flexible crs in geometries dict in aggregate_spatial (#245)

* allow crs in aggregate_spatial as int, str

* add test for crs int

* pre-commit

* update input geometry types

* update input geometry types

* run pre commit

openeo_processes_dask/process_implementations/cubes/aggregate.py

@@ -136,16 +136,31 @@ def aggregate_spatial(
                     feature["properties"] = {}
                 elif feature["properties"] is None:
                     feature["properties"] = {}
-            DEFAULT_CRS = (
-                geometries.get("crs", {}).get("properties", {}).get("name", DEFAULT_CRS)
-            )
+            if isinstance(geometries.get("crs", {}), dict):
+                DEFAULT_CRS = (
+                    geometries.get("crs", {})
+                    .get("properties", {})
+                    .get("name", DEFAULT_CRS)
+                )
+            else:
+                DEFAULT_CRS = int(geometries.get("crs", {}))
+            logger.info(f"CRS in geometries: {DEFAULT_CRS}.")
+
         if "type" in geometries and geometries["type"] == "FeatureCollection":
             gdf = gpd.GeoDataFrame.from_features(geometries, crs=DEFAULT_CRS)
         elif "type" in geometries and geometries["type"] in ["Polygon"]:
             polygon = shapely.geometry.Polygon(geometries["coordinates"][0])
             gdf = gpd.GeoDataFrame(geometry=[polygon])
             gdf.crs = DEFAULT_CRS
 
+    if isinstance(geometries, xr.Dataset):
+        if hasattr(geometries, "xvec"):
+            gdf = geometries.xvec.to_geodataframe()
+
+    if isinstance(geometries, gpd.GeoDataFrame):
+        gdf = geometries
+
+    gdf = gdf.to_crs(data.rio.crs)
     geometries = gdf.geometry.values
 
     positional_parameters = {"data": 0}

tests/test_aggregate.py

@@ -3,6 +3,8 @@
 import geopandas as gpd
 import numpy as np
 import pytest
+import xarray as xr
+import xvec
 from openeo_pg_parser_networkx.pg_schema import ParameterReference, TemporalInterval
 
 from openeo_processes_dask.process_implementations.cubes.aggregate import (
@@ -148,6 +150,34 @@ def test_aggregate_spatial(
 
     assert (output_cube.values == expected_values).all()
 
+    gdf = gpd.GeoDataFrame.from_features(polygon_geometry_small, crs="EPSG:4326")
+    gdf_equi7 = gdf.to_crs(
+        "+proj=aeqd +lat_0=53 +lon_0=24 +x_0=5837287.81977 +y_0=2121415.69617 +datum=WGS84 +units=m +no_defs"
+    )
+    output_cube_transform = aggregate_spatial(
+        data=reduced_cube, geometries=gdf_equi7, reducer=reducer
+    )
+    assert len(output_cube_transform.dims) == len(output_cube.dims)
+    assert output_cube_transform.shape == output_cube.shape
+
+    geometry_cube = xr.Dataset(
+        data_vars={"variable": (["geometry"], np.arange(len(gdf)))},
+        coords={"geometry": gdf["geometry"].values},
+    ).xvec.set_geom_indexes("geometry", crs=gdf.crs)
+    output_cube_transform = aggregate_spatial(
+        data=reduced_cube, geometries=geometry_cube, reducer=reducer
+    )
+    assert len(output_cube_transform.dims) == len(output_cube.dims)
+    assert output_cube_transform.shape == output_cube.shape
+
+    polygon_geometry_small["crs"] = 4326
+
+    output_cube = aggregate_spatial(
+        data=reduced_cube, geometries=polygon_geometry_small, reducer=reducer
+    )
+
+    assert len(output_cube.dims) < len(reduced_cube.dims)
+
     geometry_url = "https://raw.githubusercontent.com/ValentinaHutter/polygons/master/polygons_small.json"
     output_cube = aggregate_spatial(
         data=reduced_cube, geometries=geometry_url, reducer=reducer