fix rank inheritance from linked places

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

diff --git a/lib-sql/functions/placex_triggers.sql b/lib-sql/functions/placex_triggers.sql
index 6ab73a3b4143dea539d7304ce8cc464d2109e7a5..37b57b6467ac01f17bf4014b161cec5490cd4831 100644 (file)
--- a/lib-sql/functions/placex_triggers.sql
+++ b/lib-sql/functions/placex_triggers.sql
@@ -26,40 +26,47 @@ CREATE OR REPLACE FUNCTION placex_indexing_prepare(p placex)
 DECLARE
   location RECORD;
   result prepare_update_info;
 DECLARE
   location RECORD;
   result prepare_update_info;

+  extra_names HSTORE;

 BEGIN
 BEGIN

+  IF not p.address ? '_inherited' THEN
+    result.address := p.address;
+  END IF;
+

   -- For POI nodes, check if the address should be derived from a surrounding
   -- building.
   -- For POI nodes, check if the address should be derived from a surrounding
   -- building.

-  IF p.rank_search < 30 OR p.osm_type != 'N' THEN
-    result.address := p.address;
-  ELSEIF p.address is null THEN
-    -- The additional && condition works around the misguided query
-    -- planner of postgis 3.0.
-    SELECT placex.address || hstore('_inherited', '') INTO result.address
-      FROM placex
-     WHERE ST_Covers(geometry, p.centroid)
-           and geometry && p.centroid
-           and placex.address is not null
-           and (placex.address ? 'housenumber' or placex.address ? 'street' or placex.address ? 'place')
-           and rank_search = 30 AND ST_GeometryType(geometry) in ('ST_Polygon','ST_MultiPolygon')
-     LIMIT 1;
-  ELSE
-    result.address := p.address;
-    -- See if we can inherit addtional address tags from an interpolation.
-    -- These will become permanent.
-    FOR location IN
-      SELECT (address - 'interpolation'::text - 'housenumber'::text) as address
-        FROM place, planet_osm_ways w
-        WHERE place.osm_type = 'W' and place.address ? 'interpolation'
-              and place.geometry && p.geometry
-              and place.osm_id = w.id
-              and p.osm_id = any(w.nodes)
-    LOOP
-      result.address := location.address || result.address;
-    END LOOP;
+  IF p.rank_search = 30 AND p.osm_type = 'N' THEN
+    IF p.address is null THEN
+        -- The additional && condition works around the misguided query
+        -- planner of postgis 3.0.
+        SELECT placex.address || hstore('_inherited', '') INTO result.address
+          FROM placex
+         WHERE ST_Covers(geometry, p.centroid)
+               and geometry && p.centroid
+               and placex.address is not null
+               and (placex.address ? 'housenumber' or placex.address ? 'street' or placex.address ? 'place')
+               and rank_search = 30 AND ST_GeometryType(geometry) in ('ST_Polygon','ST_MultiPolygon')
+         LIMIT 1;
+    ELSE
+      -- See if we can inherit addtional address tags from an interpolation.
+      -- These will become permanent.
+      FOR location IN
+        SELECT (address - 'interpolation'::text - 'housenumber'::text) as address
+          FROM place, planet_osm_ways w
+          WHERE place.osm_type = 'W' and place.address ? 'interpolation'
+                and place.geometry && p.geometry
+                and place.osm_id = w.id
+                and p.osm_id = any(w.nodes)
+      LOOP
+        result.address := location.address || result.address;
+      END LOOP;
+    END IF;

   END IF;
 
   END IF;
 

+  -- remove internal and derived names

   result.address := result.address - '_unlisted_place'::TEXT;
   result.address := result.address - '_unlisted_place'::TEXT;

-  result.name := p.name;
+  SELECT hstore(array_agg(key), array_agg(value)) INTO result.name
+    FROM each(p.name) WHERE key not like '\_%';
+

   result.class := p.class;
   result.type := p.type;
   result.country_code := p.country_code;
   result.class := p.class;
   result.type := p.type;
   result.country_code := p.country_code;


@@ -72,8 +79,20 @@ BEGIN
   IF location.place_id is not NULL THEN
     result.linked_place_id := location.place_id;
 
   IF location.place_id is not NULL THEN
     result.linked_place_id := location.place_id;
 

-    IF NOT location.name IS NULL THEN
-      result.name := location.name || result.name;
+    IF location.name is not NULL THEN
+      {% if debug %}RAISE WARNING 'Names original: %, location: %', result.name, location.name;{% endif %}
+      -- Add all names from the place nodes that deviate from the name
+      -- in the relation with the prefix '_place_'. Deviation means that
+      -- either the value is different or a given key is missing completely
+      SELECT hstore(array_agg('_place_' || key), array_agg(value)) INTO extra_names
+        FROM each(location.name - result.name);
+      {% if debug %}RAISE WARNING 'Extra names: %', extra_names;{% endif %}
+
+      IF extra_names is not null THEN
+          result.name := result.name || extra_names;
+      END IF;
+
+      {% if debug %}RAISE WARNING 'Final names: %', result.name;{% endif %}

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


@@ -101,7 +120,8 @@ BEGIN
       IF location.members[i+1] = 'street' THEN
         FOR parent IN
           SELECT place_id from placex
       IF location.members[i+1] = 'street' THEN
         FOR parent IN
           SELECT place_id from placex

-           WHERE osm_type = 'W' and osm_id = substring(location.members[i],2)::bigint
+           WHERE osm_type = upper(substring(location.members[i], 1, 1))::char(1)
+                 and osm_id = substring(location.members[i], 2)::bigint

                  and name is not null
                  and rank_search between 26 and 27
         LOOP
                  and name is not null
                  and rank_search between 26 and 27
         LOOP


@@ -258,7 +278,7 @@ BEGIN
              OR position(bnd_name in lower(name->'name')) > 0)
         AND placex.class = 'place' AND placex.type = bnd.extratags->'place'
         AND placex.osm_type = 'N'
              OR position(bnd_name in lower(name->'name')) > 0)
         AND placex.class = 'place' AND placex.type = bnd.extratags->'place'
         AND placex.osm_type = 'N'

-        AND placex.linked_place_id is null
+        AND (placex.linked_place_id is null or placex.linked_place_id = bnd.place_id)

         AND placex.rank_search < 26 -- needed to select the right index
         AND placex.type != 'postcode'
         AND ST_Covers(bnd.geometry, placex.geometry)
         AND placex.rank_search < 26 -- needed to select the right index
         AND placex.type != 'postcode'
         AND ST_Covers(bnd.geometry, placex.geometry)


@@ -274,7 +294,7 @@ BEGIN
       WHERE placex.class = 'place' AND placex.osm_type = 'N'
         AND placex.extratags ? 'wikidata' -- needed to select right index
         AND placex.extratags->'wikidata' = bnd.extratags->'wikidata'
       WHERE placex.class = 'place' AND placex.osm_type = 'N'
         AND placex.extratags ? 'wikidata' -- needed to select right index
         AND placex.extratags->'wikidata' = bnd.extratags->'wikidata'

-        AND placex.linked_place_id is null
+        AND (placex.linked_place_id is null or placex.linked_place_id = bnd.place_id)

         AND placex.rank_search < 26
         AND _st_covers(bnd.geometry, placex.geometry)
       ORDER BY lower(name->'name') = bnd_name desc
         AND placex.rank_search < 26
         AND _st_covers(bnd.geometry, placex.geometry)
       ORDER BY lower(name->'name') = bnd_name desc


@@ -298,7 +318,7 @@ BEGIN
              OR (bnd.rank_address = 0 and placex.rank_search = bnd.rank_search))
         AND placex.osm_type = 'N'
         AND placex.class = 'place'
              OR (bnd.rank_address = 0 and placex.rank_search = bnd.rank_search))
         AND placex.osm_type = 'N'
         AND placex.class = 'place'

-        AND placex.linked_place_id is null
+        AND (placex.linked_place_id is null or placex.linked_place_id = bnd.place_id)

         AND placex.rank_search < 26 -- needed to select the right index
         AND placex.type != 'postcode'
         AND ST_Covers(bnd.geometry, placex.geometry)
         AND placex.rank_search < 26 -- needed to select the right index
         AND placex.type != 'postcode'
         AND ST_Covers(bnd.geometry, placex.geometry)


@@ -342,9 +362,10 @@ BEGIN
     WHERE s.place_id = parent_place_id;
 
   FOR addr_item IN
     WHERE s.place_id = parent_place_id;
 
   FOR addr_item IN

-    SELECT (get_addr_tag_rank(key, country)).*, key,
+    SELECT ranks.*, key,

            token_get_address_search_tokens(token_info, key) as search_tokens
            token_get_address_search_tokens(token_info, key) as search_tokens

-      FROM token_get_address_keys(token_info) as key
+      FROM token_get_address_keys(token_info) as key,
+           LATERAL get_addr_tag_rank(key, country) as ranks

       WHERE not token_get_address_search_tokens(token_info, key) <@ parent_address_vector
   LOOP
     addr_place := get_address_place(in_partition, geometry,
       WHERE not token_get_address_search_tokens(token_info, key) <@ parent_address_vector
   LOOP
     addr_place := get_address_place(in_partition, geometry,


@@ -432,6 +453,7 @@ CREATE OR REPLACE FUNCTION insert_addresslines(obj_place_id BIGINT,
                                                maxrank SMALLINT,
                                                token_info JSONB,
                                                geometry GEOMETRY,
                                                maxrank SMALLINT,
                                                token_info JSONB,
                                                geometry GEOMETRY,

+                                               centroid GEOMETRY,

                                                country TEXT,
                                                OUT parent_place_id BIGINT,
                                                OUT postcode TEXT,
                                                country TEXT,
                                                OUT parent_place_id BIGINT,
                                                OUT postcode TEXT,


@@ -456,10 +478,12 @@ BEGIN
   address_havelevel := array_fill(false, ARRAY[maxrank]);
 
   FOR location IN
   address_havelevel := array_fill(false, ARRAY[maxrank]);
 
   FOR location IN

-    SELECT (get_address_place(partition, geometry, from_rank, to_rank,
-                              extent, token_info, key)).*, key
-      FROM (SELECT (get_addr_tag_rank(key, country)).*, key
-              FROM token_get_address_keys(token_info) as key) x
+    SELECT apl.*, key
+      FROM (SELECT extra.*, key
+              FROM token_get_address_keys(token_info) as key,
+                   LATERAL get_addr_tag_rank(key, country) as extra) x,
+           LATERAL get_address_place(partition, geometry, from_rank, to_rank,
+                              extent, token_info, key) as apl

       ORDER BY rank_address, distance, isguess desc
   LOOP
     IF location.place_id is null THEN
       ORDER BY rank_address, distance, isguess desc
   LOOP
     IF location.place_id is null THEN


@@ -492,7 +516,7 @@ BEGIN
   END LOOP;
 
   FOR location IN
   END LOOP;
 
   FOR location IN

-    SELECT * FROM getNearFeatures(partition, geometry, maxrank)
+    SELECT * FROM getNearFeatures(partition, geometry, centroid, maxrank)

     WHERE not addr_place_ids @> ARRAY[place_id]
     ORDER BY rank_address, isguess asc,
              distance *
     WHERE not addr_place_ids @> ARRAY[place_id]
     ORDER BY rank_address, isguess asc,
              distance *


@@ -735,9 +759,6 @@ BEGIN
   DELETE FROM place_addressline WHERE place_id = NEW.place_id;
   result := deleteRoad(NEW.partition, NEW.place_id);
   result := deleteLocationArea(NEW.partition, NEW.place_id, NEW.rank_search);
   DELETE FROM place_addressline 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, indexed_status = 2
-         where linked_place_id = NEW.place_id;
-  -- update not necessary for osmline, cause linked_place_id does not exist

 
   NEW.extratags := NEW.extratags - 'linked_place'::TEXT;
 
 
   NEW.extratags := NEW.extratags - 'linked_place'::TEXT;
 


@@ -746,11 +767,11 @@ BEGIN
   linked_place := NEW.linked_place_id;
   NEW.linked_place_id := OLD.linked_place_id;
 
   linked_place := NEW.linked_place_id;
   NEW.linked_place_id := OLD.linked_place_id;
 

-  IF NEW.linked_place_id is not null THEN
-    NEW.token_info := null;
-    {% if debug %}RAISE WARNING 'place already linked to %', OLD.linked_place_id;{% endif %}
-    RETURN NEW;
-  END IF;
+  -- Remove linkage, if we have computed a different new linkee.
+  UPDATE placex SET linked_place_id = null, indexed_status = 2
+    WHERE linked_place_id = NEW.place_id
+          and (linked_place is null or linked_place_id != linked_place);
+  -- update not necessary for osmline, cause linked_place_id does not exist

 
   -- Postcodes are just here to compute the centroids. They are not searchable
   -- unless they are a boundary=postal_code.
 
   -- Postcodes are just here to compute the centroids. They are not searchable
   -- unless they are a boundary=postal_code.


@@ -796,6 +817,16 @@ BEGIN
                             NEW.class, NEW.type, NEW.admin_level,
                             (NEW.extratags->'capital') = 'yes',
                             NEW.address->'postcode');
                             NEW.class, NEW.type, NEW.admin_level,
                             (NEW.extratags->'capital') = 'yes',
                             NEW.address->'postcode');

+
+  -- Short-cut out for linked places. Note that this must happen after the
+  -- address rank has been recomputed. The linking might nullify a shift in
+  -- address rank.
+  IF NEW.linked_place_id is not null THEN
+    NEW.token_info := null;
+    {% if debug %}RAISE WARNING 'place already linked to %', OLD.linked_place_id;{% endif %}
+    RETURN NEW;
+  END IF;
+

   -- We must always increase the address level relative to the admin boundary.
   IF NEW.class = 'boundary' and NEW.type = 'administrative'
      and NEW.osm_type = 'R' and NEW.rank_address > 0
   -- We must always increase the address level relative to the admin boundary.
   IF NEW.class = 'boundary' and NEW.type = 'administrative'
      and NEW.osm_type = 'R' and NEW.rank_address > 0


@@ -964,15 +995,6 @@ BEGIN
       {% endif %}
 
       NEW.token_info := token_strip_info(NEW.token_info);
       {% endif %}
 
       NEW.token_info := token_strip_info(NEW.token_info);

-      -- If the address was inherited from a surrounding building,
-      -- do not add it permanently to the table.
-      IF NEW.address ? '_inherited' THEN
-        IF NEW.address ? '_unlisted_place' THEN
-          NEW.address := hstore('_unlisted_place', NEW.address->'_unlisted_place');
-        ELSE
-          NEW.address := null;
-        END IF;
-      END IF;

 
       RETURN NEW;
     END IF;
 
       RETURN NEW;
     END IF;


@@ -983,7 +1005,14 @@ BEGIN
   -- Full indexing
   {% if debug %}RAISE WARNING 'Using full index mode for % %', NEW.osm_type, NEW.osm_id;{% endif %}
   IF linked_place is not null THEN
   -- Full indexing
   {% if debug %}RAISE WARNING 'Using full index mode for % %', NEW.osm_type, NEW.osm_id;{% endif %}
   IF linked_place is not null THEN

-    SELECT * INTO location FROM placex WHERE place_id = linked_place;
+    -- Recompute the ranks here as the ones from the linked place might
+    -- have been shifted to accomodate surrounding boundaries.
+    SELECT place_id, osm_id, class, type, extratags,
+           centroid, geometry,
+           (compute_place_rank(country_code, osm_type, class, type, admin_level,
+                              (extratags->'capital') = 'yes', null)).*
+      INTO location
+      FROM placex WHERE place_id = linked_place;

 
     {% if debug %}RAISE WARNING 'Linked %', location;{% endif %}
 
 
     {% if debug %}RAISE WARNING 'Linked %', location;{% endif %}
 


@@ -994,11 +1023,11 @@ BEGIN
         NEW.centroid := geom;
     END IF;
 
         NEW.centroid := geom;
     END IF;
 

-    {% if debug %}RAISE WARNING 'parent address: % rank address: %', parent_address_level, location.rank_address;{% endif %}
-    IF location.rank_address > parent_address_level
-       and location.rank_address < 26
+    {% if debug %}RAISE WARNING 'parent address: % rank address: %', parent_address_level, location.address_rank;{% endif %}
+    IF location.address_rank > parent_address_level
+       and location.address_rank < 26

     THEN
     THEN

-      NEW.rank_address := location.rank_address;
+      NEW.rank_address := location.address_rank;

     END IF;
 
     -- merge in extra tags
     END IF;
 
     -- merge in extra tags


@@ -1044,16 +1073,22 @@ BEGIN
      AND NEW.class = 'boundary' AND NEW.type = 'administrative'
      AND NEW.country_code IS NOT NULL AND NEW.osm_type = 'R'
   THEN
      AND NEW.class = 'boundary' AND NEW.type = 'administrative'
      AND NEW.country_code IS NOT NULL AND NEW.osm_type = 'R'
   THEN

-    -- Update the list of country names. Adding an additional sanity
-    -- check here: make sure the country does overlap with the area where
-    -- we expect it to be as per static country grid.
+    -- Update the list of country names.
+    -- Only take the name from the largest area for the given country code
+    -- in the hope that this is the authoritive one.
+    -- Also replace any old names so that all mapping mistakes can
+    -- be fixed through regular OSM updates.

     FOR location IN
     FOR location IN

-      SELECT country_code FROM country_osm_grid
-       WHERE ST_Covers(geometry, NEW.centroid) and country_code = NEW.country_code
+      SELECT osm_id FROM placex
+       WHERE rank_search = 4 and osm_type = 'R'
+             and country_code = NEW.country_code
+       ORDER BY ST_Area(geometry) desc

        LIMIT 1
     LOOP
        LIMIT 1
     LOOP

-      {% if debug %}RAISE WARNING 'Updating names for country '%' with: %', NEW.country_code, NEW.name;{% endif %}
-      UPDATE country_name SET name = name || NEW.name WHERE country_code = NEW.country_code;
+      IF location.osm_id = NEW.osm_id THEN
+        {% if debug %}RAISE WARNING 'Updating names for country '%' with: %', NEW.country_code, NEW.name;{% endif %}
+        UPDATE country_name SET derived_name = NEW.name WHERE country_code = NEW.country_code;
+      END IF;

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


@@ -1081,7 +1116,8 @@ BEGIN
   END IF;
 
   SELECT * FROM insert_addresslines(NEW.place_id, NEW.partition, max_rank,
   END IF;
 
   SELECT * FROM insert_addresslines(NEW.place_id, NEW.partition, max_rank,

-                                    NEW.token_info, geom, NEW.country_code)
+                                    NEW.token_info, geom, NEW.centroid,
+                                    NEW.country_code)

     INTO NEW.parent_place_id, NEW.postcode, nameaddress_vector;
 
   {% if debug %}RAISE WARNING 'RETURN insert_addresslines: %, %, %', NEW.parent_place_id, NEW.postcode, nameaddress_vector;{% endif %}
     INTO NEW.parent_place_id, NEW.postcode, nameaddress_vector;
 
   {% if debug %}RAISE WARNING 'RETURN insert_addresslines: %, %, %', NEW.parent_place_id, NEW.postcode, nameaddress_vector;{% endif %}