Tile duplicate housenumber filtering (#1391)

This PR introduces simple filtering of duplicate housenumbers.

Simple means that filtering is done withing the tile.
Duplicates are defined as same housenumber, street, block_number[1].

Duplicates are usually caused by POIs. People like to add addresses to them. Most POIs have names so to prioritize addresses we pick features without names first.

Formula is: `row_number() OVER(PARTITION BY concat(street, block_number, housenumber) ORDER BY has_name ASC) == 1`

layers/housenumber/housenumber.sql

@@ -15,9 +15,19 @@ SELECT
     osm_id,
     geometry,
     housenumber
-FROM osm_housenumber_point
-WHERE zoom_level >= 14
-  AND geometry && bbox;
+FROM (
+    SELECT
+        osm_id,
+        geometry,
+        housenumber,
+        row_number() OVER(PARTITION BY concat(street, block_number, housenumber) ORDER BY has_name ASC) as rn
+    FROM osm_housenumber_point
+    WHERE 1=1
+        AND zoom_level >= 14
+        AND geometry && bbox
+) t
+WHERE rn = 1;
+
 $$ LANGUAGE SQL STABLE
                 -- STRICT
                 PARALLEL SAFE;

layers/housenumber/housenumber.yaml

@@ -3,6 +3,7 @@ layer:
   description: |
       Everything in OpenStreetMap which contains a `addr:housenumber` tag useful for labelling housenumbers on a map.
       This adds significant size to *z14*. For buildings the centroid of the building is used as housenumber.
+      Duplicates within a tile are dropped if they have the same street/block_number (records without name tag are prioritized for preservation).
   buffer_size: 8
   srs: +proj=merc +a=6378137 +b=6378137 +lat_ts=0.0 +lon_0=0.0 +x_0=0.0 +y_0=0.0 +k=1.0 +units=m +nadgrids=@null +wktext +no_defs +over
   fields:

layers/housenumber/housenumber_centroid.sql

@@ -22,6 +22,16 @@ $$
     WHERE (full_update OR osm_id IN (SELECT osm_id FROM housenumber.osm_ids))
         AND ST_GeometryType(geometry) <> 'ST_Point'
         AND ST_IsValid(geometry);
+
+    -- we don't need exact name just to know if it's present
+    UPDATE osm_housenumber_point
+    SET has_name = 
+            CASE
+                WHEN has_name = '' THEN '0'
+                ELSE '1'
+            END
+    WHERE (full_update OR osm_id IN (SELECT osm_id FROM housenumber.osm_ids));
+
 $$ LANGUAGE SQL;
 
 SELECT convert_housenumber_point(true);

layers/housenumber/mapping.yaml

@@ -12,6 +12,15 @@ tables:
     - name: housenumber
       key: addr:housenumber
       type: string
+    - name: street
+      key: addr:street
+      type: string
+    - name: block_number
+      key: addr:block_number
+      type: string
+    - name: has_name
+      key: name
+      type: string
     type_mappings:
       points:
         addr:housenumber: