Merge remote-tracking branch 'upstream/master'

[nominatim.git] / sql / functions.sql

diff --git a/sql/functions.sql b/sql/functions.sql
index a6fa20ea295be751f907848f663149d4b1e2b3ca..4aef74d1d28089f63a96fcc0dcaac6bc607d4351 100644 (file)
--- a/sql/functions.sql
+++ b/sql/functions.sql
@@ -791,7 +791,7 @@ LANGUAGE plpgsql;
 CREATE OR REPLACE FUNCTION create_interpolation(wayid BIGINT, interpolationtype TEXT) RETURNS INTEGER
   AS $$
 DECLARE
 CREATE OR REPLACE FUNCTION create_interpolation(wayid BIGINT, interpolationtype TEXT) RETURNS INTEGER
   AS $$
 DECLARE

-  
+

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


@@ -804,117 +804,103 @@ DECLARE
   originalnumberrange INTEGER;
   housenum INTEGER;
   linegeo GEOMETRY;
   originalnumberrange INTEGER;
   housenum INTEGER;
   linegeo GEOMETRY;

+  splitline GEOMETRY;
+  sectiongeo GEOMETRY;

   search_place_id BIGINT;
   defpostalcode TEXT;
 
   search_place_id BIGINT;
   defpostalcode TEXT;
 

-  havefirstpoint BOOLEAN;
-  linestr TEXT;

 BEGIN
 BEGIN

-  newpoints := 0;
-  IF interpolationtype = 'odd' OR interpolationtype = 'even' OR interpolationtype = 'all' THEN
-
-    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;
+  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;

 
 

---RAISE WARNING 'interpolation node % % % ',nextnode.housenumber,ST_X(nextnode.geometry),ST_Y(nextnode.geometry);
-      
-        IF havefirstpoint THEN
+  select nodes from planet_osm_ways where id = wayid INTO waynodes;

 
 

-          -- add point to the line string
-          linestr := linestr||','||ST_X(nextnode.geometry)||' '||ST_Y(nextnode.geometry);
-          endnumber := ('0'||substring(nextnode.housenumber,'[0-9]+'))::integer;
+  IF array_upper(waynodes, 1) IS NULL THEN
+    RETURN 0;
+  END IF;

 
 

-          IF startnumber IS NOT NULL and startnumber > 0 AND endnumber IS NOT NULL and endnumber > 0 AND @(startnumber - endnumber) < 1000 THEN
+  select postcode, geometry from placex where osm_type = 'W' and osm_id = wayid 
+    INTO defpostalcode, linegeo;

 
 

---RAISE WARNING 'interpolation end % % ',nextnode.place_id,endnumber;
+  IF ST_GeometryType(linegeo) != 'ST_LineString' THEN
+    RETURN 0;
+  END IF;

 
 

-            IF startnumber != endnumber THEN
+  startnumber := NULL;
+  newpoints := 0;

 
 

-              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;
+  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 NOT NULL THEN
+      select * from placex where place_id = search_place_id INTO nextnode;
+
+        IF 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);
+          IF ST_GeometryType(ST_GeometryN(splitline, 2)) = 'ST_LineString' THEN
+            linegeo := ST_GeometryN(splitline, 2);

           END IF;
           END IF;

+        ELSE
+          sectiongeo = linegeo;

         END IF;
         END IF;

+        endnumber := substring(nextnode.housenumber,'[0-9]+')::integer;

 
 

-        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;
+        IF startnumber IS NOT NULL AND endnumber IS NOT NULL
+           AND @(startnumber - endnumber) < 1000 AND startnumber != endnumber
+           AND ST_GeometryType(linegeo) = 'ST_LineString' THEN
+
+          IF (startnumber > endnumber) THEN
+            housenum := endnumber;
+            endnumber := startnumber;
+            startnumber := housenum;
+            sectiongeo := ST_Reverse(sectiongeo);

           END IF;
           END IF;

---RAISE WARNING 'interpolation start % % ',nextnode.place_id,startnumber;
+          orginalstartnumber := startnumber;
+          originalnumberrange := endnumber - startnumber;
+
+          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;
+          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 (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(sectiongeo, (housenum::float-orginalstartnumber::float)/originalnumberrange::float));
+            newpoints := newpoints + 1;
+--RAISE WARNING 'interpolation number % % ',prevnode.place_id,housenum;
+          END LOOP;

         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;
 


@@ -1023,6 +1009,13 @@ 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


@@ -1119,7 +1112,11 @@ 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;


@@ -1377,12 +1374,15 @@ BEGIN
 
     -- 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
 
     -- 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 IN select * from planet_osm_rels r where r.id = NEW.osm_id
+        FOR relation IN select * from planet_osm_rels r where r.id = NEW.osm_id and r.parts != array[]::bigint[]

         LOOP
         LOOP

-            FOR i IN relation.way_off+1..relation.rel_off LOOP
-                IF relation.members[2*i] in ('', 'main_stream') THEN
+            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.parts[i];
                   --DEBUG: RAISE WARNING 'waterway parent %, child %/%', NEW.osm_id, i, relation.parts[i];

-                  FOR location IN SELECT * FROM placex WHERE osm_type = 'W' and osm_id = relation.parts[i] and class = NEW.class and type = NEW.type
+                  FOR location IN SELECT * 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 = location.place_id;
                   END LOOP;
                   LOOP
                     UPDATE placex SET linked_place_id = NEW.place_id WHERE place_id = location.place_id;
                   END LOOP;


@@ -2097,8 +2097,8 @@ 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;


@@ -2349,29 +2349,29 @@ 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_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 $$


@@ -2503,7 +2503,7 @@ 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


@@ -2514,6 +2514,9 @@ 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;


@@ -2967,14 +2970,14 @@ 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