spatialdata.models.force_2d() silently drops interior rings (holes) of polygons

Start in src/spatialdata/models/_utils.py at force_2d, then run the supplied repro.py with uv run to confirm the loss of polygon interiors. Check both Polygon and MultiPolygon handling described in the issue. Done means converting the 3D geometry to 2D while preserving holes and the expected area, without regressing the reproduction.

[!NOTE]
This whole message is AI-generated. The issue was automatically discovered and reported by an AI agent (Claude) during an autonomous bug hunt on the spatialdata code base. It has not been verified or triaged by a human yet; the needs: triage label is set so that a maintainer can confirm it. The reproduction script below was executed by the agent in an isolated environment (see Environment) and its output is pasted verbatim.

Summary

A 3D polygon with one hole (area 96) becomes a 2D polygon without holes (area 100). gdf.geometry.force_2d() from geopandas keeps the hole.

Severity (agent's assessment): medium — geometry area changes; the function is the remedy recommended by the ShapesModel.validate warning for 3D geometries

Where: src/spatialdata/models/_utils.py::force_2d (Polygon(np.array(shape.exterior.coords.xy).T) ignores interiors; same for MultiPolygon)

Expected behaviour

Holes preserved.

Reproduction

Save as repro.py and run uv run repro.py (the PEP 723 header pins spatialdata to the commit the bug was found on; replace the URL fragment with @main to test the current main branch).

# /// script
# requires-python = ">=3.12"
# dependencies = [
#     "spatialdata @ git+https://github.com/scverse/spatialdata.git@ccf1ea048d054b6624214bf618008a9f9ae223e0",
# ]
# ///
"""spatialdata.models.force_2d() drops interior rings (holes) of polygons."""
import warnings
import geopandas as gpd
from shapely.geometry import Polygon
from spatialdata.models import force_2d

warnings.simplefilter("ignore")
outer = [(0, 0, 1), (10, 0, 1), (10, 10, 1), (0, 10, 1)]
hole = [(4, 4, 1), (6, 4, 1), (6, 6, 1), (4, 6, 1)]
gdf = gpd.GeoDataFrame({"geometry": [Polygon(outer, [hole])]})
before = gdf.geometry.iloc[0]
print(f"before: has_z={before.has_z} interiors={len(before.interiors)} area={before.area}")
force_2d(gdf)
after = gdf.geometry.iloc[0]
print(f"after : has_z={after.has_z} interiors={len(after.interiors)} area={after.area}   (expected: interiors=1, area=96.0)")
print("geopandas reference (gdf.geometry.force_2d()) area:", gpd.GeoDataFrame({"geometry": [Polygon(outer, [hole])]}).geometry.force_2d().iloc[0].area)
bug = len(after.interiors) == 0
print("VERDICT:", "BUG REPRODUCED" if bug else "NOT REPRODUCED")

Observed output

before: has_z=True interiors=1 area=96.0
after : has_z=False interiors=0 area=100.0   (expected: interiors=1, area=96.0)
geopandas reference (gdf.geometry.force_2d()) area: 96.0
VERDICT: BUG REPRODUCED

Possible fix direction (unverified)

Delegate to gdf.geometry.force_2d() / shapely.force_2d, or pass holes=[np.array(r.coords)[:, :2] for r in shape.interiors].

Environment

uv run repro.py with the PEP 723 metadata in the script (fresh, isolated environment; spatialdata built from main @ ccf1ea0 (2026-08-28); Python 3.13, latest releases of the dependencies at run time: pandas 3.0, anndata 0.13, zarr 3.3, dask 2026.8, numpy 2.5, geopandas 1.1, shapely 2.1). macOS (arm64). Also reproduced in a second environment with pandas 2.3.3 / anndata 0.12.11 / numpy 2.4.4 / zarr 3.2.1.

Automatically generated; discovered by an AI agent (Claude) and not yet reviewed by a human.