SPlit up

pipeline/download/greenspace_water.py

@@ -13,7 +13,7 @@ from pathlib import Path
 import osmium
 import polars as pl
 from pyproj import Transformer
-from shapely import wkb
+from shapely import make_valid, wkb
 from shapely.errors import GEOSException
 from shapely.geometry import MultiPolygon, Polygon
 from tqdm import tqdm
@@ -56,6 +56,22 @@ def _to_bng_polygon(geom):
     return geom
 
 
+def _polygonal_part(geom):
+    """The Polygon/MultiPolygon content of a geometry, or None if there is none."""
+    if geom.geom_type in ("Polygon", "MultiPolygon"):
+        return geom
+    if geom.geom_type == "GeometryCollection":
+        polygons = []
+        for part in geom.geoms:
+            if part.geom_type == "Polygon":
+                polygons.append(part)
+            elif part.geom_type == "MultiPolygon":
+                polygons.extend(part.geoms)
+        if polygons:
+            return MultiPolygon(polygons)
+    return None
+
+
 def _matches_tags(tags):
     """Check if an OSM element's tags match our greenspace/water criteria."""
     for key, values in GREENSPACE_TAGS.items():
@@ -91,7 +107,13 @@ class GreenspaceHandler(osmium.SimpleHandler):
             )
             return
 
-        if geom.is_empty or not geom.is_valid:
+        # Invalid geometries are often the largest, most complex park/water
+        # multipolygons (self-touching rings from OSM) — repair like pois.py
+        # rather than silently dropping them. make_valid may return a
+        # GeometryCollection with stray lines/points; keep only the polygons.
+        if not geom.is_valid:
+            geom = _polygonal_part(make_valid(geom))
+        if geom is None or geom.is_empty:
             return
 
         # Reproject to BNG for area calculation