Supply bbox on gpkg publish so GeoServer skips getBounds

orchestration/assets/products.py

@@ -161,11 +161,12 @@ def _combine_asset(
     return _combine_asset
 
 
-def _geojson_to_geopackage(geojson_path: Path, layer_name: str, out_dir: Path) -> Path:
+def _geojson_to_geopackage(geojson_path: Path, layer_name: str, out_dir: Path):
     """Convert a GeoJSON file to a GeoPackage whose layer (table) is named
     *layer_name* (so the published GeoServer layer is named *layer_name*).
     GeoPackage is a single file with no field-name length limit, unlike the
-    zipped ESRI Shapefile this replaces. Returns the path to the .gpkg."""
+    zipped ESRI Shapefile this replaces. Returns (gpkg_path, bbox) where bbox is
+    (minx, miny, maxx, maxy) in EPSG:4326."""
     gdf = gpd.read_file(geojson_path)
     if gdf.empty:
         raise ValueError(f"{layer_name}: GeoJSON has no features; nothing to publish")
@@ -177,9 +178,13 @@ def _geojson_to_geopackage(geojson_path: Path, layer_name: str, out_dir: Path) -
     # computing bounds, so flatten to 2D — elevation remains an attribute.
     gdf["geometry"] = gdf.geometry.force_2d()
 
+    # Bounds are passed to GeoServer explicitly so it never calls getBounds on
+    # the gpkg store (which still trips the 3D-CRS bug at publish time).
+    minx, miny, maxx, maxy = (float(v) for v in gdf.total_bounds)
+
     gpkg_path = out_dir / f"{layer_name}.gpkg"
     gdf.to_file(gpkg_path, driver="GPKG", layer=layer_name)
-    return gpkg_path
+    return gpkg_path, (minx, miny, maxx, maxy)
 
 
 def _build_geoserver_asset(product: dict, group: str):
@@ -203,12 +208,13 @@ def _geoserver_asset(
             with tempfile.TemporaryDirectory() as tmpdir:
                 geojson = Path(tmpdir) / f"{pid}.geojson"
                 gcs.download_latest(pid, str(geojson))
-                gpkg_path = _geojson_to_geopackage(geojson, pid, Path(tmpdir))
+                gpkg_path, bbox = _geojson_to_geopackage(geojson, pid, Path(tmpdir))
                 actions = geoserver.publish_geopackage(
                     pid,
                     str(gpkg_path),
                     title=product.get("title", pid),
                     abstract=product.get("description", ""),
+                    bbox=bbox,
                 )
         except Exception:
             error = traceback.format_exc()

orchestration/resources/geoserver.py

@@ -75,6 +75,7 @@ def publish_geopackage(
         gpkg_path: str,
         title: Optional[str] = None,
         abstract: Optional[str] = None,
+        bbox: Optional[tuple] = None,
     ) -> dict:
         """Create-or-update a native GeoPackage datastore named *layer_name*
         from the GeoPackage at *gpkg_path* and publish its layer. Idempotent —
@@ -122,19 +123,33 @@ def publish_geopackage(
         # written by geopandas (== layer_name); name is the published layer
         # name. The store was recreated above, so this POST always creates a
         # fresh featuretype. srs is set explicitly — the data is WGS84.
-        ft_url = f"{base}/rest/workspaces/{ws}/datastores/{layer_name}/featuretypes"
+        feature_type: dict = {
+            "name": layer_name,
+            "nativeName": layer_name,
+            "title": title or layer_name,
+            "abstract": abstract or "",
+            "srs": "EPSG:4326",
+            "nativeCRS": "EPSG:4326",
+        }
+        # Supply the bounding boxes so GeoServer does not call getBounds on the
+        # gpkg store — that path still trips a 3D-CRS dimension error at publish.
+        # recalculate= (empty) tells GeoServer to keep the supplied boxes as-is.
+        recalc = ""
+        if bbox is not None:
+            minx, miny, maxx, maxy = bbox
+            box = {
+                "minx": minx, "miny": miny, "maxx": maxx, "maxy": maxy,
+                "crs": "EPSG:4326",
+            }
+            feature_type["nativeBoundingBox"] = box
+            feature_type["latLonBoundingBox"] = dict(box)
+            recalc = "?recalculate="
+
+        ft_url = f"{base}/rest/workspaces/{ws}/datastores/{layer_name}/featuretypes{recalc}"
         r = requests.post(
             ft_url,
             auth=auth,
-            json={
-                "featureType": {
-                    "name": layer_name,
-                    "nativeName": layer_name,
-                    "title": title or layer_name,
-                    "abstract": abstract or "",
-                    "srs": "EPSG:4326",
-                }
-            },
+            json={"featureType": feature_type},
             headers={"Content-Type": "application/json"},
             timeout=self.timeout,
         )