Merge pull request #2553 from lonvia/revert-street-matching-to-full-names

Revert street matching to full names

diff --git a/lib-sql/tokenizer/icu_tokenizer.sql b/lib-sql/tokenizer/icu_tokenizer.sql
index 6092319a0578d338915b6890902d71df7ec90b1f..547facce1572f367d2f7164ad94ddeddad929074 100644 (file)
--- a/lib-sql/tokenizer/icu_tokenizer.sql
+++ b/lib-sql/tokenizer/icu_tokenizer.sql
@@ -51,7 +51,7 @@ $$ LANGUAGE SQL IMMUTABLE;
 CREATE OR REPLACE FUNCTION token_matches_street(info JSONB, street_tokens INTEGER[])
   RETURNS BOOLEAN
 AS $$
-  SELECT (info->>'street')::INTEGER[] <@ street_tokens
+  SELECT (info->>'street')::INTEGER[] && street_tokens
 $$ LANGUAGE SQL IMMUTABLE STRICT;
 
 

diff --git a/nominatim/tokenizer/icu_tokenizer.py b/nominatim/tokenizer/icu_tokenizer.py
index ea6e5d3cca5a9d063cd69b89c214f1d5e9699526..90caec1c9041697f02fec06c62960d08eaf01dcb 100644 (file)
--- a/nominatim/tokenizer/icu_tokenizer.py
+++ b/nominatim/tokenizer/icu_tokenizer.py
@@ -409,16 +409,18 @@ class LegacyICUNameAnalyzer(AbstractAnalyzer):
     def _process_place_address(self, token_info, address):
         hnrs = []
         addr_terms = []
+        streets = []
         for item in address:
             if item.kind == 'postcode':
                 self._add_postcode(item.name)
             elif item.kind in ('housenumber', 'streetnumber', 'conscriptionnumber'):
                 hnrs.append(item.name)
             elif item.kind == 'street':
-                token_info.add_street(self._compute_partial_tokens(item.name))
+                streets.extend(self._retrieve_full_tokens(item.name))
             elif item.kind == 'place':
-                token_info.add_place(self._compute_partial_tokens(item.name))
-            elif not item.kind.startswith('_') and \
+                if not item.suffix:
+                    token_info.add_place(self._compute_partial_tokens(item.name))
+            elif not item.kind.startswith('_') and not item.suffix and \
                  item.kind not in ('country', 'full'):
                 addr_terms.append((item.kind, self._compute_partial_tokens(item.name)))
 
@@ -429,6 +431,9 @@ class LegacyICUNameAnalyzer(AbstractAnalyzer):
         if addr_terms:
             token_info.add_address_terms(addr_terms)
 
+        if streets:
+            token_info.add_street(streets)
+
 
     def _compute_partial_tokens(self, name):
         """ Normalize the given term, split it into partial words and return
@@ -458,6 +463,26 @@ class LegacyICUNameAnalyzer(AbstractAnalyzer):
         return tokens
 
 
+    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._search_normalized(name)
+
+        # return cached if possible
+        if norm_name in self._cache.fulls:
+            return self._cache.fulls[norm_name]
+
+        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
+
+        return full
+
+
     def _compute_name_tokens(self, names):
         """ Computes the full name and partial name tokens for the given
             dictionary of names.
@@ -561,8 +586,7 @@ class _TokenInfo:
     def add_street(self, tokens):
         """ Add addr:street match terms.
         """
-        if tokens:
-            self.data['street'] = self._mk_array(tokens)
+        self.data['street'] = self._mk_array(tokens)
 
 
     def add_place(self, tokens):
@@ -591,6 +615,7 @@ class _TokenCache:
     def __init__(self):
         self.names = {}
         self.partials = {}
+        self.fulls = {}
         self.postcodes = set()
         self.housenumbers = {}
 

diff --git a/settings/import-address.style b/settings/import-address.style
index d872e2890c68a0e4abd61f507b7d57107cd04a33..c69f2c76666d23bdd91a6923f133aa35c8ecf730 100644 (file)
--- a/settings/import-address.style
+++ b/settings/import-address.style
@@ -11,8 +11,9 @@
     }
 },
 {
-    "keys" : ["name:prefix", "name:suffix", "name:prefix:*", "name:suffix:*",
-              "name:etymology", "name:signed", "name:botanical", "*wikidata"],
+    "keys" : ["*:prefix", "*:suffix", "name:prefix:*", "name:suffix:*",
+              "name:etymology", "name:signed", "name:botanical", "*:wikidata",
+              "addr:street:name", "addr:street:type"],
     "values" : {
         "" : "skip"
     }

diff --git a/settings/import-admin.style b/settings/import-admin.style
index 4e41763f276554c3e11f6b96060f12b09b20a61c..92484cb0ccddb64947748585b0327ae16caf7084 100644 (file)
--- a/settings/import-admin.style
+++ b/settings/import-admin.style
@@ -5,8 +5,9 @@
     }
 },
 {
-    "keys" : ["name:prefix", "name:suffix", "name:prefix:*", "name:suffix:*",
-              "name:etymology", "name:signed", "name:botanical", "*wikidata"],
+    "keys" : ["*:prefix", "*:suffix", "name:prefix:*", "name:suffix:*",
+              "name:etymology", "name:signed", "name:botanical", "*:wikidata",
+              "addr:street:name", "addr:street:type"],
     "values" : {
         "" : "skip"
     }

diff --git a/settings/import-extratags.style b/settings/import-extratags.style
index ce32ae1e670d936260833fdbabcb0d082c6adede..52945bbaa56c17ff26cd990b5e712b4e3e692913 100644 (file)
--- a/settings/import-extratags.style
+++ b/settings/import-extratags.style
@@ -6,8 +6,9 @@
     }
 },
 {
-    "keys" : ["name:prefix", "name:suffix", "name:prefix:*", "name:suffix:*",
-              "name:etymology", "name:signed", "name:botanical", "wikidata", "*:wikidata"],
+    "keys" : ["*:prefix", "*:suffix", "name:prefix:*", "name:suffix:*",
+              "name:etymology", "name:signed", "name:botanical", "wikidata", "*:wikidata",
+              "addr:street:name", "addr:street:type"],
     "values" : {
         "" : "extra"
     }

diff --git a/settings/import-full.style b/settings/import-full.style
index 2e2f25f0e938f8f30742e2e727b4174fe2f80cee..ce82ff83c1e1648096c0934758e41251ed81817a 100644 (file)
--- a/settings/import-full.style
+++ b/settings/import-full.style
@@ -6,8 +6,9 @@
     }
 },
 {
-    "keys" : ["name:prefix", "name:suffix", "name:prefix:*", "name:suffix:*",
-              "name:etymology", "name:signed", "name:botanical", "wikidata", "*:wikidata"],
+    "keys" : ["*:prefix", "*:suffix", "name:prefix:*", "name:suffix:*",
+              "name:etymology", "name:signed", "name:botanical", "wikidata", "*:wikidata",
+              "addr:street:name", "addr:street:type"],
     "values" : {
         "" : "extra"
     }

diff --git a/settings/import-street.style b/settings/import-street.style
index 77cf4790f3db436642ada5ef5c4cf8e59511abc0..f1bc61eeb6c541836a47f64777730cc3c3ad6590 100644 (file)
--- a/settings/import-street.style
+++ b/settings/import-street.style
@@ -5,8 +5,9 @@
     }
 },
 {
-    "keys" : ["name:prefix", "name:suffix", "name:prefix:*", "name:suffix:*",
-              "name:etymology", "name:signed", "name:botanical", "*wikidata"],
+    "keys" : ["*:prefix", "*:suffix", "name:prefix:*", "name:suffix:*",
+              "name:etymology", "name:signed", "name:botanical", "*:wikidata",
+              "addr:street:name", "addr:street:type"],
     "values" : {
         "" : "skip"
     }

diff --git a/test/bdd/.behaverc b/test/bdd/.behaverc
index 32aa6dfa79d1f9010d1dad29b47fbfcd98ca681f..1b426ec9451185bb2a106b95fdcdb7d5be73fe6a 100644 (file)
--- a/test/bdd/.behaverc
+++ b/test/bdd/.behaverc
@@ -1,3 +1,3 @@
 [behave]
 show_skipped=False
-tags=~@Fail
+default_tags=~@Fail

diff --git a/test/bdd/db/import/parenting.feature b/test/bdd/db/import/parenting.feature
index b5210f9439b55a5f328d28f0141ac695629a6334..ef25b6cc0acd6b9c4c9501109e697f1fcf158d20 100644 (file)
--- a/test/bdd/db/import/parenting.feature
+++ b/test/bdd/db/import/parenting.feature
@@ -87,6 +87,52 @@ Feature: Parenting of objects
          | N3     | W2 |
          | N4     | W1 |
 
+    @fail-legacy
+    Scenario: addr:street tag parents to appropriately named street, locale names
+        Given the scene roads-with-pois
+        And the places
+         | osm | class | type  | street| addr+street:de | geometry |
+         | N1  | place | house | south | Süd               | :p-N1 |
+         | N2  | place | house | north | Nord              | :p-N2 |
+         | N3  | place | house | south | Süd               | :p-S1 |
+         | N4  | place | house | north | Nord              | :p-S2 |
+        And the places
+         | osm | class   | type        | name  | geometry |
+         | W1  | highway | residential | Nord | :w-north |
+         | W2  | highway | residential | Süd | :w-south |
+        And the places
+         | osm | class | type   | name  | name+name:old |
+         | N5  | place | hamlet | south | north         |
+        When importing
+        Then placex contains
+         | object | parent_place_id |
+         | N1     | W2 |
+         | N2     | W1 |
+         | 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

diff --git a/test/bdd/environment.py b/test/bdd/environment.py
index f179c8f13da343283b0aaf4deb855587a471cd6f..2c27716586fd1752730421c8e043ab8b1b4aced3 100644 (file)
--- a/test/bdd/environment.py
+++ b/test/bdd/environment.py
@@ -49,3 +49,9 @@ def before_scenario(context, scenario):
 def after_scenario(context, scenario):
     if 'DB' in context.tags:
         context.nominatim.teardown_db(context)
+
+
+def before_tag(context, tag):
+    if tag == 'fail-legacy':
+        if context.config.userdata['TOKENIZER'] in (None, 'legacy'):
+            context.scenario.skip("Not implemented in legacy tokenizer")

diff --git a/test/python/tokenizer/test_icu.py b/test/python/tokenizer/test_icu.py
index 642aaceb82e0f55ac59d03327b252fa53e0d250c..83668b3936e443f11e7b0928a0eee28a588f55ae 100644 (file)
--- a/test/python/tokenizer/test_icu.py
+++ b/test/python/tokenizer/test_icu.py
@@ -471,9 +471,25 @@ class TestPlaceAddress:
 
 
     def test_process_place_street(self):
+        self.analyzer.process_place(PlaceInfo({'name': {'name' : 'Grand Road'}}))
         info = self.process_address(street='Grand Road')
 
-        assert eval(info['street']) == self.name_token_set('GRAND', 'ROAD')
+        assert eval(info['street']) == self.name_token_set('#Grand Road')
+
+
+    def test_process_place_nonexisting_street(self):
+        info = self.process_address(street='Grand Road')
+
+        assert 'street' not in info
+
+
+    def test_process_place_multiple_street_tags(self):
+        self.analyzer.process_place(PlaceInfo({'name': {'name' : 'Grand Road',
+                                                        'ref': '05989'}}))
+        info = self.process_address(**{'street': 'Grand Road',
+                                      'street:sym_ul': '05989'})
+
+        assert eval(info['street']) == self.name_token_set('#Grand Road', '#05989')
 
 
     def test_process_place_street_empty(self):
@@ -482,12 +498,28 @@ class TestPlaceAddress:
         assert 'street' not in info
 
 
+    def test_process_place_street_from_cache(self):
+        self.analyzer.process_place(PlaceInfo({'name': {'name' : 'Grand Road'}}))
+        self.process_address(street='Grand Road')
+
+        # request address again
+        info = self.process_address(street='Grand Road')
+
+        assert eval(info['street']) == self.name_token_set('#Grand Road')
+
+
     def test_process_place_place(self):
         info = self.process_address(place='Honu Lulu')
 
         assert eval(info['place']) == self.name_token_set('HONU', 'LULU')
 
 
+    def test_process_place_place_extra(self):
+        info = self.process_address(**{'place:en': 'Honu Lulu'})
+
+        assert 'place' not in info
+
+
     def test_process_place_place_empty(self):
         info = self.process_address(place='🜵')
 
@@ -507,6 +539,14 @@ class TestPlaceAddress:
         assert result == {'city': city, 'suburb': city, 'state': state}
 
 
+    def test_process_place_multiple_address_terms(self):
+        info = self.process_address(**{'city': 'Bruxelles', 'city:de': 'Brüssel'})
+
+        result = {k: eval(v) for k,v in info['addr'].items()}
+
+        assert result == {'city': self.name_token_set('Bruxelles')}
+
+
     def test_process_place_address_terms_empty(self):
         info = self.process_address(country='de', city=' ', street='Hauptstr',
                                     full='right behind the church')