computation of parent of interpolations

[nominatim.git] / sql / functions.sql

diff --git a/sql/functions.sql b/sql/functions.sql
index c2268f4379f302890ac7a9ac9912fdb1a8f3fe59..ab3571b0beb657695c466799539df1f705b65bc2 100644 (file)
--- a/sql/functions.sql
+++ b/sql/functions.sql
@@ -474,31 +474,6 @@ END;
 $$
 LANGUAGE plpgsql IMMUTABLE;
 
 $$
 LANGUAGE plpgsql IMMUTABLE;
 

-CREATE OR REPLACE FUNCTION get_word_score(wordscores wordscore[], words text[]) RETURNS integer
-  AS $$
-DECLARE
-  idxword integer;
-  idxscores integer;
-  result integer;
-BEGIN
-  IF (wordscores is null OR words is null) THEN
-    RETURN 0;
-  END IF;
-
-  result := 0;
-  FOR idxword in 1 .. array_upper(words, 1) LOOP
-    FOR idxscores in 1 .. array_upper(wordscores, 1) LOOP
-      IF wordscores[idxscores].word = words[idxword] THEN
-        result := result + wordscores[idxscores].score;
-      END IF;
-    END LOOP;
-  END LOOP;
-
-  RETURN result;
-END;
-$$
-LANGUAGE plpgsql IMMUTABLE;
-

 CREATE OR REPLACE FUNCTION get_country_code(place geometry) RETURNS TEXT
   AS $$
 DECLARE
 CREATE OR REPLACE FUNCTION get_country_code(place geometry) RETURNS TEXT
   AS $$
 DECLARE


@@ -548,21 +523,6 @@ BEGIN
     RETURN nearcountry.country_code;
   END LOOP;
 
     RETURN nearcountry.country_code;
   END LOOP;
 

-  -- WorldBoundaries data (second fallback - think there might be something broken in this data)
---  FOR nearcountry IN select country_code from country where st_covers(geometry, place_centre) limit 1
---  LOOP
---    RETURN nearcountry.country_code;
---  END LOOP;
-
---RAISE WARNING 'near country: %', ST_AsText(place_centre);
-
-  -- Still not in a country - try nearest within ~12 miles of a country
---  FOR nearcountry IN select country_code from country where st_distance(geometry, place_centre) < 0.5 
---    order by st_distance(geometry, place) limit 1
---  LOOP
---    RETURN nearcountry.country_code;
---  END LOOP;
-

   RETURN NULL;
 END;
 $$
   RETURN NULL;
 END;
 $$


@@ -787,11 +747,76 @@ END;
 $$
 LANGUAGE plpgsql;
 
 $$
 LANGUAGE plpgsql;
 

+-- find the parant road of an interpolation
+CREATE OR REPLACE FUNCTION get_interpolation_parent(wayid BIGINT, street TEXT, place TEXT,
+                                                    partition INTEGER, centroid GEOMETRY, geom GEOMETRY)
+RETURNS BIGINT AS $$
+DECLARE
+  addr_street TEXT;
+  addr_place TEXT;
+  parent_place_id BIGINT;
+  address_street_word_ids INTEGER[];
+
+  waynodes BIGINT[];
+
+  location RECORD;
+BEGIN
+  addr_street = street;
+  addr_place = place;
+
+  IF addr_street is null and addr_place is null THEN
+    select nodes from planet_osm_ways where id = wayid INTO waynodes;
+    FOR location IN SELECT street, addr_place from placex 
+                    where osm_type = 'N' and osm_id = ANY(nodes)
+                          and (street is not null or addr_place is not null)
+                          and indexed_status < 100
+                    limit 1 LOOP
+      addr_street = location.street;
+      addr_place = location.addr_place;
+    END LOOP;
+  END IF;
+
+  IF addr_street IS NOT NULL THEN
+    address_street_word_ids := get_name_ids(make_standard_name(addr_street));
+    IF address_street_word_ids IS NOT NULL THEN
+      FOR location IN SELECT place_id from getNearestNamedRoadFeature(partition, centroid, address_street_word_ids) LOOP
+        parent_place_id := location.place_id;
+      END LOOP;
+    END IF;
+  END IF;
+
+  IF parent_place_id IS NULL AND addr_place IS NOT NULL THEN
+    address_street_word_ids := get_name_ids(make_standard_name(addr_place));
+    IF address_street_word_ids IS NOT NULL THEN
+      FOR location IN SELECT place_id from getNearestNamedPlaceFeature(partition, centroid, address_street_word_ids) LOOP
+        parent_place_id := location.place_id;
+      END LOOP;
+    END IF;
+  END IF;

 
 

-CREATE OR REPLACE FUNCTION create_interpolation(wayid BIGINT, interpolationtype TEXT) RETURNS INTEGER
+  IF parent_place_id is null THEN
+    FOR location IN SELECT place_id FROM placex
+        WHERE ST_DWithin(geom, placex.geometry, 0.001) and placex.rank_search = 26
+        ORDER BY (ST_distance(placex.geometry, ST_LineInterpolatePoint(geom,0))+
+                  ST_distance(placex.geometry, ST_LineInterpolatePoint(geom,0.5))+
+                  ST_distance(placex.geometry, ST_LineInterpolatePoint(geom,1))) ASC limit 1
+    LOOP
+      parent_place_id := location.place_id;
+    END LOOP;
+  END IF;
+
+  RETURN parent_place_id;
+END;
+$$
+LANGUAGE plpgsql;
+
+CREATE OR REPLACE FUNCTION create_interpolation(wayid BIGINT, interpolationtype TEXT,
+                                                parent_place_id BIGINT, partition INTEGER,
+                                                country_code TEXT,  geometry_sector INTEGER,
+                                                defpostalcode TEXT, geom GEOMETRY) RETURNS INTEGER

   AS $$
 DECLARE
   AS $$
 DECLARE

-  
+

   newpoints INTEGER;
   waynodes BIGINT[];
   nodeid BIGINT;
   newpoints INTEGER;
   waynodes BIGINT[];
   nodeid BIGINT;


@@ -804,117 +829,100 @@ DECLARE
   originalnumberrange INTEGER;
   housenum INTEGER;
   linegeo GEOMETRY;
   originalnumberrange INTEGER;
   housenum INTEGER;
   linegeo GEOMETRY;

-  search_place_id BIGINT;
-  defpostalcode TEXT;
+  splitline GEOMETRY;
+  sectiongeo GEOMETRY;

 
 

-  havefirstpoint BOOLEAN;
-  linestr TEXT;

 BEGIN
 BEGIN

-  newpoints := 0;
-  IF interpolationtype = 'odd' OR interpolationtype = 'even' OR interpolationtype = 'all' THEN
+  IF interpolationtype = 'odd' OR interpolationtype = 'even' THEN
+    stepsize := 2;
+  ELSEIF interpolationtype = 'all' THEN
+    stepsize := 1;
+  ELSEIF interpolationtype ~ '^\d+$' THEN
+    stepsize := interpolationtype::INTEGER;
+  ELSE
+    RETURN 0;
+  END IF;

 
 

-    select postcode from placex where osm_type = 'W' and osm_id = wayid INTO defpostalcode;
-    select nodes from planet_osm_ways where id = wayid INTO waynodes;
---RAISE WARNING 'interpolation % % %',wayid,interpolationtype,waynodes;
-    IF array_upper(waynodes, 1) IS NOT NULL THEN
-
-      havefirstpoint := false;
-
-      FOR nodeidpos in 1..array_upper(waynodes, 1) LOOP
-
-        -- If there is a place of a type other than place/house, use that because
-        -- it is guaranteed to be the original node. For place/house types use the
-        -- one with the smallest id because the original node was created first.
-        -- Ignore all nodes marked for deletion. (Might happen when the type changes.)
-        select place_id from placex where osm_type = 'N' and osm_id = waynodes[nodeidpos]::BIGINT and indexed_status < 100 order by (type = 'house'),place_id limit 1 INTO search_place_id;
-        IF search_place_id IS NULL THEN
-          -- if no such node exists, create a record of the right type
-          select * from placex where osm_type = 'N' and osm_id = waynodes[nodeidpos]::BIGINT and type = 'house' limit 1 INTO nextnode;
-          select ST_SetSRID(ST_Point(lon::float/10000000,lat::float/10000000),4326) from planet_osm_nodes where id = waynodes[nodeidpos] INTO nextnode.geometry;
-          IF nextnode.geometry IS NULL THEN
-            -- we don't have any information about this point, most likely
-            -- because an excerpt was updated and the node never imported
-            -- because the interpolation is outside the region of the excerpt.
-            -- Give up.
-            RETURN newpoints;
-          END IF;
-        ELSE
-          select * from placex where place_id = search_place_id INTO nextnode;
-        END IF;
+  select nodes from planet_osm_ways where id = wayid INTO waynodes;

 
 

---RAISE WARNING 'interpolation node % % % ',nextnode.housenumber,ST_X(nextnode.geometry),ST_Y(nextnode.geometry);
-      
-        IF havefirstpoint THEN
+  IF array_upper(waynodes, 1) IS NULL THEN
+    RETURN 0;
+  END IF;

 
 

-          -- add point to the line string
-          linestr := linestr||','||ST_X(nextnode.geometry)||' '||ST_Y(nextnode.geometry);
-          endnumber := ('0'||substring(nextnode.housenumber,'[0-9]+'))::integer;
+  linegeo := geom;
+  startnumber := NULL;
+  newpoints := 0;

 
 

-          IF startnumber IS NOT NULL and startnumber > 0 AND endnumber IS NOT NULL and endnumber > 0 AND @(startnumber - endnumber) < 1000 THEN
+  FOR nodeidpos in 1..array_upper(waynodes, 1) LOOP
+
+    -- If there is a place of a type other than place/house, use that because
+    -- it is guaranteed to be the original node. For place/house types use the
+    -- one with the smallest id because the original node was created first.
+    -- Ignore all nodes marked for deletion. (Might happen when the type changes.)
+    select * from placex where osm_type = 'N' and osm_id = waynodes[nodeidpos]::BIGINT
+                               and indexed_status < 100
+                         order by (type = 'house'),place_id limit 1 INTO nextnode;
+    IF nextnode.place_id IS NOT NULL THEN
+
+        IF nodeidpos > 1 and nodeidpos < array_upper(waynodes, 1) THEN
+          -- Make sure that the point is actually on the line. That might
+          -- be a bit paranoid but ensures that the algorithm still works
+          -- should osm2pgsql attempt to repair geometries.
+          splitline := split_line_on_node(linegeo, nextnode.geometry);
+          sectiongeo := ST_GeometryN(splitline, 1);
+          linegeo := ST_GeometryN(splitline, 2);
+        ELSE
+          sectiongeo = linegeo;
+        END IF;
+        endnumber := substring(nextnode.housenumber,'[0-9]+')::integer;

 
 

---RAISE WARNING 'interpolation end % % ',nextnode.place_id,endnumber;
+        IF startnumber IS NOT NULL AND endnumber IS NOT NULL
+           AND @(startnumber - endnumber) < 1000 AND startnumber != endnumber
+           AND ST_GeometryType(sectiongeo) = 'ST_LineString' THEN

 
 

-            IF startnumber != endnumber THEN
+          IF (startnumber > endnumber) THEN
+            housenum := endnumber;
+            endnumber := startnumber;
+            startnumber := housenum;
+            sectiongeo := ST_Reverse(sectiongeo);
+          END IF;
+          orginalstartnumber := startnumber;
+          originalnumberrange := endnumber - startnumber;

 
 

-              linestr := linestr || ')';
---RAISE WARNING 'linestr %',linestr;
-              linegeo := ST_GeomFromText(linestr,4326);
-              linestr := 'LINESTRING('||ST_X(nextnode.geometry)||' '||ST_Y(nextnode.geometry);
-              IF (startnumber > endnumber) THEN
-                housenum := endnumber;
-                endnumber := startnumber;
-                startnumber := housenum;
-                linegeo := ST_Reverse(linegeo);
-              END IF;
-              orginalstartnumber := startnumber;
-              originalnumberrange := endnumber - startnumber;
-
--- Too much broken data worldwide for this test to be worth using
---              IF originalnumberrange > 500 THEN
---                RAISE WARNING 'Number block of % while processing % %', originalnumberrange, prevnode, nextnode;
---              END IF;
-
-              IF (interpolationtype = 'odd' AND startnumber%2 = 0) OR (interpolationtype = 'even' AND startnumber%2 = 1) THEN
-                startnumber := startnumber + 1;
-                stepsize := 2;
-              ELSE
-                IF (interpolationtype = 'odd' OR interpolationtype = 'even') THEN
-                  startnumber := startnumber + 2;
-                  stepsize := 2;
-                ELSE -- everything else assumed to be 'all'
-                  startnumber := startnumber + 1;
-                  stepsize := 1;
-                END IF;
-              END IF;
-              endnumber := endnumber - 1;
-              delete from placex where osm_type = 'N' and osm_id = prevnode.osm_id and type = 'house' and place_id != prevnode.place_id;
-              FOR housenum IN startnumber..endnumber BY stepsize LOOP
-                -- this should really copy postcodes but it puts a huge burdon on the system for no big benefit
-                -- ideally postcodes should move up to the way
-                insert into placex (osm_type, osm_id, class, type, admin_level, housenumber, street, addr_place, isin, postcode,
-                  country_code, parent_place_id, rank_address, rank_search, indexed_status, geometry)
-                  values ('N',prevnode.osm_id, 'place', 'house', prevnode.admin_level, housenum, prevnode.street, prevnode.addr_place, prevnode.isin, coalesce(prevnode.postcode, defpostalcode),
-                  prevnode.country_code, prevnode.parent_place_id, prevnode.rank_address, prevnode.rank_search, 1, ST_LineInterpolatePoint(linegeo, (housenum::float-orginalstartnumber::float)/originalnumberrange::float));
-                newpoints := newpoints + 1;
---RAISE WARNING 'interpolation number % % ',prevnode.place_id,housenum;
-              END LOOP;
-            END IF;
-            havefirstpoint := false;
+          startnumber := startnumber + stepsize;
+          -- correct for odd/even
+          IF (interpolationtype = 'odd' AND startnumber%2 = 0) OR (interpolationtype = 'even' AND startnumber%2 = 1) THEN
+            startnumber := startnumber - 1;

           END IF;
           END IF;

+          endnumber := endnumber - 1;
+
+          delete from placex where osm_type = 'N' and osm_id = prevnode.osm_id and type = 'house' and place_id != prevnode.place_id;
+          FOR housenum IN startnumber..endnumber BY stepsize LOOP
+            -- this should really copy postcodes but it puts a huge burden on
+            -- the system for no big benefit ideally postcodes should move up to the way
+            insert into placex (place_id, partition, osm_type, osm_id, class, type, admin_level,
+              housenumber, street, addr_place, isin, postcode,
+              country_code, parent_place_id, rank_address, rank_search,
+              indexed_status, geometry)
+              values (nextval('seq_place'), partition, 'N', prevnode.osm_id, 'place', 'house', prevnode.admin_level,
+              housenum, prevnode.street, prevnode.addr_place, prevnode.isin, coalesce(prevnode.postcode, defpostalcode),
+              prevnode.country_code, prevnode.parent_place_id, prevnode.rank_address, prevnode.rank_search,
+              0, ST_LineInterpolatePoint(sectiongeo, (housenum::float-orginalstartnumber::float)/originalnumberrange::float));
+            newpoints := newpoints + 1;
+--RAISE WARNING 'interpolation number % % ',prevnode.place_id,housenum;
+          END LOOP;

         END IF;
 
         END IF;
 

-        IF NOT havefirstpoint THEN
-          startnumber := ('0'||substring(nextnode.housenumber,'[0-9]+'))::integer;
-          IF startnumber IS NOT NULL AND startnumber > 0 THEN
-            havefirstpoint := true;
-            linestr := 'LINESTRING('||ST_X(nextnode.geometry)||' '||ST_Y(nextnode.geometry);
-            prevnode := nextnode;
-          END IF;
---RAISE WARNING 'interpolation start % % ',nextnode.place_id,startnumber;
+        -- early break if we are out of line string,
+        -- might happen when a line string loops back on itself
+        IF ST_GeometryType(linegeo) != 'ST_LineString' THEN
+            RETURN newpoints;

         END IF;
         END IF;

-      END LOOP;
+
+        startnumber := substring(nextnode.housenumber,'[0-9]+')::integer;
+        prevnode := nextnode;

     END IF;
     END IF;

-  END IF;
+  END LOOP;

 
 --RAISE WARNING 'interpolation points % ',newpoints;
 
 
 --RAISE WARNING 'interpolation points % ',newpoints;
 


@@ -936,6 +944,10 @@ DECLARE
 BEGIN
   --DEBUG: RAISE WARNING '% %',NEW.osm_type,NEW.osm_id;
 
 BEGIN
   --DEBUG: RAISE WARNING '% %',NEW.osm_type,NEW.osm_id;
 

+  IF place_id is not null THEN
+    RETURN NEW; -- already indexed (happens for interpolated addresses
+  END IF;
+

   -- just block these
   IF NEW.class in ('landuse','natural') and NEW.name is null THEN
 --    RAISE WARNING 'empty landuse %',NEW.osm_id;
   -- just block these
   IF NEW.class in ('landuse','natural') and NEW.name is null THEN
 --    RAISE WARNING 'empty landuse %',NEW.osm_id;


@@ -1018,6 +1030,13 @@ BEGIN
             NEW.rank_address := 5;
           END IF;
 
             NEW.rank_address := 5;
           END IF;
 

+        ELSEIF NEW.calculated_country_code = 'sg' THEN
+
+          IF NEW.postcode ~ '^([0-9]{6})$' THEN
+            NEW.rank_search := 25;
+            NEW.rank_address := 11;
+          END IF;
+

         ELSEIF NEW.calculated_country_code = 'de' THEN
 
           IF NEW.postcode ~ '^([0-9]{5})$' THEN
         ELSEIF NEW.calculated_country_code = 'de' THEN
 
           IF NEW.postcode ~ '^([0-9]{5})$' THEN


@@ -1095,8 +1114,6 @@ BEGIN
         NEW.rank_address := NEW.rank_search;
       ELSEIF NEW.type in ('houses') THEN
         -- can't guarantee all required nodes loaded yet due to caching in osm2pgsql
         NEW.rank_address := NEW.rank_search;
       ELSEIF NEW.type in ('houses') THEN
         -- can't guarantee all required nodes loaded yet due to caching in osm2pgsql

-        -- insert new point into place for each derived building
-        --i := create_interpolation(NEW.osm_id, NEW.housenumber);

         NEW.rank_search := 28;
         NEW.rank_address := 0;
       END IF;
         NEW.rank_search := 28;
         NEW.rank_address := 0;
       END IF;


@@ -1114,7 +1131,11 @@ BEGIN
       END IF;
     ELSEIF NEW.class = 'landuse' AND ST_GeometryType(NEW.geometry) in ('ST_Polygon','ST_MultiPolygon') THEN
       NEW.rank_search := 22;
       END IF;
     ELSEIF NEW.class = 'landuse' AND ST_GeometryType(NEW.geometry) in ('ST_Polygon','ST_MultiPolygon') THEN
       NEW.rank_search := 22;

-      NEW.rank_address := NEW.rank_search;
+      IF NEW.type in ('residential', 'farm', 'farmyard', 'industrial', 'commercial', 'allotments', 'retail') THEN
+        NEW.rank_address := NEW.rank_search;
+      ELSE
+        NEW.rank_address := 0;
+      END IF;

     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 in ('peak','volcano','mountain_range') THEN
       NEW.rank_search := 18;
       NEW.rank_address := 0;


@@ -1126,7 +1147,7 @@ BEGIN
     -- 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;
     -- 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;

-      NEW.rank_address := NEW.rank_search;
+      NEW.rank_address := 0;

     ELSEIF NEW.class = 'highway' AND NEW.name is NULL AND 
            NEW.type in ('service','cycleway','path','footway','steps','bridleway','track','byway','motorway_link','primary_link','trunk_link','secondary_link','tertiary_link') THEN
 --      RAISE WARNING 'unnamed minor feature %',NEW.osm_id;
     ELSEIF NEW.class = 'highway' AND NEW.name is NULL AND 
            NEW.type in ('service','cycleway','path','footway','steps','bridleway','track','byway','motorway_link','primary_link','trunk_link','secondary_link','tertiary_link') THEN
 --      RAISE WARNING 'unnamed minor feature %',NEW.osm_id;


@@ -1136,7 +1157,11 @@ BEGIN
     ELSEIF NEW.class = 'waterway' AND NEW.name is NULL THEN
       RETURN NULL;
     ELSEIF NEW.class = 'waterway' THEN
     ELSEIF NEW.class = 'waterway' AND NEW.name is NULL THEN
       RETURN NULL;
     ELSEIF NEW.class = 'waterway' THEN

-      NEW.rank_search := 17;
+      IF NEW.osm_type = 'R' THEN
+        NEW.rank_search := 16;
+      ELSE
+        NEW.rank_search := 17;
+      END IF;

       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 := 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;


@@ -1258,8 +1283,6 @@ DECLARE
   search_maxdistance FLOAT[];
   search_mindistance FLOAT[];
   address_havelevel BOOLEAN[];
   search_maxdistance FLOAT[];
   search_mindistance FLOAT[];
   address_havelevel BOOLEAN[];

---  search_scores wordscore[];
---  search_scores_pos INTEGER;

 
   i INTEGER;
   iMax FLOAT;
 
   i INTEGER;
   iMax FLOAT;


@@ -1284,6 +1307,7 @@ DECLARE
   location_distance FLOAT;
   location_parent GEOMETRY;
   location_isaddress BOOLEAN;
   location_distance FLOAT;
   location_parent GEOMETRY;
   location_isaddress BOOLEAN;

+  location_keywords INTEGER[];

 
   tagpairid INTEGER;
 
 
   tagpairid INTEGER;
 


@@ -1324,7 +1348,6 @@ BEGIN
 
     result := deleteSearchName(NEW.partition, NEW.place_id);
     DELETE FROM place_addressline WHERE place_id = NEW.place_id;
 
     result := deleteSearchName(NEW.partition, NEW.place_id);
     DELETE FROM place_addressline WHERE place_id = NEW.place_id;

-    DELETE FROM place_boundingbox where place_id = NEW.place_id;

     result := deleteRoad(NEW.partition, NEW.place_id);
     result := deleteLocationArea(NEW.partition, NEW.place_id, NEW.rank_search);
     UPDATE placex set linked_place_id = null where linked_place_id = NEW.place_id;
     result := deleteRoad(NEW.partition, NEW.place_id);
     result := deleteLocationArea(NEW.partition, NEW.place_id, NEW.rank_search);
     UPDATE placex set linked_place_id = null where linked_place_id = NEW.place_id;


@@ -1333,17 +1356,12 @@ BEGIN
       RETURN NEW;
     END IF;
 
       RETURN NEW;
     END IF;
 

-    IF NEW.class = 'place' AND NEW.type = 'houses' THEN
-      i := create_interpolation(NEW.osm_id, NEW.housenumber);
-      RETURN NEW;
-    END IF;
-

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

-    -- reclaculate country and partition
+    -- recalculate country and partition

     IF NEW.rank_search = 4 THEN
       -- for countries, believe the mapped country code,
       -- so that we remain in the right partition if the boundaries
     IF NEW.rank_search = 4 THEN
       -- for countries, believe the mapped country code,
       -- so that we remain in the right partition if the boundaries


@@ -1366,6 +1384,37 @@ BEGIN
     END IF;
     NEW.geometry_sector := geometry_sector(NEW.partition, place_centroid);
 
     END IF;
     NEW.geometry_sector := geometry_sector(NEW.partition, place_centroid);
 

+    -- interpolations XXXXX
+    IF NEW.class = 'place' AND NEW.type = 'houses'THEN
+      IF osm_type = 'W' and ST_GeometryType(NEW.geometry) = 'ST_LineString' THEN
+        NEW.parent_place_id := get_interpolation_parent(NEW.osm_id, NEW.street, NEW.addr_place,
+                                                        NEW.partition, place_centroid, NEW.geometry);
+        IF NEW.parent_place_id is not null THEN
+          i := create_interpolation(NEW.osm_id, NEW.housenumber);
+        END IF;
+      END IF;
+      RETURN NEW;
+    END IF;
+
+    -- waterway ways are linked when they are part of a relation and have the same class/type
+    IF NEW.osm_type = 'R' and NEW.class = 'waterway' THEN
+        FOR relation_members IN select members from planet_osm_rels r where r.id = NEW.osm_id and r.parts != array[]::bigint[]
+        LOOP
+            FOR i IN 1..array_upper(relation_members, 1) BY 2 LOOP
+                IF relation_members[i+1] in ('', 'main_stream', 'side_stream') AND substring(relation_members[i],1,1) = 'w' THEN
+                  --DEBUG: RAISE WARNING 'waterway parent %, child %/%', NEW.osm_id, i, relation.members[i];
+                  FOR linked_node_id IN SELECT place_id FROM placex
+                    WHERE osm_type = 'W' and osm_id = substring(relation_members[i],2,200)::bigint
+                    and class = NEW.class and type = NEW.type
+                    and ( relation_members[i+1] != 'side_stream' or NEW.name->'name' = name->'name')
+                  LOOP
+                    UPDATE placex SET linked_place_id = NEW.place_id WHERE place_id = linked_node_id;
+                  END LOOP;
+                END IF;
+            END LOOP;
+        END LOOP;
+    END IF;
+

     -- Adding ourselves to the list simplifies address calculations later
     INSERT INTO place_addressline VALUES (NEW.place_id, NEW.place_id, true, true, 0, NEW.rank_address); 
 
     -- Adding ourselves to the list simplifies address calculations later
     INSERT INTO place_addressline VALUES (NEW.place_id, NEW.place_id, true, true, 0, NEW.rank_address); 
 


@@ -1478,13 +1527,10 @@ BEGIN
           END IF;    
           
           -- If the way contains an explicit name of a street copy it
           END IF;    
           
           -- If the way contains an explicit name of a street copy it

-          IF NEW.street IS NULL AND NEW.addr_place IS NULL AND location.street IS NOT NULL THEN
+          -- Slightly less strict then above because data is copied from any object.
+          IF NEW.street IS NULL AND NEW.addr_place IS NULL THEN

 --RAISE WARNING 'node in way that has a streetname %',location;
             NEW.street := location.street;
 --RAISE WARNING 'node in way that has a streetname %',location;
             NEW.street := location.street;

-          END IF;
-
-          -- IF the way contains an explicit name of a place copy it
-          IF NEW.addr_place IS NULL AND NEW.street IS NULL AND location.addr_place IS NOT NULL THEN

             NEW.addr_place := location.addr_place;
           END IF;
 
             NEW.addr_place := location.addr_place;
           END IF;
 


@@ -1815,10 +1861,26 @@ BEGIN
 
       IF location.rank_address != location_rank_search THEN
         location_rank_search := location.rank_address;
 
       IF location.rank_address != location_rank_search THEN
         location_rank_search := location.rank_address;

-        location_distance := location.distance * 1.5;
+        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;
+        END IF;
+      ELSE
+        CONTINUE WHEN location.keywords <@ location_keywords;

       END IF;
 
       IF location.distance < location_distance OR NOT location.isguess THEN
       END IF;
 
       IF location.distance < location_distance OR NOT location.isguess THEN

+        location_keywords := location.keywords;

 
         location_isaddress := NOT address_havelevel[location.rank_address];
         IF location_isaddress AND location.isguess AND location_parent IS NOT NULL THEN
 
         location_isaddress := NOT address_havelevel[location.rank_address];
         IF location_isaddress AND location.isguess AND location_parent IS NOT NULL THEN


@@ -2070,8 +2132,8 @@ BEGIN
   -- Handle a place changing type by removing the old data
   -- My generated 'place' types are causing havok because they overlap with real keys
   -- TODO: move them to their own special purpose key/class to avoid collisions
   -- Handle a place changing type by removing the old data
   -- My generated 'place' types are causing havok because they overlap with real keys
   -- TODO: move them to their own special purpose key/class to avoid collisions

-  IF existing.osm_type IS NULL AND (NEW.type not in ('postcode','house','houses')) THEN
-    DELETE FROM place where osm_type = NEW.osm_type and osm_id = NEW.osm_id and class = NEW.class and type not in ('postcode','house','houses');
+  IF existing.osm_type IS NULL THEN
+    DELETE FROM place where osm_type = NEW.osm_type and osm_id = NEW.osm_id and class = NEW.class;

   END IF;
 
   --DEBUG: RAISE WARNING 'Existing: %',existing.osm_id;
   END IF;
 
   --DEBUG: RAISE WARNING 'Existing: %',existing.osm_id;


@@ -2316,29 +2378,6 @@ END;
 $$
 LANGUAGE plpgsql IMMUTABLE;
 
 $$
 LANGUAGE plpgsql IMMUTABLE;
 

-CREATE OR REPLACE FUNCTION get_connected_ways(way_ids INTEGER[]) RETURNS SETOF planet_osm_ways
-  AS $$
-DECLARE
-  searchnodes INTEGER[];
-  location RECORD;
-  j INTEGER;
-BEGIN
-
-  searchnodes := '{}';
-  FOR j IN 1..array_upper(way_ids, 1) LOOP
-    FOR location IN 
-      select nodes from planet_osm_ways where id = way_ids[j] LIMIT 1
-    LOOP
-      IF not (ARRAY[location.nodes] <@ searchnodes) THEN
-        searchnodes := searchnodes || location.nodes;
-      END IF;
-    END LOOP;
-  END LOOP;
-
-  RETURN QUERY select * from planet_osm_ways where nodes && searchnodes and NOT ARRAY[id] <@ way_ids;
-END;
-$$
-LANGUAGE plpgsql IMMUTABLE;

 
 CREATE OR REPLACE FUNCTION get_address_postcode(for_place_id BIGINT) RETURNS TEXT
   AS $$
 
 CREATE OR REPLACE FUNCTION get_address_postcode(for_place_id BIGINT) RETURNS TEXT
   AS $$


@@ -2470,7 +2509,7 @@ BEGIN
       CASE WHEN class = 'place' and type = 'postcode' THEN hstore('name', postcode) ELSE name END as name,
       class, type, admin_level, true as fromarea, true as isaddress,
       CASE WHEN rank_address = 0 THEN 100 WHEN rank_address = 11 THEN 5 ELSE rank_address END as rank_address,
       CASE WHEN class = 'place' and type = 'postcode' THEN hstore('name', postcode) ELSE name END as name,
       class, type, admin_level, true as fromarea, true as isaddress,
       CASE WHEN rank_address = 0 THEN 100 WHEN rank_address = 11 THEN 5 ELSE rank_address END as rank_address,

-      0 as distance, calculated_country_code
+      0 as distance, calculated_country_code, postcode

       from placex
       where place_id = for_place_id 
   LOOP
       from placex
       where place_id = for_place_id 
   LOOP


@@ -2481,6 +2520,9 @@ BEGIN
     IF searchpostcode IS NOT NULL and location.type = 'postcode' THEN
       location.isaddress := FALSE;
     END IF;
     IF searchpostcode IS NOT NULL and location.type = 'postcode' THEN
       location.isaddress := FALSE;
     END IF;

+    IF searchpostcode IS NULL and location.postcode IS NOT NULL THEN
+      searchpostcode := location.postcode;
+    END IF;

     IF location.rank_address = 4 AND location.isaddress THEN
       hadcountry := true;
     END IF;
     IF location.rank_address = 4 AND location.isaddress THEN
       hadcountry := true;
     END IF;


@@ -2575,86 +2617,10 @@ END;
 $$
 LANGUAGE plpgsql;
 
 $$
 LANGUAGE plpgsql;
 

-CREATE OR REPLACE FUNCTION get_place_boundingbox(search_place_id BIGINT) RETURNS place_boundingbox
-  AS $$
-DECLARE
-  result place_boundingbox;
-  numfeatures integer;
-BEGIN
-  select * from place_boundingbox into result where place_id = search_place_id;
-  IF result.place_id IS NULL THEN
--- remove  isaddress = true because if there is a matching polygon it always wins
-    select count(*) from place_addressline where address_place_id = search_place_id into numfeatures;
-    insert into place_boundingbox select place_id,
-             ST_Y(ST_PointN(ExteriorRing(ST_Box2D(geometry)),4)),ST_Y(ST_PointN(ExteriorRing(ST_Box2D(geometry)),2)),
-             ST_X(ST_PointN(ExteriorRing(ST_Box2D(geometry)),1)),ST_X(ST_PointN(ExteriorRing(ST_Box2D(geometry)),3)),
-             numfeatures, ST_Area(geometry),
-             geometry as area from location_area where place_id = search_place_id;
-    select * from place_boundingbox into result where place_id = search_place_id;
-  END IF;
-  IF result.place_id IS NULL THEN
--- TODO 0.0001
-    insert into place_boundingbox select address_place_id,
-             min(ST_Y(ST_Centroid(geometry))) as minlon,max(ST_Y(ST_Centroid(geometry))) as maxlon,
-             min(ST_X(ST_Centroid(geometry))) as minlat,max(ST_X(ST_Centroid(geometry))) as maxlat,
-             count(*), ST_Area(ST_Buffer(ST_Convexhull(ST_Collect(geometry)),0.0001)) as area,
-             ST_Buffer(ST_Convexhull(ST_Collect(geometry)),0.0001) as boundary 
-             from (select * from place_addressline where address_place_id = search_place_id order by cached_rank_address limit 4000) as place_addressline join placex using (place_id) 
-             where address_place_id = search_place_id
---               and (isaddress = true OR place_id = search_place_id)
-               and (st_length(geometry) < 0.01 or place_id = search_place_id)
-             group by address_place_id limit 1;
-    select * from place_boundingbox into result where place_id = search_place_id;
-  END IF;
-  return result;
-END;
-$$
-LANGUAGE plpgsql;
-
--- don't do the operation if it would be slow
-CREATE OR REPLACE FUNCTION get_place_boundingbox_quick(search_place_id BIGINT) RETURNS place_boundingbox
-  AS $$
-DECLARE
-  result place_boundingbox;
-  numfeatures integer;
-  rank integer;
-BEGIN
-  select * from place_boundingbox into result where place_id = search_place_id;
-  IF result IS NULL AND rank > 14 THEN
-    select count(*) from place_addressline where address_place_id = search_place_id and isaddress = true into numfeatures;
-    insert into place_boundingbox select place_id,
-             ST_Y(ST_PointN(ExteriorRing(ST_Box2D(geometry)),4)),ST_Y(ST_PointN(ExteriorRing(ST_Box2D(geometry)),2)),
-             ST_X(ST_PointN(ExteriorRing(ST_Box2D(geometry)),1)),ST_X(ST_PointN(ExteriorRing(ST_Box2D(geometry)),3)),
-             numfeatures, ST_Area(geometry),
-             geometry as area from location_area where place_id = search_place_id;
-    select * from place_boundingbox into result where place_id = search_place_id;
-  END IF;
-  IF result IS NULL THEN
-    select rank_search from placex where place_id = search_place_id into rank;
-    IF rank > 20 THEN
--- TODO 0.0001
-      insert into place_boundingbox select address_place_id,
-             min(ST_Y(ST_Centroid(geometry))) as minlon,max(ST_Y(ST_Centroid(geometry))) as maxlon,
-             min(ST_X(ST_Centroid(geometry))) as minlat,max(ST_X(ST_Centroid(geometry))) as maxlat,
-             count(*), ST_Area(ST_Buffer(ST_Convexhull(ST_Collect(geometry)),0.0001)) as area,
-             ST_Buffer(ST_Convexhull(ST_Collect(geometry)),0.0001) as boundary 
-             from place_addressline join placex using (place_id) 
-             where address_place_id = search_place_id 
-               and (isaddress = true OR place_id = search_place_id)
-               and (st_length(geometry) < 0.01 or place_id = search_place_id)
-             group by address_place_id limit 1;
-      select * from place_boundingbox into result where place_id = search_place_id;
-    END IF;
-  END IF;
-  return result;
-END;
-$$
-LANGUAGE plpgsql;

 
 CREATE OR REPLACE FUNCTION update_place(search_place_id BIGINT) RETURNS BOOLEAN
   AS $$
 DECLARE
 
 CREATE OR REPLACE FUNCTION update_place(search_place_id BIGINT) RETURNS BOOLEAN
   AS $$
 DECLARE

-  result place_boundingbox;

   numfeatures integer;
 BEGIN
   update placex set 
   numfeatures integer;
 BEGIN
   update placex set 


@@ -2934,14 +2900,14 @@ BEGIN
       wiki_article := regexp_replace(wiki_article,E'^(.*?)([a-z]{2,3}).wikipedia.org/wiki/',E'\\2:');
       wiki_article := regexp_replace(wiki_article,E'^(.*?)([a-z]{2,3}).wikipedia.org/w/index.php\\?title=',E'\\2:');
       wiki_article := regexp_replace(wiki_article,E'^(.*?)/([a-z]{2,3})/wiki/',E'\\2:');
       wiki_article := regexp_replace(wiki_article,E'^(.*?)([a-z]{2,3}).wikipedia.org/wiki/',E'\\2:');
       wiki_article := regexp_replace(wiki_article,E'^(.*?)([a-z]{2,3}).wikipedia.org/w/index.php\\?title=',E'\\2:');
       wiki_article := regexp_replace(wiki_article,E'^(.*?)/([a-z]{2,3})/wiki/',E'\\2:');

-      wiki_article := regexp_replace(wiki_article,E'^(.*?)([a-z]{2,3})[=:]',E'\\2:');
+      --wiki_article := regexp_replace(wiki_article,E'^(.*?)([a-z]{2,3})[=:]',E'\\2:');

       wiki_article := replace(wiki_article,' ','_');
       wiki_article := replace(wiki_article,' ','_');

-      wiki_article_title := trim(split_part(wiki_article, ':', 2));
-      IF wiki_article_title IS NULL OR wiki_article_title = '' THEN
+      IF strpos(wiki_article, ':') IN (3,4) THEN
+        wiki_article_language := lower(trim(split_part(wiki_article, ':', 1)));
+        wiki_article_title := trim(substr(wiki_article, strpos(wiki_article, ':')+1));
+      ELSE

         wiki_article_title := trim(wiki_article);
         wiki_article_language := CASE WHEN langs[i] = 'english' THEN 'en' WHEN langs[i] = 'country' THEN get_country_language_code(country_code) ELSE langs[i] END;
         wiki_article_title := trim(wiki_article);
         wiki_article_language := CASE WHEN langs[i] = 'english' THEN 'en' WHEN langs[i] = 'country' THEN get_country_language_code(country_code) ELSE langs[i] END;

-      ELSE
-        wiki_article_language := lower(trim(split_part(wiki_article, ':', 1)));

       END IF;
 
       select wikipedia_article.language,wikipedia_article.title,wikipedia_article.importance
       END IF;
 
       select wikipedia_article.language,wikipedia_article.title,wikipedia_article.importance