Only log a warning if a wrong input is detected on the wiki while importing special...

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

diff --git a/nominatim/tools/special_phrases.py b/nominatim/tools/special_phrases.py
index 3bb240ab88d9b0e90c36c714de89112395f2742b..d07ce570d4f882f47337758506ca13c586f8785a 100644 (file)
--- a/nominatim/tools/special_phrases.py
+++ b/nominatim/tools/special_phrases.py
@@ -27,7 +27,7 @@ class SpecialPhrasesImporter():
         self.black_list, self.white_list = self._load_white_and_black_lists()
         #Compile the regex here to increase performances.
         self.occurence_pattern = re.compile(
-            r'\| ([^\|]+) \|\| ([^\|]+) \|\| ([^\|]+) \|\| ([^\|]+) \|\| ([\-YN])'
+            r'\| *([^\|]+) *\|\| *([^\|]+) *\|\| *([^\|]+) *\|\| *([^\|]+) *\|\| *([\-YN])'
         )
         self.sanity_check_pattern = re.compile(r'^\w+$')
         self.transliterator = Transliterator.createFromRules("special-phrases normalizer",
@@ -65,7 +65,7 @@ class SpecialPhrasesImporter():
         if self.config.PHRASE_CONFIG:
             settings_path = self._convert_php_settings_if_needed(self.config.PHRASE_CONFIG)
 
-        with open(settings_path, "r") as json_settings:
+        with settings_path.open("r") as json_settings:
             settings = json.load(json_settings)
         return settings['blackList'], settings['whiteList']
 
@@ -102,8 +102,10 @@ class SpecialPhrasesImporter():
         class_matchs = self.sanity_check_pattern.findall(phrase_class)
 
         if len(class_matchs) < 1 or len(type_matchs) < 1:
-            raise UsageError("Bad class/type for language {}: {}={}".format(
-                lang, phrase_class, phrase_type))
+            LOG.warning("Bad class/type for language %s: %s=%s. It will not be imported",
+                        lang, phrase_class, phrase_type)
+            return False
+        return True
 
     def _process_xml_content(self, xml_content, lang):
         """