split addr: tags into words before adding to the search index

[nominatim.git] / sql / functions / normalization.sql

diff --git a/sql/functions/normalization.sql b/sql/functions/normalization.sql
index 53a81588e09e9f2f79c57c92b969cd165294f333..a5560cb98dbb719de4bfe54397e9889442374d57 100644 (file)
--- a/sql/functions/normalization.sql
+++ b/sql/functions/normalization.sql
@@ -81,7 +81,8 @@ BEGIN
   lookup_word := upper(trim(postcode));
   lookup_token := ' ' || make_standard_name(lookup_word);
   SELECT min(word_id) FROM word
   lookup_word := upper(trim(postcode));
   lookup_token := ' ' || make_standard_name(lookup_word);
   SELECT min(word_id) FROM word

-    WHERE word_token = lookup_token and class='place' and type='postcode'
+    WHERE word_token = lookup_token and word = lookup_word
+          and class='place' and type='postcode'

     INTO return_word_id;
   IF return_word_id IS NULL THEN
     return_word_id := nextval('seq_word');
     INTO return_word_id;
   IF return_word_id IS NULL THEN
     return_word_id := nextval('seq_word');


@@ -201,56 +202,58 @@ END;
 $$
 LANGUAGE plpgsql;
 
 $$
 LANGUAGE plpgsql;
 

-
-CREATE OR REPLACE FUNCTION get_word_id(lookup_word TEXT)
-  RETURNS INTEGER
+-- Normalize a string and lookup its word ids (partial words).
+CREATE OR REPLACE FUNCTION addr_ids_from_name(lookup_word TEXT)
+  RETURNS INTEGER[]

   AS $$
 DECLARE
   AS $$
 DECLARE

-  lookup_token TEXT;
-  return_word_id INTEGER;
+  words TEXT[];
+  id INTEGER;
+  return_word_id INTEGER[];
+  word_ids INTEGER[];
+  j INTEGER;

 BEGIN
 BEGIN

-  lookup_token := trim(lookup_word);
-  SELECT min(word_id) FROM word
-    WHERE word_token = lookup_token and class is null and type is null
-    INTO return_word_id;
-  RETURN return_word_id;
-END;
-$$
-LANGUAGE plpgsql STABLE;
-
+  words := string_to_array(make_standard_name(lookup_word), ' ');
+  IF array_upper(words, 1) IS NOT NULL THEN
+    FOR j IN 1..array_upper(words, 1) LOOP
+      IF (words[j] != '') THEN
+        SELECT array_agg(word_id) INTO word_ids
+          FROM word
+         WHERE word_token = words[j] and class is null and type is null;
+
+        IF word_ids IS NULL THEN
+          id := nextval('seq_word');
+          INSERT INTO word VALUES (id, words[j], null, null, null, null, 0);
+          return_word_id := return_word_id || id;
+        ELSE
+          return_word_id := array_merge(return_word_id, word_ids);
+        END IF;
+      END IF;
+    END LOOP;
+  END IF;

 
 

-CREATE OR REPLACE FUNCTION get_name_id(lookup_word TEXT)
-  RETURNS INTEGER
-  AS $$
-DECLARE
-  lookup_token TEXT;
-  return_word_id INTEGER;
-BEGIN
-  lookup_token := ' '||trim(lookup_word);
-  SELECT min(word_id) FROM word
-    WHERE word_token = lookup_token and class is null and type is null
-    INTO return_word_id;

   RETURN return_word_id;
 END;
 $$
   RETURN return_word_id;
 END;
 $$

-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql;

 
 
 
 

-CREATE OR REPLACE FUNCTION get_name_ids(lookup_word TEXT)
+-- Normalize a string and look up its name ids (full words).
+CREATE OR REPLACE FUNCTION word_ids_from_name(lookup_word TEXT)

   RETURNS INTEGER[]
   AS $$
 DECLARE
   lookup_token TEXT;
   return_word_ids INTEGER[];
 BEGIN
   RETURNS INTEGER[]
   AS $$
 DECLARE
   lookup_token TEXT;
   return_word_ids INTEGER[];
 BEGIN

-  lookup_token := ' '||trim(lookup_word);
+  lookup_token := ' '|| make_standard_name(lookup_word);

   SELECT array_agg(word_id) FROM word
     WHERE word_token = lookup_token and class is null and type is null
     INTO return_word_ids;
   RETURN return_word_ids;
 END;
 $$
   SELECT array_agg(word_id) FROM word
     WHERE word_token = lookup_token and class is null and type is null
     INTO return_word_ids;
   RETURN return_word_ids;
 END;
 $$

-LANGUAGE plpgsql STABLE;
+LANGUAGE plpgsql STABLE STRICT;

 
 
 CREATE OR REPLACE FUNCTION create_country(src HSTORE, country_code varchar(2))
 
 
 CREATE OR REPLACE FUNCTION create_country(src HSTORE, country_code varchar(2))


@@ -421,3 +424,124 @@ BEGIN
 END;
 $$
 LANGUAGE plpgsql;
 END;
 $$
 LANGUAGE plpgsql;

+
+
+CREATE OR REPLACE FUNCTION create_poi_search_terms(obj_place_id BIGINT,
+                                                   in_partition SMALLINT,
+                                                   parent_place_id BIGINT,
+                                                   address HSTORE,
+                                                   country TEXT,
+                                                   housenumber TEXT,
+                                                   initial_name_vector INTEGER[],
+                                                   geometry GEOMETRY,
+                                                   OUT name_vector INTEGER[],
+                                                   OUT nameaddress_vector INTEGER[])
+  AS $$
+DECLARE
+  parent_name_vector INTEGER[];
+  parent_address_vector INTEGER[];
+  addr_place_ids INTEGER[];
+
+  addr_item RECORD;
+  parent_address_place_ids BIGINT[];
+  filtered_address HSTORE;
+BEGIN
+  nameaddress_vector := '{}'::INTEGER[];
+
+  SELECT s.name_vector, s.nameaddress_vector
+    INTO parent_name_vector, parent_address_vector
+    FROM search_name s
+    WHERE s.place_id = parent_place_id;
+
+  -- Find all address tags that don't appear in the parent search names.
+  SELECT hstore(array_agg(ARRAY[k, v])) INTO filtered_address
+    FROM (SELECT skeys(address) as k, svals(address) as v) a
+   WHERE not addr_ids_from_name(v) && parent_address_vector
+         AND k not in ('country', 'street', 'place', 'postcode',
+                       'housenumber', 'streetnumber', 'conscriptionnumber');
+
+  -- Compute all search terms from the addr: tags.
+  IF filtered_address IS NOT NULL THEN
+    FOR addr_item IN
+      SELECT * FROM
+        get_places_for_addr_tags(in_partition, geometry, filtered_address, country)
+    LOOP
+        IF addr_item.place_id is null THEN
+            nameaddress_vector := array_merge(nameaddress_vector,
+                                              addr_item.keywords);
+            CONTINUE;
+        END IF;
+
+        IF parent_address_place_ids is null THEN
+            SELECT array_agg(parent_place_id) INTO parent_address_place_ids
+              FROM place_addressline
+             WHERE place_id = parent_place_id;
+        END IF;
+
+        IF not parent_address_place_ids @> ARRAY[addr_item.place_id] THEN
+            nameaddress_vector := array_merge(nameaddress_vector,
+                                              addr_item.keywords);
+
+            INSERT INTO place_addressline (place_id, address_place_id, fromarea,
+                                           isaddress, distance, cached_rank_address)
+            VALUES (obj_place_id, addr_item.place_id, not addr_item.isguess,
+                    true, addr_item.distance, addr_item.rank_address);
+        END IF;
+    END LOOP;
+  END IF;
+
+
+  -- If the POI is named, simply mix in all address terms and be done.
+  IF array_length(initial_name_vector, 1) is not NULL THEN
+    -- Cheating here by not recomputing all terms but simply using the ones
+    -- from the parent object.
+    name_vector := initial_name_vector;
+    nameaddress_vector := array_merge(nameaddress_vector, parent_name_vector);
+    nameaddress_vector := array_merge(nameaddress_vector, parent_address_vector);
+
+    IF not address ? 'street' and address ? 'place' THEN
+      -- make sure addr:place terms are always searchable
+      nameaddress_vector := array_merge(nameaddress_vector,
+                                        addr_ids_from_name(address->'place'));
+    END IF;
+
+    RETURN;
+  END IF;
+
+  ----- unnamed POIS
+
+  IF (array_length(nameaddress_vector, 1) is null
+      and (address ? 'street'or not address ? 'place'))
+     or housenumber is null
+  THEN
+    RETURN;
+  END IF;
+
+  -- Check if the parent covers all address terms.
+  -- If not, create a search name entry with the house number as the name.
+  -- This is unusual for the search_name table but prevents that the place
+  -- is returned when we only search for the street/place.
+
+  IF not nameaddress_vector <@ parent_address_vector THEN
+    name_vector := ARRAY[getorcreate_name_id(housenumber)];
+  END IF;
+
+  IF not address ? 'street' and address ? 'place' THEN
+    addr_place_ids := addr_ids_from_name(address->'place');
+    IF not addr_place_ids <@ parent_name_vector THEN
+      -- addr:place tag exists without a corresponding place. Mix in addr:place
+      -- in the address and drop the name from the parent. This would only be
+      -- the street name of the nearest street.
+      nameaddress_vector := array_merge(nameaddress_vector, addr_place_ids);
+      name_vector := ARRAY[getorcreate_name_id(housenumber)];
+    END IF;
+  ELSE
+    nameaddress_vector := array_merge(nameaddress_vector, parent_name_vector);
+  END IF;
+
+  -- The address vector always gets merged in.
+  nameaddress_vector := array_merge(nameaddress_vector, parent_address_vector);
+
+END;
+$$
+LANGUAGE plpgsql;