From: Sarah Hoffmann <lonvia@denofr.de>
Date: Wed, 8 Dec 2021 20:58:43 +0000 (+0100)
Subject: correctly match abbreviated addr:street
X-Git-Tag: v4.1.0~104^2
X-Git-Url: https://git.openstreetmap.org./nominatim.git/commitdiff_plain/f9b56a8581be3663239d23ee1194891dde9a3857

correctly match abbreviated addr:street

This only works when addr:street is abbreviated and the street
name isn't. It does not work the other way around.
---

diff --git a/nominatim/tokenizer/icu_tokenizer.py b/nominatim/tokenizer/icu_tokenizer.py
index 33f05cc4..90caec1c 100644
--- a/nominatim/tokenizer/icu_tokenizer.py
+++ b/nominatim/tokenizer/icu_tokenizer.py
@@ -416,9 +416,7 @@ class LegacyICUNameAnalyzer(AbstractAnalyzer):
             elif item.kind in ('housenumber', 'streetnumber', 'conscriptionnumber'):
                 hnrs.append(item.name)
             elif item.kind == 'street':
-                token = self._retrieve_full_token(item.name)
-                if token:
-                    streets.append(token)
+                streets.extend(self._retrieve_full_tokens(item.name))
             elif item.kind == 'place':
                 if not item.suffix:
                     token_info.add_place(self._compute_partial_tokens(item.name))
@@ -465,25 +463,20 @@ class LegacyICUNameAnalyzer(AbstractAnalyzer):
         return tokens
 
 
-    def _retrieve_full_token(self, name):
+    def _retrieve_full_tokens(self, name):
         """ Get the full name token for the given name, if it exists.
             The name is only retrived for the standard analyser.
         """
-        norm_name = self._normalized(name)
+        norm_name = self._search_normalized(name)
 
         # return cached if possible
         if norm_name in self._cache.fulls:
             return self._cache.fulls[norm_name]
 
-        # otherwise compute
-        full, _ = self._cache.names.get(norm_name, (None, None))
-
-        if full is None:
-            with self.conn.cursor() as cur:
-                cur.execute("SELECT word_id FROM word WHERE word = %s and type = 'W' LIMIT 1",
-                            (norm_name, ))
-                if cur.rowcount > 0:
-                    full = cur.fetchone()[0]
+        with self.conn.cursor() as cur:
+            cur.execute("SELECT word_id FROM word WHERE word_token = %s and type = 'W'",
+                        (norm_name, ))
+            full = [row[0] for row in cur]
 
         self._cache.fulls[norm_name] = full
 
diff --git a/test/bdd/db/import/parenting.feature b/test/bdd/db/import/parenting.feature
index f6d88ca8..ef25b6cc 100644
--- a/test/bdd/db/import/parenting.feature
+++ b/test/bdd/db/import/parenting.feature
@@ -111,6 +111,28 @@ Feature: Parenting of objects
          | N3     | W2 |
          | N4     | W1 |
 
+    Scenario: addr:street tag parents to appropriately named street with abbreviation
+        Given the scene roads-with-pois
+        And the places
+         | osm | class | type  | street| geometry |
+         | N1  | place | house | south st | :p-N1 |
+         | N2  | place | house | north st | :p-N2 |
+         | N3  | place | house | south st | :p-S1 |
+         | N4  | place | house | north st | :p-S2 |
+        And the places
+         | osm | class   | type        | name+name:en  | geometry |
+         | W1  | highway | residential | north street | :w-north |
+         | W2  | highway | residential | south street | :w-south |
+        When importing
+        Then placex contains
+         | object | parent_place_id |
+         | N1     | W2 |
+         | N2     | W1 |
+         | N3     | W2 |
+         | N4     | W1 |
+
+
+
     Scenario: addr:street tag parents to next named street
         Given the scene roads-with-pois
         And the places