overhaul address computation

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

diff --git a/sql/functions/placex_triggers.sql b/sql/functions/placex_triggers.sql
index 3ddb97c04240225d87345360bbddf7e3ddbd52d7..aac7cafcdca0cb24768ba2ec874b3ec0c61b9a22 100644 (file)
--- a/sql/functions/placex_triggers.sql
+++ b/sql/functions/placex_triggers.sql
@@ -92,7 +92,18 @@ BEGIN
     END IF;
 
     IF fallback THEN
     END IF;
 
     IF fallback THEN

-      IF ST_Area(bbox) < 0.005 THEN
+      IF addr_street is null and addr_place is not null THEN
+        -- The address is attached to a place we don't know.
+        -- Instead simply use the containing area with the largest rank.
+        FOR location IN
+          SELECT place_id FROM placex
+            WHERE bbox @ geometry AND _ST_Covers(geometry, ST_Centroid(bbox))
+                  AND rank_address between 5 and 25
+            ORDER BY rank_address desc
+        LOOP
+            RETURN location.place_id;
+        END LOOP;
+      ELSEIF ST_Area(bbox) < 0.005 THEN

         -- for smaller features get the nearest road
         SELECT getNearestRoadPlaceId(poi_partition, bbox) INTO parent_place_id;
         --DEBUG: RAISE WARNING 'Checked for nearest way (%)', parent_place_id;
         -- for smaller features get the nearest road
         SELECT getNearestRoadPlaceId(poi_partition, bbox) INTO parent_place_id;
         --DEBUG: RAISE WARNING 'Checked for nearest way (%)', parent_place_id;


@@ -200,7 +211,11 @@ BEGIN
     FOR linked_placex IN
       SELECT placex.* from placex
       WHERE make_standard_name(name->'name') = bnd_name
     FOR linked_placex IN
       SELECT placex.* from placex
       WHERE make_standard_name(name->'name') = bnd_name

-        AND ((bnd.rank_address > 0 and placex.rank_address = bnd.rank_address)
+        AND ((bnd.rank_address > 0
+              and bnd.rank_address = (compute_place_rank(placex.country_code,
+                                                         'N', placex.class,
+                                                         placex.type, 15::SMALLINT,
+                                                         false, placex.postcode)).address_rank)

              OR (bnd.rank_address = 0 and placex.rank_search = bnd.rank_search))
         AND placex.osm_type = 'N'
         AND placex.rank_search < 26 -- needed to select the right index
              OR (bnd.rank_address = 0 and placex.rank_search = bnd.rank_search))
         AND placex.osm_type = 'N'
         AND placex.rank_search < 26 -- needed to select the right index


@@ -224,7 +239,7 @@ LANGUAGE plpgsql STABLE;
 -- \param maxrank       Rank of the place. All address features must have
 --                      a search rank lower than the given rank.
 -- \param address       Address terms for the place.
 -- \param maxrank       Rank of the place. All address features must have
 --                      a search rank lower than the given rank.
 -- \param address       Address terms for the place.

--- \param geoemtry      Geometry to which the address objects should be close.
+-- \param geometry      Geometry to which the address objects should be close.

 --
 -- \retval parent_place_id  Place_id of the address object that is the direct
 --                          ancestor.
 --
 -- \retval parent_place_id  Place_id of the address object that is the direct
 --                          ancestor.


@@ -244,21 +259,16 @@ CREATE OR REPLACE FUNCTION insert_addresslines(obj_place_id BIGINT,
                                                OUT nameaddress_vector INT[])
   AS $$
 DECLARE
                                                OUT nameaddress_vector INT[])
   AS $$
 DECLARE

-  current_rank_address INTEGER := 0;
-  location_distance FLOAT := 0;
-  location_parent GEOMETRY := NULL;
-  parent_place_id_rank SMALLINT := 0;
+  address_havelevel BOOLEAN[];

 
   location_isaddress BOOLEAN;
 
   location_isaddress BOOLEAN;

-
-  address_havelevel BOOLEAN[];
-  location_keywords INT[];
+  current_boundary GEOMETRY := NULL;
+  current_node_area GEOMETRY := NULL;

 
   location RECORD;
   addr_item RECORD;
 
   isin_tokens INT[];
 
   location RECORD;
   addr_item RECORD;
 
   isin_tokens INT[];

-  isin TEXT[];

 BEGIN
   parent_place_id := 0;
   nameaddress_vector := '{}'::int[];
 BEGIN
   parent_place_id := 0;
   nameaddress_vector := '{}'::int[];


@@ -281,100 +291,77 @@ BEGIN
         END IF;
       END IF;
     END LOOP;
         END IF;
       END IF;
     END LOOP;

-
-    IF address ? 'is_in' THEN
-      -- is_in items need splitting
-      isin := regexp_split_to_array(address->'is_in', E'[;,]');
-      IF array_upper(isin, 1) IS NOT NULL THEN
-        FOR i IN 1..array_upper(isin, 1) LOOP
-          isin_tokens := array_merge(isin_tokens,
-                                     word_ids_from_name(isin[i]));
-
-          -- merge word into address vector
-          IF NOT %REVERSE-ONLY% THEN
-            nameaddress_vector := array_merge(nameaddress_vector,
-                                              addr_ids_from_name(isin[i]));
-          END IF;
-        END LOOP;
-      END IF;
-    END IF;

   END IF;
   IF NOT %REVERSE-ONLY% THEN
     nameaddress_vector := array_merge(nameaddress_vector, isin_tokens);
   END IF;
 
   ---- now compute the address terms
   END IF;
   IF NOT %REVERSE-ONLY% THEN
     nameaddress_vector := array_merge(nameaddress_vector, isin_tokens);
   END IF;
 
   ---- now compute the address terms

-  FOR i IN 1..28 LOOP
+  FOR i IN 1..maxrank LOOP

     address_havelevel[i] := false;
   END LOOP;
 
   FOR location IN
     address_havelevel[i] := false;
   END LOOP;
 
   FOR location IN

-    SELECT * FROM getNearFeatures(partition, geometry, maxrank, isin_tokens)
+    SELECT * FROM getNearFeatures(partition, geometry, maxrank)
+    ORDER BY rank_address, isin_tokens && keywords desc, isguess asc,
+             distance *
+               CASE WHEN rank_address = 16 AND rank_search = 15 THEN 0.2
+                    WHEN rank_address = 16 AND rank_search = 16 THEN 0.25
+                    WHEN rank_address = 16 AND rank_search = 18 THEN 0.5
+                    ELSE 1 END ASC

   LOOP
   LOOP

-    IF location.rank_address != current_rank_address THEN
-      current_rank_address := location.rank_address;
-      IF location.isguess THEN
-        location_distance := location.distance * 1.5;
-      ELSE
-        IF location.rank_address <= 12 THEN
-          -- for county and above, if we have an area consider that exact
-          -- (It would be nice to relax the constraint for places close to
-          --  the boundary but we'd need the exact geometry for that. Too
-          --  expensive.)
-          location_distance = 0;
-        ELSE
-          -- Below county level remain slightly fuzzy.
-          location_distance := location.distance * 0.5;
-        END IF;
+    -- Ignore all place nodes that do not fit in a lower level boundary.
+    CONTINUE WHEN location.isguess
+                  and current_boundary is not NULL
+                  and not ST_Contains(current_boundary, location.centroid);
+
+    -- If this is the first item in the rank, then assume it is the address.
+    location_isaddress := not address_havelevel[location.rank_address];
+
+    -- Further sanity checks to ensure that the address forms a sane hierarchy.
+    IF location_isaddress THEN
+      IF location.isguess and current_node_area is not NULL THEN
+        location_isaddress := ST_Contains(current_node_area, location.centroid);
+      END IF;
+      IF not location.isguess and current_boundary is not NULL
+         and location.rank_address != 11 AND location.rank_address != 5 THEN
+        location_isaddress := ST_Contains(current_boundary, location.centroid);

       END IF;
       END IF;

-    ELSE
-      CONTINUE WHEN location.keywords <@ location_keywords;

     END IF;
 
     END IF;
 

-    IF location.distance < location_distance OR NOT location.isguess THEN
-      location_keywords := location.keywords;
-
-      location_isaddress := NOT address_havelevel[location.rank_address];
-      --DEBUG: RAISE WARNING 'should be address: %, is guess: %, rank: %', location_isaddress, location.isguess, location.rank_address;
-      IF location_isaddress AND location.isguess AND location_parent IS NOT NULL THEN
-          location_isaddress := ST_Contains(location_parent, location.centroid);
-      END IF;
+    IF location_isaddress THEN
+      address_havelevel[location.rank_address] := true;
+      parent_place_id := location.place_id;

 
 

-      --DEBUG: RAISE WARNING '% isaddress: %', location.place_id, location_isaddress;
-      -- Add it to the list of search terms
-      IF NOT %REVERSE-ONLY% THEN
-          nameaddress_vector := array_merge(nameaddress_vector,
-                                            location.keywords::integer[]);
+      -- Set postcode if we have one.
+      -- (Returned will be the highest ranking one.)
+      IF location.postcode is not NULL THEN
+        postcode = location.postcode;

       END IF;
 
       END IF;
 

-      INSERT INTO place_addressline (place_id, address_place_id, fromarea,
-                                     isaddress, distance, cached_rank_address)
-        VALUES (obj_place_id, location.place_id, true,
-                location_isaddress, location.distance, location.rank_address);
-
-      IF location_isaddress THEN
-        -- add postcode if we have one
-        -- (If multiple postcodes are available, we end up with the highest ranking one.)
-        IF location.postcode is not null THEN
-            postcode = location.postcode;
-        END IF;
-
-        address_havelevel[location.rank_address] := true;
-        -- add a hack against postcode ranks
-        IF NOT location.isguess
-           AND location.rank_address != 11 AND location.rank_address != 5
-        THEN
+      -- Recompute the areas we need for hierarchy sanity checks.
+      IF location.rank_address != 11 AND location.rank_address != 5 THEN
+        IF location.isguess THEN
+          current_node_area := place_node_fuzzy_area(location.centroid,
+                                                     location.rank_search);
+        ELSE
+          current_node_area := NULL;

           SELECT p.geometry FROM placex p
           SELECT p.geometry FROM placex p

-            WHERE p.place_id = location.place_id INTO location_parent;
-        END IF;
-
-        IF location.rank_address > parent_place_id_rank THEN
-          parent_place_id = location.place_id;
-          parent_place_id_rank = location.rank_address;
+              WHERE p.place_id = location.place_id INTO current_boundary;

         END IF;
       END IF;
     END IF;
 
         END IF;
       END IF;
     END IF;
 

+    -- Add it to the list of search terms
+    IF NOT %REVERSE-ONLY% THEN
+      nameaddress_vector := array_merge(nameaddress_vector,
+                                        location.keywords::integer[]);
+    END IF;
+
+    INSERT INTO place_addressline (place_id, address_place_id, fromarea,
+                                     isaddress, distance, cached_rank_address)
+        VALUES (obj_place_id, location.place_id, true,
+                location_isaddress, location.distance, location.rank_address);

   END LOOP;
 END;
 $$
   END LOOP;
 END;
 $$


@@ -417,7 +404,12 @@ BEGIN
 
       NEW.name := hstore('ref', NEW.address->'postcode');
 
 
       NEW.name := hstore('ref', NEW.address->'postcode');
 

-    ELSEIF NEW.class = 'boundary' AND NOT is_area THEN
+    ELSEIF NEW.class = 'highway' AND is_area AND NEW.name is null
+           AND NEW.extratags ? 'area' AND NEW.extratags->'area' = 'yes'
+    THEN
+        RETURN NULL;
+    ELSEIF NEW.class = 'boundary' AND NOT is_area
+    THEN

         RETURN NULL;
     ELSEIF NEW.class = 'boundary' AND NEW.type = 'administrative'
            AND NEW.admin_level <= 4 AND NEW.osm_type = 'W'
         RETURN NULL;
     ELSEIF NEW.class = 'boundary' AND NEW.type = 'administrative'
            AND NEW.admin_level <= 4 AND NEW.osm_type = 'W'


@@ -453,11 +445,16 @@ BEGIN
         -- mark items within the geometry for re-indexing
   --    RAISE WARNING 'placex poly insert: % % % %',NEW.osm_type,NEW.osm_id,NEW.class,NEW.type;
 
         -- 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 address ? 'place'));
-        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 address ? 'place'));
+        UPDATE placex SET indexed_status = 2
+         WHERE ST_Intersects(NEW.geometry, placex.geometry)
+               and indexed_status = 0
+               and ((rank_address = 0 and rank_search > NEW.rank_address)
+                    or rank_address > NEW.rank_address
+                    or (class = 'place' and osm_type = 'N')
+                   )
+               and (rank_search < 28
+                    or name is not null
+                    or (NEW.rank_address >= 16 and address ? 'place'));

       END IF;
     ELSE
       -- mark nearby items for re-indexing, where 'nearby' depends on the features rank_search and is a complete guess :(
       END IF;
     ELSE
       -- mark nearby items for re-indexing, where 'nearby' depends on the features rank_search and is a complete guess :(


@@ -541,6 +538,9 @@ DECLARE
 
   name_vector INTEGER[];
   nameaddress_vector INTEGER[];
 
   name_vector INTEGER[];
   nameaddress_vector INTEGER[];

+  addr_nameaddress_vector INTEGER[];
+
+  inherited_address HSTORE;

 
   linked_node_id BIGINT;
   linked_importance FLOAT;
 
   linked_node_id BIGINT;
   linked_importance FLOAT;


@@ -575,12 +575,22 @@ BEGIN
   -- update not necessary for osmline, cause linked_place_id does not exist
 
   NEW.extratags := NEW.extratags - 'linked_place'::TEXT;
   -- update not necessary for osmline, cause linked_place_id does not exist
 
   NEW.extratags := NEW.extratags - 'linked_place'::TEXT;

+  NEW.address := NEW.address - '_unlisted_place'::TEXT;

 
   IF NEW.linked_place_id is not null THEN
     --DEBUG: RAISE WARNING 'place already linked to %', NEW.linked_place_id;
     RETURN NEW;
   END IF;
 
 
   IF NEW.linked_place_id is not null THEN
     --DEBUG: RAISE WARNING 'place already linked to %', NEW.linked_place_id;
     RETURN NEW;
   END IF;
 

+  -- Postcodes are just here to compute the centroids. They are not searchable
+  -- unless they are a boundary=postal_code.
+  -- There was an error in the style so that boundary=postal_code used to be
+  -- imported as place=postcode. That's why relations are allowed to pass here.
+  -- This can go away in a couple of versions.
+  IF NEW.class = 'place'  and NEW.type = 'postcode' and NEW.osm_type != 'R' THEN
+    RETURN NEW;
+  END IF;
+

   -- Speed up searches - just use the centroid of the feature
   -- cheaper but less acurate
   NEW.centroid := ST_PointOnSurface(NEW.geometry);
   -- Speed up searches - just use the centroid of the feature
   -- cheaper but less acurate
   NEW.centroid := ST_PointOnSurface(NEW.geometry);


@@ -607,6 +617,21 @@ BEGIN
         NEW.rank_address := parent_address_level + 2;
       END IF;
     END IF;
         NEW.rank_address := parent_address_level + 2;
       END IF;
     END IF;

+  -- If a place node is contained in a admin boundary with the same address level
+  -- and has not been linked, then make the node a subpart by increasing the
+  -- address rank (city level and above).
+  ELSEIF NEW.class = 'place' and NEW.osm_type = 'N'
+     and NEW.rank_address between 16 and 23
+  THEN
+    FOR location IN
+        SELECT rank_address FROM placex
+        WHERE osm_type = 'R' and class = 'boundary' and type = 'administrative'
+              and rank_address = NEW.rank_address
+              and geometry && NEW.centroid and _ST_Covers(geometry, NEW.centroid)
+        LIMIT 1
+    LOOP
+      NEW.rank_address = NEW.rank_address + 2;
+    END LOOP;

   ELSE
     parent_address_level := 3;
   END IF;
   ELSE
     parent_address_level := 3;
   END IF;


@@ -634,7 +659,7 @@ BEGIN
       addr_street := NEW.address->'street';
       addr_place := NEW.address->'place';
 
       addr_street := NEW.address->'street';
       addr_place := NEW.address->'place';
 

-      IF NEW.address ? 'postcode' and NEW.address->'postcode' not similar to '%(,|;)%' THEN
+      IF NEW.address ? 'postcode' and NEW.address->'postcode' not similar to '%(:|,|;)%' THEN

         i := getorcreate_postcode_id(NEW.address->'postcode');
       END IF;
   END IF;
         i := getorcreate_postcode_id(NEW.address->'postcode');
       END IF;
   END IF;


@@ -694,13 +719,14 @@ BEGIN
 
   -- ---------------------------------------------------------------------------
   -- For low level elements we inherit from our parent road
 
   -- ---------------------------------------------------------------------------
   -- For low level elements we inherit from our parent road

-  IF (NEW.rank_search > 27 OR (NEW.type = 'postcode' AND NEW.rank_search = 25)) THEN
+  IF NEW.rank_search > 27 THEN

 
     --DEBUG: RAISE WARNING 'finding street for % %', NEW.osm_type, NEW.osm_id;
     NEW.parent_place_id := null;
 
     -- if we have a POI and there is no address information,
     -- see if we can get it from a surrounding building
 
     --DEBUG: RAISE WARNING 'finding street for % %', NEW.osm_type, NEW.osm_id;
     NEW.parent_place_id := null;
 
     -- if we have a POI and there is no address information,
     -- see if we can get it from a surrounding building

+    inherited_address := ''::HSTORE;

     IF NEW.osm_type = 'N' AND addr_street IS NULL AND addr_place IS NULL
        AND NEW.housenumber IS NULL THEN
       FOR location IN
     IF NEW.osm_type = 'N' AND addr_street IS NULL AND addr_place IS NULL
        AND NEW.housenumber IS NULL THEN
       FOR location IN


@@ -715,6 +741,7 @@ BEGIN
         NEW.housenumber := location.address->'housenumber';
         addr_street := location.address->'street';
         addr_place := location.address->'place';
         NEW.housenumber := location.address->'housenumber';
         addr_street := location.address->'street';
         addr_place := location.address->'place';

+        inherited_address := location.address;

       END LOOP;
     END IF;
 
       END LOOP;
     END IF;
 


@@ -729,9 +756,18 @@ BEGIN
     IF NEW.parent_place_id is not null THEN
 
       -- Get the details of the parent road
     IF NEW.parent_place_id is not null THEN
 
       -- Get the details of the parent road

-      SELECT p.country_code, p.postcode FROM placex p
+      SELECT p.country_code, p.postcode, p.name FROM placex p

        WHERE p.place_id = NEW.parent_place_id INTO location;
 
        WHERE p.place_id = NEW.parent_place_id INTO location;
 

+      IF addr_street is null and addr_place is not null THEN
+        -- Check if the addr:place tag is part of the parent name
+        SELECT count(*) INTO i
+          FROM svals(location.name) AS pname WHERE pname = addr_place;
+        IF i = 0 THEN
+          NEW.address = NEW.address || hstore('_unlisted_place', addr_place);
+        END IF;
+      END IF;
+

       NEW.country_code := location.country_code;
       --DEBUG: RAISE WARNING 'Got parent details from search name';
 
       NEW.country_code := location.country_code;
       --DEBUG: RAISE WARNING 'Got parent details from search name';
 


@@ -745,33 +781,26 @@ BEGIN
         NEW.postcode := get_nearest_postcode(NEW.country_code, NEW.geometry);
       END IF;
 
         NEW.postcode := get_nearest_postcode(NEW.country_code, NEW.geometry);
       END IF;
 

-      -- If there is no name it isn't searchable, don't bother to create a search record
-      IF NEW.name is NULL THEN
-        --DEBUG: RAISE WARNING 'Not a searchable place % %', NEW.osm_type, NEW.osm_id;
-        return NEW;
-      END IF;
+      IF NEW.name is not NULL THEN
+          NEW.name := add_default_place_name(NEW.country_code, NEW.name);
+          name_vector := make_keywords(NEW.name);

 
 

-      NEW.name := add_default_place_name(NEW.country_code, NEW.name);
-      name_vector := make_keywords(NEW.name);
+          IF NEW.rank_search <= 25 and NEW.rank_address > 0 THEN
+            result := add_location(NEW.place_id, NEW.country_code, NEW.partition,
+                                   name_vector, NEW.rank_search, NEW.rank_address,
+                                   upper(trim(NEW.address->'postcode')), NEW.geometry);
+            --DEBUG: RAISE WARNING 'Place added to location table';
+          END IF;

 
 

-      -- Performance, it would be more acurate to do all the rest of the import
-      -- process but it takes too long
-      -- Just be happy with inheriting from parent road only
-      IF NEW.rank_search <= 25 and NEW.rank_address > 0 THEN
-        result := add_location(NEW.place_id, NEW.country_code, NEW.partition, name_vector, NEW.rank_search, NEW.rank_address, upper(trim(NEW.address->'postcode')), NEW.geometry);
-        --DEBUG: RAISE WARNING 'Place added to location table';

       END IF;
 
       END IF;
 

-      result := insertSearchName(NEW.partition, NEW.place_id, name_vector,
-                                 NEW.rank_search, NEW.rank_address, NEW.geometry);
-

       IF NOT %REVERSE-ONLY% THEN
       IF NOT %REVERSE-ONLY% THEN

-          -- Merge address from parent
-          SELECT array_merge(s.name_vector, s.nameaddress_vector)
-            INTO nameaddress_vector
-            FROM search_name s
-            WHERE s.place_id = NEW.parent_place_id;
+        SELECT * INTO name_vector, nameaddress_vector
+          FROM create_poi_search_terms(NEW.parent_place_id,
+                                       inherited_address || NEW.address,
+                                       NEW.housenumber, name_vector);

 
 

+        IF array_length(name_vector, 1) is not NULL THEN

           INSERT INTO search_name (place_id, search_rank, address_rank,
                                    importance, country_code, name_vector,
                                    nameaddress_vector, centroid)
           INSERT INTO search_name (place_id, search_rank, address_rank,
                                    importance, country_code, name_vector,
                                    nameaddress_vector, centroid)


@@ -780,8 +809,9 @@ BEGIN
                          nameaddress_vector, NEW.centroid);
           --DEBUG: RAISE WARNING 'Place added to search table';
         END IF;
                          nameaddress_vector, NEW.centroid);
           --DEBUG: RAISE WARNING 'Place added to search table';
         END IF;

+      END IF;

 
 

-      return NEW;
+      RETURN NEW;

     END IF;
 
   END IF;
     END IF;
 
   END IF;