move linking of places to the preparation stage

Linked places may bring in extra names. These names need to be
processed by the tokenizer. That means that the linking needs
to be done before the data is handed to the tokenizer. Move finding
the linked place into the preparation stage and update the name
fields. Everything else is still done in the indexing stage.

diff --git a/lib-sql/functions/placex_triggers.sql b/lib-sql/functions/placex_triggers.sql
index e3be9503aa84eadca65e2f74d07e6e7aa5835d4b..fa7156ec904c396a8376b780d03c09a37045fb43 100644 (file)
--- a/lib-sql/functions/placex_triggers.sql
+++ b/lib-sql/functions/placex_triggers.sql
@@ -11,8 +11,11 @@
 CREATE OR REPLACE FUNCTION placex_prepare_update(p placex,
                                                  OUT name HSTORE,
                                                  OUT address HSTORE,
-                                                 OUT country_feature VARCHAR)
+                                                 OUT country_feature VARCHAR,
+                                                 OUT linked_place_id BIGINT)
   AS $$
+DECLARE
+  location RECORD;
 BEGIN
   -- For POI nodes, check if the address should be derived from a surrounding
   -- building.
@@ -34,6 +37,18 @@ BEGIN
   address := address - '_unlisted_place'::TEXT;
   name := p.name;
 
+  -- Names of linked places need to be merged in, so search for a linkable
+  -- place already here.
+  SELECT * INTO location FROM find_linked_place(p);
+
+  IF location.place_id is not NULL THEN
+    linked_place_id := location.place_id;
+
+    IF NOT location.name IS NULL THEN
+      name := location.name || name;
+    END IF;
+  END IF;
+
   country_feature := CASE WHEN p.admin_level = 2
                                and p.class = 'boundary' and p.type = 'administrative'
                                and p.osm_type = 'R'
@@ -683,6 +698,8 @@ DECLARE
   nameaddress_vector INTEGER[];
   addr_nameaddress_vector INTEGER[];
 
+  linked_place BIGINT;
+
   linked_node_id BIGINT;
   linked_importance FLOAT;
   linked_wikipedia TEXT;
@@ -718,9 +735,14 @@ BEGIN
 
   NEW.extratags := NEW.extratags - 'linked_place'::TEXT;
 
+  -- NEW.linked_place_id contains the precomputed linkee. Save this and restore
+  -- the previous link status.
+  linked_place := NEW.linked_place_id;
+  NEW.linked_place_id := OLD.linked_place_id;
+
   IF NEW.linked_place_id is not null THEN
     NEW.token_info := null;
-    {% if debug %}RAISE WARNING 'place already linked to %', NEW.linked_place_id;{% endif %}
+    {% if debug %}RAISE WARNING 'place already linked to %', OLD.linked_place_id;{% endif %}
     RETURN NEW;
   END IF;
 
@@ -956,8 +978,9 @@ BEGIN
   -- ---------------------------------------------------------------------------
   -- Full indexing
   {% if debug %}RAISE WARNING 'Using full index mode for % %', NEW.osm_type, NEW.osm_id;{% endif %}
-  SELECT * INTO location FROM find_linked_place(NEW);
-  IF location.place_id is not null THEN
+  IF linked_place is not null THEN
+    SELECT * INTO location FROM placex WHERE place_id = linked_place;
+
     {% if debug %}RAISE WARNING 'Linked %', location;{% endif %}
 
     -- Use the linked point as the centre point of the geometry,
@@ -974,11 +997,6 @@ BEGIN
       NEW.rank_address := location.rank_address;
     END IF;
 
-    -- merge in the label name
-    IF NOT location.name IS NULL THEN
-      NEW.name := location.name || NEW.name;
-    END IF;
-
     -- merge in extra tags
     NEW.extratags := hstore('linked_' || location.class, location.type)
                      || coalesce(location.extratags, ''::hstore)

diff --git a/nominatim/indexer/runners.py b/nominatim/indexer/runners.py
index 068d7d0fa63edba9a45fe7aef9195d06d164996a..29261ee50ebe59aac57a260e560a6016036e5139 100644 (file)
--- a/nominatim/indexer/runners.py
+++ b/nominatim/indexer/runners.py
@@ -16,6 +16,7 @@ class AbstractPlacexRunner:
     """ Returns SQL commands for indexing of the placex table.
     """
     SELECT_SQL = pysql.SQL('SELECT place_id FROM placex ')
+    UPDATE_LINE = "(%s, %s::hstore, %s::hstore, %s::int, %s::jsonb)"
 
     def __init__(self, rank, analyzer):
         self.rank = rank
@@ -27,10 +28,11 @@ class AbstractPlacexRunner:
     def _index_sql(num_places):
         return pysql.SQL(
             """ UPDATE placex
-                SET indexed_status = 0, address = v.addr, token_info = v.ti
-                FROM (VALUES {}) as v(id, addr, ti)
+                SET indexed_status = 0, address = v.addr, token_info = v.ti,
+                    name = v.name, linked_place_id = v.linked_place_id
+                FROM (VALUES {}) as v(id, name, addr, linked_place_id, ti)
                 WHERE place_id = v.id
-            """).format(_mk_valuelist("(%s, %s::hstore, %s::jsonb)", num_places))
+            """).format(_mk_valuelist(AbstractPlacexRunner.UPDATE_LINE, num_places))
 
 
     @staticmethod
@@ -43,7 +45,8 @@ class AbstractPlacexRunner:
     def index_places(self, worker, places):
         values = []
         for place in places:
-            values.extend((place[x] for x in ('place_id', 'address')))
+            for field in ('place_id', 'name', 'address', 'linked_place_id'):
+                values.append(place[field])
             values.append(psycopg2.extras.Json(self.analyzer.process_place(place)))
 
         worker.perform(self._index_sql(len(places)), values)

diff --git a/test/bdd/db/query/linking.feature b/test/bdd/db/query/linking.feature
new file mode 100644 (file)
index 0000000..4e6c47d
--- /dev/null
+++ b/test/bdd/db/query/linking.feature
@@ -0,0 +1,22 @@
+@DB
+Feature: Searching linked places
+    Tests that information from linked places can be searched correctly
+
+    Scenario: Additional names from linked places are searchable
+        Given the places
+         | osm  | class    | type           | admin | name  | geometry |
+         | R13  | boundary | administrative | 6     | Garbo | poly-area:0.1 |
+        Given the places
+         | osm  | class    | type           | admin | name+name:it | geometry |
+         | N2   | place    | hamlet         | 15    | Vario        | 0.006 0.00001 |
+        And the relations
+         | id | members       | tags+type |
+         | 13 | N2:label      | boundary |
+        When importing
+        Then placex contains
+         | object  | linked_place_id |
+         | N2      | R13 |
+        When sending search query "Vario"
+        Then results contain
+         | osm |
+         | R13 |

diff --git a/test/python/test_indexing.py b/test/python/test_indexing.py
index 9873e7d72673a8aaeb3e356276a4b558d97be3e8..60ad0bc4cbd3c5705891c60c847809ba74d7b985 100644 (file)
--- a/test/python/test_indexing.py
+++ b/test/python/test_indexing.py
@@ -19,8 +19,10 @@ class IndexerTestDB:
         with self.conn.cursor() as cur:
             cur.execute('CREATE EXTENSION hstore')
             cur.execute("""CREATE TABLE placex (place_id BIGINT,
+                                                name HSTORE,
                                                 class TEXT,
                                                 type TEXT,
+                                                linked_place_id BIGINT,
                                                 rank_address SMALLINT,
                                                 rank_search SMALLINT,
                                                 indexed_status SMALLINT,
@@ -55,11 +57,12 @@ class IndexerTestDB:
             cur.execute("""CREATE OR REPLACE FUNCTION placex_prepare_update(p placex,
                                                       OUT name HSTORE,
                                                       OUT address HSTORE,
-                                                      OUT country_feature VARCHAR)
+                                                      OUT country_feature VARCHAR,
+                                                      OUT linked_place_id BIGINT)
                            AS $$
                            BEGIN
                             address := p.address;
-                            name := p.address;
+                            name := p.name;
                            END;
                            $$ LANGUAGE plpgsql STABLE;
                         """)