optimize indexes for interpolation lines

[nominatim.git] / nominatim / tools / migration.py

diff --git a/nominatim/tools/migration.py b/nominatim/tools/migration.py
index 756a1d3a903326a8540dcbee9ee3b545bb11d843..94058f334e94ef9954e0f6ea15a0bbbafd875d9f 100644 (file)
--- a/nominatim/tools/migration.py
+++ b/nominatim/tools/migration.py
@@ -1,13 +1,20 @@
+# 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 database migration to newer software versions.
 """
 import logging
 
 """
 Functions for database migration to newer software versions.
 """
 import logging
 

-from ..db import properties
-from ..db.connection import connect
-from ..version import NOMINATIM_VERSION
-from . import refresh, database_import
-from ..errors import UsageError
+from nominatim.db import properties
+from nominatim.db.connection import connect
+from nominatim.version import NOMINATIM_VERSION
+from nominatim.tools import refresh
+from nominatim.tokenizer import factory as tokenizer_factory
+from nominatim.errors import UsageError

 
 LOG = logging.getLogger()
 
 
 LOG = logging.getLogger()
 


@@ -25,7 +32,7 @@ def migrate(config, paths):
 
         if db_version_str is not None:
             parts = db_version_str.split('.')
 
         if db_version_str is not None:
             parts = db_version_str.split('.')

-            db_version = tuple([int(x) for x in parts[:2] + parts[2].split('-')])
+            db_version = tuple(int(x) for x in parts[:2] + parts[2].split('-'))

 
             if db_version == NOMINATIM_VERSION:
                 LOG.warning("Database already at latest version (%s)", db_version_str)
 
             if db_version == NOMINATIM_VERSION:
                 LOG.warning("Database already at latest version (%s)", db_version_str)


@@ -43,11 +50,14 @@ def migrate(config, paths):
                             '{0[0]}.{0[1]}.{0[2]}-{0[3]}'.format(version))
                 kwargs = dict(conn=conn, config=config, paths=paths)
                 func(**kwargs)
                             '{0[0]}.{0[1]}.{0[2]}-{0[3]}'.format(version))
                 kwargs = dict(conn=conn, config=config, paths=paths)
                 func(**kwargs)

+                conn.commit()

                 has_run_migration = True
 
         if has_run_migration:
             LOG.warning('Updating SQL functions.')
                 has_run_migration = True
 
         if has_run_migration:
             LOG.warning('Updating SQL functions.')

-            refresh.create_functions(conn, config, paths.sqllib_dir)
+            refresh.create_functions(conn, config)
+            tokenizer = tokenizer_factory.get_tokenizer_for_db(config)
+            tokenizer.update_sql_functions(config)

 
         properties.set_property(conn, 'database_version',
                                 '{0[0]}.{0[1]}.{0[2]}-{0[3]}'.format(NOMINATIM_VERSION))
 
         properties.set_property(conn, 'database_version',
                                 '{0[0]}.{0[1]}.{0[2]}-{0[3]}'.format(NOMINATIM_VERSION))


@@ -92,6 +102,7 @@ def _migration(major, minor, patch=0, dbpatch=0):
     """
     def decorator(func):
         _MIGRATION_FUNCTIONS.append(((major, minor, patch, dbpatch), func))
     """
     def decorator(func):
         _MIGRATION_FUNCTIONS.append(((major, minor, patch, dbpatch), func))

+        return func

 
     return decorator
 
 
     return decorator
 


@@ -108,17 +119,6 @@ def import_status_timestamp_change(conn, **_):
                        TYPE timestamp with time zone;""")
 
 
                        TYPE timestamp with time zone;""")
 
 

-@_migration(3, 5, 0, 99)
-def install_database_module_in_project_directory(conn, config, paths, **_):
-    """ Install database module in project directory.
-
-        The database module needs to be present in the project directory
-        since those were introduced.
-    """
-    database_import.install_module(paths.module_dir, paths.project_dir,
-                                   config.DATABASE_MODULE_PATH, conn=conn)
-
-

 @_migration(3, 5, 0, 99)
 def add_nominatim_property_table(conn, config, **_):
     """ Add nominatim_property table.
 @_migration(3, 5, 0, 99)
 def add_nominatim_property_table(conn, config, **_):
     """ Add nominatim_property table.


@@ -130,3 +130,89 @@ def add_nominatim_property_table(conn, config, **_):
                                value TEXT);
                            GRANT SELECT ON TABLE nominatim_properties TO "{}";
                         """.format(config.DATABASE_WEBUSER))
                                value TEXT);
                            GRANT SELECT ON TABLE nominatim_properties TO "{}";
                         """.format(config.DATABASE_WEBUSER))

+
+@_migration(3, 6, 0, 0)
+def change_housenumber_transliteration(conn, **_):
+    """ Transliterate housenumbers.
+
+        The database schema switched from saving raw housenumbers in
+        placex.housenumber to saving transliterated ones.
+
+        Note: the function create_housenumber_id() has been dropped in later
+              versions.
+    """
+    with conn.cursor() as cur:
+        cur.execute("""CREATE OR REPLACE FUNCTION create_housenumber_id(housenumber TEXT)
+                       RETURNS TEXT AS $$
+                       DECLARE
+                         normtext TEXT;
+                       BEGIN
+                         SELECT array_to_string(array_agg(trans), ';')
+                           INTO normtext
+                           FROM (SELECT lookup_word as trans,
+                                        getorcreate_housenumber_id(lookup_word)
+                                 FROM (SELECT make_standard_name(h) as lookup_word
+                                       FROM regexp_split_to_table(housenumber, '[,;]') h) x) y;
+                         return normtext;
+                       END;
+                       $$ LANGUAGE plpgsql STABLE STRICT;""")
+        cur.execute("DELETE FROM word WHERE class = 'place' and type = 'house'")
+        cur.execute("""UPDATE placex
+                       SET housenumber = create_housenumber_id(housenumber)
+                       WHERE housenumber is not null""")
+
+
+@_migration(3, 7, 0, 0)
+def switch_placenode_geometry_index(conn, **_):
+    """ Replace idx_placex_geometry_reverse_placeNode index.
+
+        Make the index slightly more permissive, so that it can also be used
+        when matching up boundaries and place nodes. It makes the index
+        idx_placex_adminname index unnecessary.
+    """
+    with conn.cursor() as cur:
+        cur.execute(""" CREATE INDEX IF NOT EXISTS idx_placex_geometry_placenode ON placex
+                        USING GIST (geometry)
+                        WHERE osm_type = 'N' and rank_search < 26
+                              and class = 'place' and type != 'postcode'
+                              and linked_place_id is null""")
+        cur.execute(""" DROP INDEX IF EXISTS idx_placex_adminname """)
+
+
+@_migration(3, 7, 0, 1)
+def install_legacy_tokenizer(conn, config, **_):
+    """ Setup legacy tokenizer.
+
+        If no other tokenizer has been configured yet, then create the
+        configuration for the backwards-compatible legacy tokenizer
+    """
+    if properties.get_property(conn, 'tokenizer') is None:
+        with conn.cursor() as cur:
+            for table in ('placex', 'location_property_osmline'):
+                has_column = cur.scalar("""SELECT count(*) FROM information_schema.columns
+                                           WHERE table_name = %s
+                                           and column_name = 'token_info'""",
+                                        (table, ))
+                if has_column == 0:
+                    cur.execute('ALTER TABLE {} ADD COLUMN token_info JSONB'.format(table))
+        tokenizer = tokenizer_factory.create_tokenizer(config, init_db=False,
+                                                       module_name='legacy')
+
+        tokenizer.migrate_database(config)
+
+
+@_migration(4, 0, 99, 0)
+def create_tiger_housenumber_index(conn, **_):
+    """ Create idx_location_property_tiger_parent_place_id with included
+        house number.
+
+        The inclusion is needed for efficient lookup of housenumbers in
+        full address searches.
+    """
+    if conn.server_version_tuple() >= (11, 0, 0):
+        with conn.cursor() as cur:
+            cur.execute(""" CREATE INDEX IF NOT EXISTS
+                                idx_location_property_tiger_housenumber_migrated
+                            ON location_property_tiger
+                            USING btree(parent_place_id)
+                            INCLUDE (startnumber, endnumber) """)