Refactoring loading of external special phrases and importation process by introducin...

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

diff --git a/nominatim/clicmd/special_phrases.py b/nominatim/clicmd/special_phrases.py
index 1eb25bdc156ea8dc9bc0ce53441a1d598620de0c..0de8b23957b5b0c0259d8734a6dcec9e1abdfe1a 100644 (file)
--- a/nominatim/clicmd/special_phrases.py
+++ b/nominatim/clicmd/special_phrases.py
@@ -2,13 +2,18 @@
     Implementation of the 'special-phrases' command.
 """
 import logging
-from nominatim.tools import SpecialPhrasesImporter
+from nominatim.errors import UsageError
+from pathlib import Path
+from nominatim.tools import SPWikiLoader
+from nominatim.tools import SPImporter
 from nominatim.db.connection import connect
 
 LOG = logging.getLogger()
 
 # Do not repeat documentation of subcommand classes.
 # pylint: disable=C0111
+# Using non-top-level imports to avoid eventually unused imports.
+# pylint: disable=E0012,C0415
 
 class ImportSpecialPhrases:
     """\
@@ -19,13 +24,23 @@ class ImportSpecialPhrases:
         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.')
+        group.add_argument('--csv-file', metavar='FILE',
+                    help='CSV file containing phrases to import.')
 
     @staticmethod
     def run(args):
+        from ..tokenizer import factory as tokenizer_factory
+
         if args.import_from_wiki:
-            LOG.warning('Special phrases importation starting')
+            tokenizer = tokenizer_factory.get_tokenizer_for_db(args.config)
             with connect(args.config.get_libpq_dsn()) as db_connection:
-                SpecialPhrasesImporter(
-                    args.config, args.phplib_dir, db_connection
-                ).import_from_wiki()
+                SPImporter(
+                    args.config, args.phplib_dir, db_connection, SPWikiLoader(args.config)
+                ).import_phrases(tokenizer)
+
+        if args.csv_file:
+            if not Path(args.csv_file).is_file():
+                LOG.fatal("CSV file '%s' does not exist.", args.csv_file)
+                raise UsageError('Cannot access file.')
+
         return 0