reinit the tokenizer directory on website refresh

[nominatim.git] / nominatim / clicmd / refresh.py

diff --git a/nominatim/clicmd/refresh.py b/nominatim/clicmd/refresh.py
index fbc23350c7212478f8c8b0c9fa484ec154f9f070..3c245cd46e84b385a39288f537bb50813f8f284f 100644 (file)
--- a/nominatim/clicmd/refresh.py
+++ b/nominatim/clicmd/refresh.py
@@ -1,3 +1,9 @@
+# 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.
 """
 Implementation of 'refresh' subcommand.
 """
@@ -17,7 +23,13 @@ class UpdateRefresh:
     """\
     Recompute auxiliary data used by the indexing process.
 
-    These functions must not be run in parallel with other update commands.
+    This sub-commands updates various static data and functions in the database.
+    It usually needs to be run after changing various aspects of the
+    configuration. The configuration documentation will mention the exact
+    command to use in such case.
+
+    Warning: the 'update' command must not be run in parallel with other update
+             commands like 'replication' or 'add-data'.
     """
     def __init__(self):
         self.tokenizer = None
@@ -27,6 +39,8 @@ class UpdateRefresh:
         group = parser.add_argument_group('Data arguments')
         group.add_argument('--postcodes', action='store_true',
                            help='Update postcode centroid table')
+        group.add_argument('--word-tokens', action='store_true',
+                           help='Clean up search terms')
         group.add_argument('--word-counts', action='store_true',
                            help='Compute frequency of full-word search terms')
         group.add_argument('--address-levels', action='store_true',
@@ -34,7 +48,7 @@ class UpdateRefresh:
         group.add_argument('--functions', action='store_true',
                            help='Update the PL/pgSQL functions in the database')
         group.add_argument('--wiki-data', action='store_true',
-                           help='Update Wikipedia/data importance numbers.')
+                           help='Update Wikipedia/data importance numbers')
         group.add_argument('--importance', action='store_true',
                            help='Recompute place importances (expensive!)')
         group.add_argument('--website', action='store_true',
@@ -61,18 +75,22 @@ class UpdateRefresh:
                                   args.threads or 1)
                 indexer.index_postcodes()
             else:
-                LOG.error("The place table doesn\'t exist. " \
+                LOG.error("The place table doesn't exist. "
                           "Postcode updates on a frozen database is not possible.")
 
+        if args.word_tokens:
+            LOG.warning('Updating word tokens')
+            tokenizer = self._get_tokenizer(args.config)
+            tokenizer.update_word_tokens()
+
         if args.word_counts:
-            LOG.warning('Recompute frequency of full-word search terms')
-            refresh.recompute_word_counts(args.config.get_libpq_dsn(), args.sqllib_dir)
+            LOG.warning('Recompute word statistics')
+            self._get_tokenizer(args.config).update_statistics()
 
         if args.address_levels:
-            cfg = Path(args.config.ADDRESS_LEVEL_CONFIG)
-            LOG.warning('Updating address levels from %s', cfg)
+            LOG.warning('Updating address levels')
             with connect(args.config.get_libpq_dsn()) as conn:
-                refresh.load_address_levels_from_file(conn, cfg)
+                refresh.load_address_levels_from_config(conn, args.config)
 
         if args.functions:
             LOG.warning('Create functions')
@@ -99,6 +117,10 @@ class UpdateRefresh:
         if args.website:
             webdir = args.project_dir / 'website'
             LOG.warning('Setting up website directory at %s', webdir)
+            # This is a little bit hacky: call the tokenizer setup, so that
+            # the tokenizer directory gets repopulated as well, in case it
+            # wasn't there yet.
+            self._get_tokenizer(args.config)
             with connect(args.config.get_libpq_dsn()) as conn:
                 refresh.setup_website(webdir, args.config, conn)