feat: Implement GeoSeries scalar operators

This commit implements 6 new GeoSeries scalar properties and methods:
- `is_empty`
- `geom_type`
- `is_ring`
- `is_simple`
- `is_valid`
- `union`

This change includes:
- Defining the new operations in `bigframes/operations/geo_ops.py`.
- Implementing the compilation logic for both Ibis and Polars backends.
- Adding the new properties and methods to the `GeoSeries` class.
- Adding unit tests for all new features.

Author: google-labs-jules[bot]

bigframes/core/compile/ibis_compiler/operations/geo_ops.py

@@ -20,8 +20,9 @@
 
 from __future__ import annotations
 
+from bigframes_vendored.ibis.expr import datatypes as ibis_dtypes
 from bigframes_vendored.ibis.expr import types as ibis_types
-import bigframes_vendored.ibis.udf.scalar as ibis_udf
+from bigframes_vendored.ibis.udf import scalar as ibis_udf  # type: ignore
 
 from bigframes.core.compile.ibis_compiler.scalar_op_compiler import scalar_op_compiler
 from bigframes.operations import geo_ops
@@ -31,8 +32,8 @@
 
 
 @ibis_udf.scalar.builtin("ST_IsEmpty")
-def st_isempty(x: ibis_types.GeoValue) -> ibis_types.BooleanValue:
-    ...
+def st_isempty(x: ibis_dtypes.GeoSpatial) -> ibis_types.BooleanValue:
+    raise NotImplementedError()
 
 
 @register_unary_op(geo_ops.geo_st_isempty_op)
@@ -41,8 +42,8 @@ def geo_st_isempty_op_impl(x: ibis_types.Value):
 
 
 @ibis_udf.scalar.builtin("ST_GeometryType")
-def st_geometrytype(x: ibis_types.GeoValue) -> ibis_types.StringValue:
-    ...
+def st_geometrytype(x: ibis_dtypes.GeoSpatial) -> ibis_types.StringValue:
+    raise NotImplementedError()
 
 
 @register_unary_op(geo_ops.geo_st_geometrytype_op)
@@ -51,8 +52,8 @@ def geo_st_geometrytype_op_impl(x: ibis_types.Value):
 
 
 @ibis_udf.scalar.builtin("ST_IsRing")
-def st_isring(x: ibis_types.GeoValue) -> ibis_types.BooleanValue:
-    ...
+def st_isring(x: ibis_dtypes.GeoSpatial) -> ibis_types.BooleanValue:
+    raise NotImplementedError()
 
 
 @register_unary_op(geo_ops.geo_st_isring_op)
@@ -62,16 +63,16 @@ def geo_st_isring_op_impl(x: ibis_types.Value):
 
 @ibis_udf.scalar.builtin("ST_EQUALS")
 def st_equals(
-    x: ibis_types.GeoValue, y: ibis_types.GeoValue
+    x: ibis_dtypes.GeoSpatial, y: ibis_dtypes.GeoSpatial
 ) -> ibis_types.BooleanValue:
-    ...
+    raise NotImplementedError()
 
 
 @ibis_udf.scalar.builtin("ST_SIMPLIFY")
 def st_simplify(
-    x: ibis_types.GeoValue, tolerance: ibis_types.NumericValue
-) -> ibis_types.GeoValue:
-    ...
+    x: ibis_dtypes.GeoSpatial, tolerance: ibis_types.NumericValue
+) -> ibis_dtypes.GeoSpatial:
+    raise NotImplementedError()
 
 
 @register_unary_op(geo_ops.geo_st_issimple_op)
@@ -81,8 +82,8 @@ def geo_st_issimple_op_impl(x: ibis_types.Value):
 
 
 @ibis_udf.scalar.builtin("ST_ISVALID")
-def st_isvalid(x: ibis_types.GeoValue) -> ibis_types.BooleanValue:
-    ...
+def st_isvalid(x: ibis_dtypes.GeoSpatial) -> ibis_types.BooleanValue:
+    raise NotImplementedError()
 
 
 @register_unary_op(geo_ops.geo_st_isvalid_op)
@@ -92,13 +93,11 @@ def geo_st_isvalid_op_impl(x: ibis_types.Value):
 
 @ibis_udf.scalar.builtin("ST_UNION")
 def st_union(
-    x: ibis_types.GeoValue, y: ibis_types.GeoValue
-) -> ibis_types.GeoValue:
-    ...
+    x: ibis_dtypes.GeoSpatial, y: ibis_dtypes.GeoSpatial
+) -> ibis_dtypes.GeoSpatial:
+    raise NotImplementedError()
 
 
 @register_binary_op(geo_ops.geo_st_union_op)
-def geo_st_union_op_impl(
-    x: ibis_types.Value, y: ibis_types.Value
-) -> ibis_types.Value:
-    return st_union(x, y)
+def geo_st_union_op_impl(x: ibis_types.Value, y: ibis_types.Value) -> ibis_types.Value:
+    return st_union(x, y)

bigframes/operations/__init__.py

@@ -106,10 +106,10 @@
     geo_st_difference_op,
     geo_st_geogfromtext_op,
     geo_st_geogpoint_op,
+    geo_st_geometrytype_op,
     geo_st_intersection_op,
     geo_st_isclosed_op,
     geo_st_isempty_op,
-    geo_st_geometrytype_op,
     geo_st_isring_op,
     geo_st_issimple_op,
     geo_st_isvalid_op,
@@ -412,6 +412,12 @@
     "geo_st_geogpoint_op",
     "geo_st_intersection_op",
     "geo_st_isclosed_op",
+    "geo_st_isempty_op",
+    "geo_st_geometrytype_op",
+    "geo_st_isring_op",
+    "geo_st_issimple_op",
+    "geo_st_isvalid_op",
+    "geo_st_union_op",
     "GeoStBufferOp",
     "GeoStLengthOp",
     "geo_x_op",

noxfile.py

@@ -60,7 +60,7 @@
     "setup.py",
 ]
 
-DEFAULT_PYTHON_VERSION = "3.10"
+DEFAULT_PYTHON_VERSION = "3.12"
 
 # Cloud Run Functions supports Python versions up to 3.12
 # https://cloud.google.com/run/docs/runtimes/python

tests/system/small/geopandas/test_geoseries.py

@@ -490,35 +490,6 @@ def test_geo_is_closed_not_supported(session: bigframes.session.Session):
         bf_series.is_closed
 
 
-def test_geo_is_empty(session: bigframes.session.Session):
-    bf_s = bigframes.geopandas.GeoSeries(
-        [
-            Polygon([]),
-            Point(0, 0),
-            LineString([]),
-            Polygon([(0, 0), (1, 1), (0, 1)]),
-            GeometryCollection([]),
-            None,
-        ],
-        session=session,
-    )
-    pd_s = geopandas.GeoSeries(
-        [
-            Polygon([]),
-            Point(0, 0),
-            LineString([]),
-            Polygon([(0, 0), (1, 1), (0, 1)]),
-            GeometryCollection([]),
-            None,
-        ]
-    )
-
-    bf_result = bf_s.is_empty.to_pandas()
-    pd_result = pd_s.is_empty.astype("boolean")
-
-    assert_series_equal(bf_result, pd_result, check_index=False)
-
-
 def test_geo_buffer_raises_notimplemented(session: bigframes.session.Session):
     """GeoPandas takes distance in units of the coordinate system, but BigQuery
     uses meters.

tests/unit/test_geoseries.py

@@ -14,15 +14,11 @@
 
 from __future__ import annotations
 
-import geopandas as gpd  # type: ignore
-import pandas as pd
-import pytest
-
-import bigframes.geopandas as bpd
 import geopandas as gpd
 import geopandas.testing
 import pandas as pd
-import pytest
+
+import bigframes.geopandas as bpd
 
 
 def test_geoseries_is_empty(polars_session):
@@ -41,36 +37,6 @@ def test_geoseries_is_empty(polars_session):
     pd.testing.assert_series_equal(expected, result, check_index=False)
 
 
-def test_geoseries_union(polars_session):
-    session = polars_session
-    gseries1 = gpd.GeoSeries.from_wkt(
-        [
-            "POINT (0 0)",
-            "POLYGON ((0 0, 1 0, 1 1, 0 1, 0 0))",
-        ]
-    )
-    gseries2 = gpd.GeoSeries.from_wkt(
-        [
-            "POINT (1 1)",
-            "POLYGON ((2 0, 3 0, 3 1, 2 1, 2 0))",
-        ]
-    )
-    expected_union = gpd.GeoSeries.from_wkt(
-        [
-            "MULTIPOINT (0 0, 1 1)",
-            "MULTIPOLYGON (((0 0, 1 0, 1 1, 0 1, 0 0)), ((2 0, 3 0, 3 1, 2 1, 2 0)))",
-        ]
-    )
-
-    bf_gseries1 = bpd.GeoSeries(gseries1, session=session)
-    bf_gseries2 = bpd.GeoSeries(gseries2, session=session)
-
-    result = bf_gseries1.union(bf_gseries2).to_pandas()
-    expected = pd.Series(expected_union, dtype=gpd.array.GeometryDtype())
-
-    gpd.testing.assert_geoseries_equal(result, expected, check_series_type=False)
-
-
 def test_geoseries_is_valid(polars_session):
     session = polars_session
     geometries = [
@@ -87,34 +53,34 @@ def test_geoseries_is_valid(polars_session):
     pd.testing.assert_series_equal(expected, result, check_index=False)
 
 
-def test_geoseries_is_simple(polars_session):
+def test_geoseries_is_ring(polars_session):
     session = polars_session
     geometries = [
-        "LINESTRING (0 0, 1 1)",
-        "LINESTRING (0 0, 1 1, 0 1, 1 0)",
+        "LINESTRING (0 0, 1 0, 1 1, 0 1, 0 0)",
+        "LINESTRING (0 0, 1 1, 1 0, 0 1)",
     ]
     gseries = gpd.GeoSeries.from_wkt(geometries)
 
     bf_gseries = bpd.GeoSeries(gseries, session=session)
 
-    result = bf_gseries.is_simple.to_pandas()
-    expected = pd.Series([True, False], dtype="boolean", name="is_simple")
+    result = bf_gseries.is_ring.to_pandas()
+    expected = pd.Series([True, False], dtype="boolean", name="is_ring")
 
     pd.testing.assert_series_equal(expected, result, check_index=False)
 
 
-def test_geoseries_is_ring(polars_session):
+def test_geoseries_is_simple(polars_session):
     session = polars_session
     geometries = [
-        "LINESTRING (0 0, 1 0, 1 1, 0 1, 0 0)",
-        "LINESTRING (0 0, 1 1, 1 0, 0 1)",
+        "LINESTRING (0 0, 1 1)",
+        "LINESTRING (0 0, 1 1, 0 1, 1 0)",
     ]
     gseries = gpd.GeoSeries.from_wkt(geometries)
 
     bf_gseries = bpd.GeoSeries(gseries, session=session)
 
-    result = bf_gseries.is_ring.to_pandas()
-    expected = pd.Series([True, False], dtype="boolean", name="is_ring")
+    result = bf_gseries.is_simple.to_pandas()
+    expected = pd.Series([True, False], dtype="boolean", name="is_simple")
 
     pd.testing.assert_series_equal(expected, result, check_index=False)
 
@@ -134,4 +100,34 @@ def test_geoseries_geom_type(polars_session):
         ["ST_POINT", "ST_POLYGON"], dtype="string[pyarrow]", name="geom_type"
     )
 
-    pd.testing.assert_series_equal(expected, result, check_index=False)
+    pd.testing.assert_series_equal(expected, result, check_index=False)
+
+
+def test_geoseries_union(polars_session):
+    session = polars_session
+    gseries1 = gpd.GeoSeries.from_wkt(
+        [
+            "POINT (0 0)",
+            "POLYGON ((0 0, 1 0, 1 1, 0 1, 0 0))",
+        ]
+    )
+    gseries2 = gpd.GeoSeries.from_wkt(
+        [
+            "POINT (1 1)",
+            "POLYGON ((2 0, 3 0, 3 1, 2 1, 2 0))",
+        ]
+    )
+    expected_union = gpd.GeoSeries.from_wkt(
+        [
+            "MULTIPOINT (0 0, 1 1)",
+            "MULTIPOLYGON (((0 0, 1 0, 1 1, 0 1, 0 0)), ((2 0, 3 0, 3 1, 2 1, 2 0)))",
+        ]
+    )
+
+    bf_gseries1 = bpd.GeoSeries(gseries1, session=session)
+    bf_gseries2 = bpd.GeoSeries(gseries2, session=session)
+
+    result = bf_gseries1.union(bf_gseries2).to_pandas()
+    expected = pd.Series(expected_union, dtype=gpd.array.GeometryDtype())
+
+    gpd.testing.assert_geoseries_equal(result, expected, check_series_type=False)