Merge remote-tracking branch 'upstream/master'

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

diff --git a/nominatim/clicmd/special_phrases.py b/nominatim/clicmd/special_phrases.py
index 66fb41ed4557fedbf972d58b84139c3f450744d7..a2c346deb84edd38ce58011d190ee8ab735f7e9f 100644 (file)
--- a/nominatim/clicmd/special_phrases.py
+++ b/nominatim/clicmd/special_phrases.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 the 'special-phrases' command.
 """
 """
     Implementation of the 'special-phrases' command.
 """


@@ -19,14 +25,42 @@ LOG = logging.getLogger()
 class ImportSpecialPhrases:
     """\
     Import special phrases.
 class ImportSpecialPhrases:
     """\
     Import special phrases.

+
+    Special phrases are search terms that narrow down the type of object
+    that should be searched. For example, you might want to search for
+    'Hotels in Barcelona'. The OSM wiki has a selection of special phrases
+    in many languages, which can be imported with this command.
+
+    You can also provide your own phrases in a CSV file. The file needs to have
+    the following five columns:
+     * phrase - the term expected for searching
+     * class - the OSM tag key of the object type
+     * type - the OSM tag value of the object type
+     * operator - the kind of search to be done (one of: in, near, name, -)
+     * plural - whether the term is a plural or not (Y/N)
+
+    An example file can be found in the Nominatim sources at
+    'test/testdb/full_en_phrases_test.csv'.
+
+    The import can be further configured to ignore specific key/value pairs.
+    This is particularly useful when importing phrases from the wiki. The
+    default configuration excludes some very common tags like building=yes.
+    The configuration can be customized by putting a file `phrase-settings.json`
+    with custom rules into the project directory or by using the `--config`
+    option to point to another configuration file.

     """
     @staticmethod
     def add_args(parser):
         group = parser.add_argument_group('Input arguments')
         group.add_argument('--import-from-wiki', action='store_true',
     """
     @staticmethod
     def add_args(parser):
         group = parser.add_argument_group('Input arguments')
         group.add_argument('--import-from-wiki', action='store_true',

-                           help='Import special phrases from the OSM wiki to the database.')
+                           help='Import special phrases from the OSM wiki to the database')

         group.add_argument('--import-from-csv', metavar='FILE',
         group.add_argument('--import-from-csv', metavar='FILE',

-                           help='Import special phrases from a CSV file.')
+                           help='Import special phrases from a CSV file')
+        group.add_argument('--no-replace', action='store_true',
+                           help='Keep the old phrases and only add the new ones')
+        group.add_argument('--config', action='store',
+                           help='Configuration file for black/white listing '
+                                '(default: phrase-settings.json)')

 
     @staticmethod
     def run(args):
 
     @staticmethod
     def run(args):


@@ -51,7 +85,8 @@ class ImportSpecialPhrases:
         from ..tokenizer import factory as tokenizer_factory
 
         tokenizer = tokenizer_factory.get_tokenizer_for_db(args.config)
         from ..tokenizer import factory as tokenizer_factory
 
         tokenizer = tokenizer_factory.get_tokenizer_for_db(args.config)

+        should_replace = not args.no_replace

         with connect(args.config.get_libpq_dsn()) as db_connection:
             SPImporter(
         with connect(args.config.get_libpq_dsn()) as db_connection:
             SPImporter(

-                args.config, args.phplib_dir, db_connection, loader
-            ).import_phrases(tokenizer)
+                args.config, db_connection, loader
+            ).import_phrases(tokenizer, should_replace)