Merge pull request #2761 from lonvia/repair-index-analysis

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

diff --git a/nominatim/tools/admin.py b/nominatim/tools/admin.py
index 2e18cb6f854b2c06698d225f377bcbbc5982e241..1bf217e2d81e77fd2526f3f3b3b98ec03a236a05 100644 (file)
--- a/nominatim/tools/admin.py
+++ b/nominatim/tools/admin.py
@@ -1,49 +1,83 @@
+# 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 analysis and maintenance.
 """
 import logging
 
+from psycopg2.extras import Json, register_hstore
+
+from nominatim.db.connection import connect
+from nominatim.tokenizer import factory as tokenizer_factory
 from nominatim.errors import UsageError
+from nominatim.data.place_info import PlaceInfo
 
 LOG = logging.getLogger()
 
-def analyse_indexing(conn, osm_id=None, place_id=None):
+def _get_place_info(cursor, osm_id, place_id):
+    sql = """SELECT place_id, extra.*
+             FROM placex, LATERAL placex_indexing_prepare(placex) as extra
+          """
+
+    if osm_id:
+        osm_type = osm_id[0].upper()
+        if osm_type not in 'NWR' or not osm_id[1:].isdigit():
+            LOG.fatal('OSM ID must be of form <N|W|R><id>. Got: %s', osm_id)
+            raise UsageError("OSM ID parameter badly formatted")
+
+        sql += ' WHERE placex.osm_type = %s AND placex.osm_id = %s'
+        values = (osm_type, int(osm_id[1:]))
+    elif place_id is not None:
+        sql += ' WHERE placex.place_id = %s'
+        values = (place_id, )
+    else:
+        LOG.fatal("No OSM object given to index.")
+        raise UsageError("OSM object not found")
+
+    cursor.execute(sql + ' LIMIT 1', values)
+
+    if cursor.rowcount < 1:
+        LOG.fatal("OSM object %s not found in database.", osm_id)
+        raise UsageError("OSM object not found")
+
+    return cursor.fetchone()
+
+
+def analyse_indexing(config, osm_id=None, place_id=None):
     """ Analyse indexing of a single Nominatim object.
     """
-    with conn.cursor() as cur:
-        if osm_id:
-            osm_type = osm_id[0].upper()
-            if osm_type not in 'NWR' or not osm_id[1:].isdigit():
-                LOG.fatal('OSM ID must be of form <N|W|R><id>. Got: %s', osm_id)
-                raise UsageError("OSM ID parameter badly formatted")
-            cur.execute('SELECT place_id FROM placex WHERE osm_type = %s AND osm_id = %s',
-                        (osm_type, osm_id[1:]))
-
-            if cur.rowcount < 1:
-                LOG.fatal("OSM object %s not found in database.", osm_id)
-                raise UsageError("OSM object not found")
-
-            place_id = cur.fetchone()[0]
-
-        if place_id is None:
-            LOG.fatal("No OSM object given to index.")
-            raise UsageError("OSM object not found")
-
-        cur.execute("update placex set indexed_status = 2 where place_id = %s",
-                    (place_id, ))
-
-        cur.execute("""SET auto_explain.log_min_duration = '0';
-                       SET auto_explain.log_analyze = 'true';
-                       SET auto_explain.log_nested_statements = 'true';
-                       LOAD 'auto_explain';
-                       SET client_min_messages = LOG;
-                       SET log_min_messages = FATAL""")
-
-        cur.execute("update placex set indexed_status = 0 where place_id = %s",
-                    (place_id, ))
-
-    # we do not want to keep the results
-    conn.rollback()
-
-    for msg in conn.notices:
-        print(msg)
+    with connect(config.get_libpq_dsn()) as conn:
+        register_hstore(conn)
+        with conn.cursor() as cur:
+            place = _get_place_info(cur, osm_id, place_id)
+
+            cur.execute("update placex set indexed_status = 2 where place_id = %s",
+                        (place['place_id'], ))
+
+            cur.execute("""SET auto_explain.log_min_duration = '0';
+                           SET auto_explain.log_analyze = 'true';
+                           SET auto_explain.log_nested_statements = 'true';
+                           LOAD 'auto_explain';
+                           SET client_min_messages = LOG;
+                           SET log_min_messages = FATAL""")
+
+            tokenizer = tokenizer_factory.get_tokenizer_for_db(config)
+
+            with tokenizer.name_analyzer() as analyzer:
+                cur.execute("""UPDATE placex
+                               SET indexed_status = 0, address = %s, token_info = %s,
+                               name = %s, linked_place_id = %s
+                               WHERE place_id = %s""",
+                            (place['address'],
+                             Json(analyzer.process_place(PlaceInfo(place))),
+                             place['name'], place['linked_place_id'], place['place_id']))
+
+        # we do not want to keep the results
+        conn.rollback()
+
+        for msg in conn.notices:
+            print(msg)