[boundaries] adding check for valid name key in formatted places, and removing short_name from the Sao Paulo relation as well

resources/boundaries/names/global.yaml

@@ -86,6 +86,15 @@ names:
                 probability: 0.09
               - alternative: official_name
                 probability: 0.01
+        - id: 298285 # Sao Paulo (relation)
+          type: relation
+          default: name
+          probability: 0.9
+          alternatives:
+              - alternative: alt_name
+                probability: 0.09
+              - alternative: official_name
+                probability: 0.01
         - id: 556706 # New Zealand
           type: relation
           default: name:en

scripts/geodata/boundaries/names.py

@@ -110,6 +110,9 @@ class BoundaryNames(object):
     def _string_as_regex(self, s):
         return safe_decode(s).replace(six.u('.'), six.u('\\.'))
 
+    def valid_name(self, object_type, object_id, name):
+        return name in self.exceptions.get((object_type, object_id), ((), ()))[0]
+
     def name_key_dist(self, props, component):
         object_type = props.get('type')
         object_id = safe_encode(props.get('id', ''))

scripts/geodata/osm/formatter.py

@@ -606,7 +606,7 @@ class OSMAddressFormatter(object):
         except Exception:
             return (), None
 
-        if 'name' not in tags:
+        if 'name' not in tags and not any((t.startswith('name:') for t in tags)):
             return (), None
 
         osm_components = self.components.osm_reverse_geocoded_components(latitude, longitude)
@@ -732,7 +732,13 @@ class OSMAddressFormatter(object):
 
         revised_tags = self.fix_component_encodings(revised_tags)
 
+        object_type = tags.get('type')
+        object_id = tags.get('id')
+
         for name_tag in ('name', 'alt_name', 'loc_name', 'short_name', 'int_name', 'name:simple', 'official_name'):
+            if not boundary_names.valid_name(object_type, object_id, name_tag):
+                continue
+
             if more_than_one_official_language:
                 name = tags.get(name_tag)
                 language_suffix = ''