str] = None,
+                       **params: Any) -> SearchResults:
+        """ Find an address using structured search.
+        """
+        return self._loop.run_until_complete(
+                   self._async_api.search_address(amenity, street, city, county,
+                                                  state, country, postalcode, **params))
+
+
+    def search_category(self, categories: List[Tuple[str, str]],
+                        near_query: Optional[str] = None,
+                        **params: Any) -> SearchResults:
+        """ Find an object of a certain category near another place.
+            The near place may either be given as an unstructured search
+            query in itself or as a geographic area through the
+            viewbox or near parameters.
+        """

         return self._loop.run_until_complete(
         return self._loop.run_until_complete(

-                   self._async_api.reverse(coord, max_rank, layer, details))
+                   self._async_api.search_category(categories, near_query, **params))