X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/f6c7c4fcef6e64b4a29482daddde5be885d7dd04..163b488591181e05a1ca0f80923ff7d5d309c71b:/sql/functions.sql

diff --git a/sql/functions.sql b/sql/functions.sql
index 45681610..775ff168 100644
--- a/sql/functions.sql
+++ b/sql/functions.sql
@@ -271,6 +271,20 @@ END;
 $$
 LANGUAGE plpgsql IMMUTABLE;
 
+CREATE OR REPLACE FUNCTION get_name_ids(lookup_word TEXT)
+  RETURNS INTEGER[]
+  AS $$
+DECLARE
+  lookup_token TEXT;
+  return_word_ids INTEGER[];
+BEGIN
+  lookup_token := ' '||trim(lookup_word);
+  SELECT array_agg(word_id) FROM word WHERE word_token = lookup_token and class is null and type is null into return_word_ids;
+  RETURN return_word_ids;
+END;
+$$
+LANGUAGE plpgsql IMMUTABLE;
+
 CREATE OR REPLACE FUNCTION array_merge(a INTEGER[], b INTEGER[])
   RETURNS INTEGER[]
   AS $$
@@ -582,10 +596,9 @@ END;
 $$
 LANGUAGE plpgsql IMMUTABLE;
 
-CREATE OR REPLACE FUNCTION get_partition(place geometry, in_country_code VARCHAR(10)) RETURNS INTEGER
+CREATE OR REPLACE FUNCTION get_partition(in_country_code VARCHAR(10)) RETURNS INTEGER
   AS $$
 DECLARE
-  place_centre GEOMETRY;
   nearcountry RECORD;
 BEGIN
   FOR nearcountry IN select partition from country_name where country_code = in_country_code
@@ -923,6 +936,11 @@ DECLARE
 BEGIN
   --DEBUG: RAISE WARNING '% %',NEW.osm_type,NEW.osm_id;
 
+  -- remove operator tag for most places, messes too much with search_name indexes
+  IF NEW.class not in ('amenity', 'shop') THEN
+    NEW.name := delete(NEW.name, 'operator');
+  END IF;
+
   -- just block these
   IF NEW.class in ('landuse','natural') and NEW.name is null THEN
 --    RAISE WARNING 'empty landuse %',NEW.osm_id;
@@ -953,7 +971,7 @@ BEGIN
 
   NEW.calculated_country_code := lower(get_country_code(NEW.geometry));
 
-  NEW.partition := get_partition(NEW.geometry, NEW.calculated_country_code);
+  NEW.partition := get_partition(NEW.calculated_country_code);
   NEW.geometry_sector := geometry_sector(NEW.partition, NEW.geometry);
 
   -- copy 'name' to or from the default language (if there is a default language)
@@ -1064,18 +1082,6 @@ BEGIN
       ELSEIF NEW.type in ('village','hamlet','municipality','district','unincorporated_area','borough') THEN
         NEW.rank_search := 19;
         NEW.rank_address := 16;
-      ELSEIF NEW.type in ('airport') AND ST_GeometryType(NEW.geometry) in ('ST_Polygon','ST_MultiPolygon') THEN
-        NEW.rank_search := 18;
-        NEW.rank_address := 17;
-      ELSEIF NEW.type in ('moor') AND ST_GeometryType(NEW.geometry) in ('ST_Polygon','ST_MultiPolygon') THEN
-        NEW.rank_search := 17;
-        NEW.rank_address := 18;
-      ELSEIF NEW.type in ('moor') THEN
-        NEW.rank_search := 17;
-        NEW.rank_address := 0;
-      ELSEIF NEW.type in ('national_park') THEN
-        NEW.rank_search := 18;
-        NEW.rank_address := 18;
       ELSEIF NEW.type in ('suburb','croft','subdivision','isolated_dwelling') THEN
         NEW.rank_search := 20;
         NEW.rank_address := NEW.rank_search;
@@ -1086,12 +1092,9 @@ BEGIN
         IF (NEW.extratags -> 'locality') = 'townland' THEN
           NEW.rank_address := 20;
         END IF;
-      ELSEIF NEW.type in ('hall_of_residence','neighbourhood','housing_estate','nature_reserve') THEN
+      ELSEIF NEW.type in ('neighbourhood') THEN
         NEW.rank_search := 22;
         NEW.rank_address := 22;
-      ELSEIF NEW.type in ('airport','street') THEN
-        NEW.rank_search := 26;
-        NEW.rank_address := NEW.rank_search;
       ELSEIF NEW.type in ('house','building') THEN
         NEW.rank_search := 30;
         NEW.rank_address := NEW.rank_search;
@@ -1120,6 +1123,11 @@ BEGIN
     ELSEIF NEW.class = 'natural' and NEW.type in ('peak','volcano','mountain_range') THEN
       NEW.rank_search := 18;
       NEW.rank_address := 0;
+    ELSEIF NEW.class = 'natural' and NEW.type = 'sea' THEN
+      NEW.rank_search := 4;
+      NEW.rank_address := NEW.rank_search;
+    ELSEIF NEW.class = 'natural' and NEW.type in ('coastline') THEN
+      RETURN NULL;
     -- any feature more than 5 square miles is probably worth indexing
     ELSEIF ST_GeometryType(NEW.geometry) in ('ST_Polygon','ST_MultiPolygon') AND ST_Area(NEW.geometry) > 0.1 THEN
       NEW.rank_search := 22;
@@ -1133,18 +1141,14 @@ BEGIN
     ELSEIF NEW.class = 'waterway' AND NEW.name is NULL THEN
       RETURN NULL;
     ELSEIF NEW.class = 'waterway' THEN
-      NEW.rank_address := 17;
+      NEW.rank_search := 17;
+      NEW.rank_address := 0;
     ELSEIF NEW.class = 'highway' AND NEW.osm_type != 'N' AND NEW.type in ('service','cycleway','path','footway','steps','bridleway','motorway_link','primary_link','trunk_link','secondary_link','tertiary_link') THEN
       NEW.rank_search := 27;
       NEW.rank_address := NEW.rank_search;
     ELSEIF NEW.class = 'highway' AND NEW.osm_type != 'N' THEN
       NEW.rank_search := 26;
       NEW.rank_address := NEW.rank_search;
-    ELSEIF NEW.class = 'natural' and NEW.type = 'sea' THEN
-      NEW.rank_search := 4;
-      NEW.rank_address := NEW.rank_search;
-    ELSEIF NEW.class = 'natural' and NEW.type in ('coastline') THEN
-      RETURN NULL;
     ELSEIF NEW.class = 'mountain_pass' THEN
         NEW.rank_search := 20;
         NEW.rank_address := 0;
@@ -1165,7 +1169,7 @@ BEGIN
   END IF;
 
   -- a country code make no sense below rank 4 (country)
-  IF NEW.rank_address < 4 THEN
+  IF NEW.rank_search < 4 THEN
     NEW.calculated_country_code := NULL;
   END IF;
 
@@ -1178,53 +1182,54 @@ BEGIN
 
   RETURN NEW; -- @DIFFUPDATES@ The following is not needed until doing diff updates, and slows the main index process down
 
-  IF (ST_GeometryType(NEW.geometry) in ('ST_Polygon','ST_MultiPolygon') AND ST_IsValid(NEW.geometry)) THEN
-    -- Performance: We just can't handle re-indexing for country level changes
-    IF st_area(NEW.geometry) < 1 THEN
-      -- mark items within the geometry for re-indexing
---    RAISE WARNING 'placex poly insert: % % % %',NEW.osm_type,NEW.osm_id,NEW.class,NEW.type;
-
-      -- work around bug in postgis, this may have been fixed in 2.0.0 (see http://trac.osgeo.org/postgis/ticket/547)
-      update placex set indexed_status = 2 where (st_covers(NEW.geometry, placex.geometry) OR ST_Intersects(NEW.geometry, placex.geometry)) 
-       AND rank_search > NEW.rank_search and indexed_status = 0 and ST_geometrytype(placex.geometry) = 'ST_Point' and (rank_search < 28 or name is not null or (NEW.rank_search >= 16 and addr_place is not null));
-      update placex set indexed_status = 2 where (st_covers(NEW.geometry, placex.geometry) OR ST_Intersects(NEW.geometry, placex.geometry)) 
-       AND rank_search > NEW.rank_search and indexed_status = 0 and ST_geometrytype(placex.geometry) != 'ST_Point' and (rank_search < 28 or name is not null or (NEW.rank_search >= 16 and addr_place is not null));
-    END IF;
-  ELSE
-    -- mark nearby items for re-indexing, where 'nearby' depends on the features rank_search and is a complete guess :(
-    diameter := 0;
-    -- 16 = city, anything higher than city is effectively ignored (polygon required!)
-    IF NEW.type='postcode' THEN
-      diameter := 0.05;
-    ELSEIF NEW.rank_search < 16 THEN
+  IF NEW.rank_address > 0 THEN
+    IF (ST_GeometryType(NEW.geometry) in ('ST_Polygon','ST_MultiPolygon') AND ST_IsValid(NEW.geometry)) THEN
+      -- Performance: We just can't handle re-indexing for country level changes
+      IF st_area(NEW.geometry) < 1 THEN
+        -- mark items within the geometry for re-indexing
+  --    RAISE WARNING 'placex poly insert: % % % %',NEW.osm_type,NEW.osm_id,NEW.class,NEW.type;
+
+        -- work around bug in postgis, this may have been fixed in 2.0.0 (see http://trac.osgeo.org/postgis/ticket/547)
+        update placex set indexed_status = 2 where (st_covers(NEW.geometry, placex.geometry) OR ST_Intersects(NEW.geometry, placex.geometry)) 
+         AND rank_search > NEW.rank_search and indexed_status = 0 and ST_geometrytype(placex.geometry) = 'ST_Point' and (rank_search < 28 or name is not null or (NEW.rank_search >= 16 and addr_place is not null));
+        update placex set indexed_status = 2 where (st_covers(NEW.geometry, placex.geometry) OR ST_Intersects(NEW.geometry, placex.geometry)) 
+         AND rank_search > NEW.rank_search and indexed_status = 0 and ST_geometrytype(placex.geometry) != 'ST_Point' and (rank_search < 28 or name is not null or (NEW.rank_search >= 16 and addr_place is not null));
+      END IF;
+    ELSE
+      -- mark nearby items for re-indexing, where 'nearby' depends on the features rank_search and is a complete guess :(
       diameter := 0;
-    ELSEIF NEW.rank_search < 18 THEN
-      diameter := 0.1;
-    ELSEIF NEW.rank_search < 20 THEN
-      diameter := 0.05;
-    ELSEIF NEW.rank_search = 21 THEN
-      diameter := 0.001;
-    ELSEIF NEW.rank_search < 24 THEN
-      diameter := 0.02;
-    ELSEIF NEW.rank_search < 26 THEN
-      diameter := 0.002; -- 100 to 200 meters
-    ELSEIF NEW.rank_search < 28 THEN
-      diameter := 0.001; -- 50 to 100 meters
-    END IF;
-    IF diameter > 0 THEN
---      RAISE WARNING 'placex point insert: % % % % %',NEW.osm_type,NEW.osm_id,NEW.class,NEW.type,diameter;
-      IF NEW.rank_search >= 26 THEN
-        -- roads may cause reparenting for >27 rank places
-        update placex set indexed_status = 2 where indexed_status = 0 and rank_search > NEW.rank_search and ST_DWithin(placex.geometry, NEW.geometry, diameter);
-      ELSEIF NEW.rank_search >= 16 THEN
-        -- up to rank 16, street-less addresses may need reparenting
-        update placex set indexed_status = 2 where indexed_status = 0 and rank_search > NEW.rank_search and ST_DWithin(placex.geometry, NEW.geometry, diameter) and (rank_search < 28 or name is not null or addr_place is not null);
-      ELSE
-        -- for all other places the search terms may change as well
-        update placex set indexed_status = 2 where indexed_status = 0 and rank_search > NEW.rank_search and ST_DWithin(placex.geometry, NEW.geometry, diameter) and (rank_search < 28 or name is not null);
+      -- 16 = city, anything higher than city is effectively ignored (polygon required!)
+      IF NEW.type='postcode' THEN
+        diameter := 0.05;
+      ELSEIF NEW.rank_search < 16 THEN
+        diameter := 0;
+      ELSEIF NEW.rank_search < 18 THEN
+        diameter := 0.1;
+      ELSEIF NEW.rank_search < 20 THEN
+        diameter := 0.05;
+      ELSEIF NEW.rank_search = 21 THEN
+        diameter := 0.001;
+      ELSEIF NEW.rank_search < 24 THEN
+        diameter := 0.02;
+      ELSEIF NEW.rank_search < 26 THEN
+        diameter := 0.002; -- 100 to 200 meters
+      ELSEIF NEW.rank_search < 28 THEN
+        diameter := 0.001; -- 50 to 100 meters
+      END IF;
+      IF diameter > 0 THEN
+  --      RAISE WARNING 'placex point insert: % % % % %',NEW.osm_type,NEW.osm_id,NEW.class,NEW.type,diameter;
+        IF NEW.rank_search >= 26 THEN
+          -- roads may cause reparenting for >27 rank places
+          update placex set indexed_status = 2 where indexed_status = 0 and rank_search > NEW.rank_search and ST_DWithin(placex.geometry, NEW.geometry, diameter);
+        ELSEIF NEW.rank_search >= 16 THEN
+          -- up to rank 16, street-less addresses may need reparenting
+          update placex set indexed_status = 2 where indexed_status = 0 and rank_search > NEW.rank_search and ST_DWithin(placex.geometry, NEW.geometry, diameter) and (rank_search < 28 or name is not null or addr_place is not null);
+        ELSE
+          -- for all other places the search terms may change as well
+          update placex set indexed_status = 2 where indexed_status = 0 and rank_search > NEW.rank_search and ST_DWithin(placex.geometry, NEW.geometry, diameter) and (rank_search < 28 or name is not null);
+        END IF;
       END IF;
     END IF;
-
   END IF;
 
    -- add to tables for special search
@@ -1274,6 +1279,7 @@ DECLARE
   search_maxrank INTEGER;
   address_maxrank INTEGER;
   address_street_word_id INTEGER;
+  address_street_word_ids INTEGER[];
   parent_place_id_rank BIGINT;
   
   isin TEXT[];
@@ -1343,12 +1349,26 @@ BEGIN
     NEW.centroid := null;
 
     -- reclaculate country and partition
-    IF NEW.rank_search >= 4 THEN
-      NEW.calculated_country_code := lower(get_country_code(place_centroid));
+    IF NEW.rank_search = 4 THEN
+      -- for countries, believe the mapped country code,
+      -- so that we remain in the right partition if the boundaries
+      -- suddenly expand.
+      NEW.partition := get_partition(lower(NEW.country_code));
+      IF NEW.partition = 0 THEN
+        NEW.calculated_country_code := lower(get_country_code(place_centroid));
+        NEW.partition := get_partition(NEW.calculated_country_code);
+      ELSE
+        NEW.calculated_country_code := lower(NEW.country_code);
+      END IF;
     ELSE
-      NEW.calculated_country_code := NULL;
+      IF NEW.rank_search > 4 THEN
+        --NEW.calculated_country_code := lower(get_country_code(NEW.geometry, NEW.country_code));
+        NEW.calculated_country_code := lower(get_country_code(place_centroid));
+      ELSE
+        NEW.calculated_country_code := NULL;
+      END IF;
+      NEW.partition := get_partition(NEW.calculated_country_code);
     END IF;
-    NEW.partition := get_partition(place_centroid, NEW.calculated_country_code);
     NEW.geometry_sector := geometry_sector(NEW.partition, place_centroid);
 
     -- Adding ourselves to the list simplifies address calculations later
@@ -1416,7 +1436,7 @@ BEGIN
               IF NEW.parent_place_id IS NULL AND relation.members[i+1] = 'street' THEN
 --RAISE WARNING 'node in relation %',relation;
                 SELECT place_id from placex where osm_type='W' and osm_id = substring(relation.members[i],2,200)::bigint 
-                  and rank_search = 26 INTO NEW.parent_place_id;
+                  and rank_search = 26 and name is not null INTO NEW.parent_place_id;
               END IF;
             END LOOP;
           END IF;
@@ -1445,7 +1465,7 @@ BEGIN
                     IF NEW.parent_place_id IS NULL AND relation.members[i+1] = 'street' THEN
     --RAISE WARNING 'node in way that is in a relation %',relation;
                       SELECT place_id from placex where osm_type='W' and osm_id = substring(relation.members[i],2,200)::bigint 
-                        and rank_search = 26 INTO NEW.parent_place_id;
+                        and rank_search = 26 and name is not null INTO NEW.parent_place_id;
                     END IF;
                   END LOOP;
                 END IF;
@@ -1495,7 +1515,7 @@ BEGIN
               IF NEW.parent_place_id IS NULL AND relation.members[i+1] = 'street' THEN
 --RAISE WARNING 'way that is in a relation %',relation;
                 SELECT place_id from placex where osm_type='W' and osm_id = substring(relation.members[i],2,200)::bigint
-                  and rank_search = 26 INTO NEW.parent_place_id;
+                  and rank_search = 26 and name is not null INTO NEW.parent_place_id;
               END IF;
             END LOOP;
           END IF;
@@ -1505,18 +1525,18 @@ BEGIN
 --RAISE WARNING 'x3 %',NEW.parent_place_id;
 
       IF NEW.parent_place_id IS NULL AND NEW.street IS NOT NULL THEN
-        address_street_word_id := get_name_id(make_standard_name(NEW.street));
-        IF address_street_word_id IS NOT NULL THEN
-          FOR location IN SELECT * from getNearestNamedRoadFeature(NEW.partition, place_centroid, address_street_word_id) LOOP
+        address_street_word_ids := get_name_ids(make_standard_name(NEW.street));
+        IF address_street_word_ids IS NOT NULL THEN
+          FOR location IN SELECT * from getNearestNamedRoadFeature(NEW.partition, place_centroid, address_street_word_ids) LOOP
               NEW.parent_place_id := location.place_id;
           END LOOP;
         END IF;
       END IF;
 
       IF NEW.parent_place_id IS NULL AND NEW.addr_place IS NOT NULL THEN
-        address_street_word_id := get_name_id(make_standard_name(NEW.addr_place));
-        IF address_street_word_id IS NOT NULL THEN
-          FOR location IN SELECT * from getNearestNamedPlaceFeature(NEW.partition, place_centroid, address_street_word_id) LOOP
+        address_street_word_ids := get_name_id(make_standard_name(NEW.addr_place));
+        IF address_street_word_ids IS NOT NULL THEN
+          FOR location IN SELECT * from getNearestNamedPlaceFeature(NEW.partition, place_centroid, address_street_word_ids) LOOP
             NEW.parent_place_id := location.place_id;
           END LOOP;
         END IF;
@@ -1600,17 +1620,13 @@ BEGIN
             END IF;
 
             -- merge in extra tags
-            IF NOT linkedPlacex.extratags IS NULL THEN
-              NEW.extratags := linkedPlacex.extratags || NEW.extratags;
-            END IF;
-
-            IF NOT NEW.extratags ? linkedPlacex.class THEN
-              NEW.extratags := NEW.extratags || hstore(linkedPlacex.class, linkedPlacex.type);
-            END IF;
+            NEW.extratags := hstore(linkedPlacex.class, linkedPlacex.type) || coalesce(linkedPlacex.extratags, ''::hstore) || coalesce(NEW.extratags, ''::hstore);
 
             -- mark the linked place (excludes from search results)
             UPDATE placex set linked_place_id = NEW.place_id where place_id = linkedPlacex.place_id;
 
+            -- keep a note of the node id in case we need it for wikipedia in a bit
+            linked_node_id := linkedPlacex.osm_id;
           END LOOP;
 
         END LOOP;
@@ -1639,13 +1655,7 @@ BEGIN
                 END IF;
 
                 -- merge in extra tags
-                IF NOT linkedPlacex.extratags IS NULL THEN
-                  NEW.extratags := linkedPlacex.extratags || NEW.extratags;
-                END IF;
-
-                IF NOT NEW.extratags ? linkedPlacex.class THEN
-                  NEW.extratags := NEW.extratags || hstore(linkedPlacex.class, linkedPlacex.type);
-                END IF;
+                NEW.extratags := hstore(linkedPlacex.class, linkedPlacex.type) || coalesce(linkedPlacex.extratags, ''::hstore) || coalesce(NEW.extratags, ''::hstore);
 
                 -- mark the linked place (excludes from search results)
                 UPDATE placex set linked_place_id = NEW.place_id where place_id = linkedPlacex.place_id;
@@ -1687,11 +1697,7 @@ BEGIN
           name_vector := make_keywords(NEW.name);
 
           -- merge in extra tags
-          NEW.extratags := linkedPlacex.extratags || NEW.extratags;
-
-          IF NOT NEW.extratags ? linkedPlacex.class THEN
-            NEW.extratags := NEW.extratags || hstore(linkedPlacex.class, linkedPlacex.type);
-          END IF;
+          NEW.extratags := hstore(linkedPlacex.class, linkedPlacex.type) || coalesce(linkedPlacex.extratags, ''::hstore) || coalesce(NEW.extratags, ''::hstore);
 
           -- mark the linked place (excludes from search results)
           UPDATE placex set linked_place_id = NEW.place_id where place_id = linkedPlacex.place_id;
@@ -1988,15 +1994,18 @@ LANGUAGE plpgsql;
 CREATE OR REPLACE FUNCTION place_delete() RETURNS TRIGGER
   AS $$
 DECLARE
-  placeid BIGINT;
+  has_rank BOOLEAN;
 BEGIN
 
   --DEBUG: RAISE WARNING 'delete: % % % %',OLD.osm_type,OLD.osm_id,OLD.class,OLD.type;
 
   -- deleting large polygons can have a massive effect on the system - require manual intervention to let them through
   IF st_area(OLD.geometry) > 2 and st_isvalid(OLD.geometry) THEN
-    insert into import_polygon_delete values (OLD.osm_type,OLD.osm_id,OLD.class,OLD.type);
-    RETURN NULL;
+    SELECT bool_or(not (rank_address = 0 or rank_address > 26)) as ranked FROM placex WHERE osm_type = OLD.osm_type and osm_id = OLD.osm_id and class = OLD.class and type = OLD.type INTO has_rank;
+    IF has_rank THEN
+      insert into import_polygon_delete values (OLD.osm_type,OLD.osm_id,OLD.class,OLD.type);
+      RETURN NULL;
+    END IF;
   END IF;
 
   -- mark for delete
@@ -2028,6 +2037,11 @@ BEGIN
     --DEBUG: RAISE WARNING '%', existingplacex;
   END IF;
 
+  -- remove operator tag for most places, messes too much with search_name indexes
+  IF NEW.class not in ('amenity', 'shop') THEN
+    NEW.name := delete(NEW.name, 'operator');
+  END IF;
+
   -- Just block these - lots and pointless
   IF NEW.class in ('landuse','natural') and NEW.name is null THEN
     -- if the name tag was removed, older versions might still be lurking in the place table
@@ -2218,6 +2232,12 @@ BEGIN
 
   END IF;
 
+  -- refuse to update multiplpoygons with too many objects, too much of a performance hit
+  IF ST_NumGeometries(NEW.geometry) > 2000 THEN
+    RAISE WARNING 'Dropping update of % % because of geometry complexity.', NEW.osm_type, NEW.osm_id;
+    RETURN NULL;
+  END IF;
+
   IF coalesce(existing.name::text, '') != coalesce(NEW.name::text, '')
      OR coalesce(existing.extratags::text, '') != coalesce(NEW.extratags::text, '')
      OR coalesce(existing.housenumber, '') != coalesce(NEW.housenumber, '')
@@ -2803,7 +2823,7 @@ DECLARE
 BEGIN
 
   place_centroid := ST_Centroid(pointgeo);
-  out_partition := get_partition(place_centroid, in_countrycode);
+  out_partition := get_partition(in_countrycode);
   out_parent_place_id := null;
 
   address_street_word_id := get_name_id(make_standard_name(in_street));
@@ -3077,6 +3097,7 @@ DECLARE
   diameter FLOAT;
   rank INTEGER;
 BEGIN
+  UPDATE placex SET indexed_status = 2 WHERE place_id = placeid;
   SELECT geometry, rank_search FROM placex WHERE place_id = placeid INTO placegeom, rank;
   IF placegeom IS NOT NULL AND ST_IsValid(placegeom) THEN
     IF ST_GeometryType(placegeom) in ('ST_Polygon','ST_MultiPolygon') THEN