Added --no-replace command for special phrases importation and added corresponding tests

nominatim/clicmd/special_phrases.py

@@ -27,6 +27,8 @@ class ImportSpecialPhrases:
                            help='Import special phrases from the OSM wiki to the database.')
         group.add_argument('--import-from-csv', metavar='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.')
 
     @staticmethod
     def run(args):
@@ -51,7 +53,8 @@ class ImportSpecialPhrases:
         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(
                 args.config, args.phplib_dir, db_connection, loader
-            ).import_phrases(tokenizer)
+            ).import_phrases(tokenizer, should_replace)

nominatim/tokenizer/legacy_icu_tokenizer.py

@@ -306,7 +306,7 @@ class LegacyICUNameAnalyzer:
             #                WHERE word_id is null and type = 'postcode'""")
 
 
-    def update_special_phrases(self, phrases):
+    def update_special_phrases(self, phrases, should_replace):
         """ Replace the search index for special phrases with the new phrases.
         """
         norm_phrases = set(((self.normalize(p[0]), p[1], p[2], p[3])
@@ -345,7 +345,7 @@ class LegacyICUNameAnalyzer:
                               columns=['word', 'word_token', 'class', 'type',
                                        'operator', 'search_name_count'])
 
-            if to_delete:
+            if to_delete and should_replace:
                 psycopg2.extras.execute_values(
                     cur,
                     """ DELETE FROM word USING (VALUES %s) as v(name, in_class, in_type, op)

nominatim/tokenizer/legacy_tokenizer.py

@@ -314,7 +314,7 @@ class LegacyNameAnalyzer:
                                  FROM location_postcode) x""")
 
 
-    def update_special_phrases(self, phrases):
+    def update_special_phrases(self, phrases, should_replace):
         """ Replace the search index for special phrases with the new phrases.
         """
         norm_phrases = set(((self.normalize(p[0]), p[1], p[2], p[3])
@@ -343,7 +343,7 @@ class LegacyNameAnalyzer:
                            FROM (VALUES %s) as v(name, class, type, op))""",
                     to_add)
 
-            if to_delete:
+            if to_delete and should_replace:
                 psycopg2.extras.execute_values(
                     cur,
                     """ DELETE FROM word USING (VALUES %s) as v(name, in_class, in_type, op)

nominatim/tools/special_phrases/sp_importer.py

@@ -23,7 +23,7 @@ LOG = logging.getLogger()
 class SPImporter():
     # pylint: disable-msg=too-many-instance-attributes
     """
-        Class handling the process of special phrases importations into the database.
+        Class handling the process of special phrases importation into the database.
 
         Take a sp loader which load the phrases from an external source.
     """
@@ -42,10 +42,14 @@ class SPImporter():
         #special phrases class/type on the wiki.
         self.table_phrases_to_delete = set()
 
-    def import_phrases(self, tokenizer):
+    def import_phrases(self, tokenizer, should_replace):
         """
-            Iterate through all specified languages and
-            extract corresponding special phrases from the wiki.
+            Iterate through all SpecialPhrases extracted from the
+            loader and import them into the database.
+
+            If should_replace is set to True only the loaded phrases
+            will be kept into the database. All other phrases already
+            in the database will be removed.
         """
         LOG.warning('Special phrases importation starting')
         self._fetch_existing_place_classtype_tables()
@@ -60,11 +64,12 @@ class SPImporter():
                     class_type_pairs.update(result)
 
         self._create_place_classtype_table_and_indexes(class_type_pairs)
-        self._remove_non_existent_tables_from_db()
+        if should_replace:
+            self._remove_non_existent_tables_from_db()
         self.db_connection.commit()
 
         with tokenizer.name_analyzer() as analyzer:
-            analyzer.update_special_phrases(self.word_phrases)
+            analyzer.update_special_phrases(self.word_phrases, should_replace)
 
         LOG.warning('Import done.')
         self.statistics_handler.notify_import_done()