[GH-2004] Geopandas.GeoSeries: Implement Test Framework (#2005)

* Fix small nit in series __repr__()

* Add test_non_geom_fails()

* test_constructor on all different geometry types

* Change Series.area return type to pd.Series to match gpd behavior and add area tests

* Fix GeoSeries.to_pandas() and fix refactor tests

* pre-commit

* Test if sgpd_res equals sedona result and gpd result

* Remove run_sedona_sql test

* Rename test_geoseries.py to test_match_geopandas_series.py

* Make area( return ps.Series instead of pd.Series

* Add new test_geoseries to mimic the scala tests

* Use smaller tests for test_geoseries and hard-code expected results

* Remove check_less_precise for version compatibility

Author: petern48

python/sedona/geopandas/geoseries.py

@@ -50,8 +50,8 @@ def __repr__(self) -> str:
         Return a string representation of the GeoSeries in WKT format.
         """
         try:
-            pandas_series = self.to_geopandas()
-            return gpd.GeoSeries(pandas_series).__repr__()
+            gpd_series = self.to_geopandas()
+            return gpd_series.__repr__()
 
         except Exception as e:
             # Fallback to parent's representation if conversion fails
@@ -176,7 +176,7 @@ def _process_geometry_column(
             A GeoSeries with the operation applied to the geometry column.
         """
         # Find the first column with BinaryType or GeometryType
-        first_col = self.get_first_geometry_column()
+        first_col = self.get_first_geometry_column()  # TODO: fixme
 
         if first_col:
             data_type = self._internal.spark_frame.schema[first_col].dataType
@@ -230,9 +230,16 @@ def to_geopandas(self) -> gpd.GeoSeries:
         return self._to_geopandas()
 
     def _to_geopandas(self) -> gpd.GeoSeries:
-        return gpd.GeoSeries(
-            self._to_internal_pandas().map(lambda wkb: shapely.wkb.loads(bytes(wkb)))
-        )
+        pd_series = self._to_internal_pandas()
+        try:
+            return gpd.GeoSeries(
+                pd_series.map(lambda wkb: shapely.wkb.loads(bytes(wkb)))
+            )
+        except Exception as e:
+            return gpd.GeoSeries(pd_series)
+
+    def to_spark_pandas(self) -> pspd.Series:
+        return pspd.Series(self._to_internal_pandas())
 
     @property
     def geometry(self) -> "GeoSeries":
@@ -274,7 +281,7 @@ def copy(self, deep=False):
             return self
 
     @property
-    def area(self) -> "GeoSeries":
+    def area(self) -> pspd.Series:
         """
         Returns a Series containing the area of each geometry in the GeoSeries expressed in the units of the CRS.
 
@@ -295,7 +302,7 @@ def area(self) -> "GeoSeries":
         1    4.0
         dtype: float64
         """
-        return self._process_geometry_column("ST_Area", rename="area")
+        return self._process_geometry_column("ST_Area", rename="area").to_spark_pandas()
 
     @property
     def crs(self):
@@ -521,7 +528,7 @@ def buffer(
         mitre_limit=5.0,
         single_sided=False,
         **kwargs,
-    ):
+    ) -> "GeoSeries":
         """
         Returns a GeoSeries of geometries representing all points within a given distance of each geometric object.
 

python/tests/geopandas/test_geoseries.py

@@ -14,109 +14,55 @@
 # KIND, either express or implied.  See the License for the
 # specific language governing permissions and limitations
 # under the License.
-import os
-import shutil
-import tempfile
-from geopandas.testing import assert_geoseries_equal
 
-from shapely.geometry import (
-    Point,
-    Polygon,
-)
-
-from sedona.geopandas import GeoSeries
+import pandas as pd
+import geopandas as gpd
+import sedona.geopandas as sgpd
 from tests.test_base import TestBase
-import pyspark.pandas as ps
+from shapely import wkt
+from shapely.geometry import Point, LineString, Polygon, GeometryCollection
+from pandas.testing import assert_series_equal
 
 
-class TestSeries(TestBase):
+class TestGeoSeries(TestBase):
     def setup_method(self):
-        self.tempdir = tempfile.mkdtemp()
-        self.t1 = Polygon([(0, 0), (1, 0), (1, 1)])
-        self.t2 = Polygon([(0, 0), (1, 1), (0, 1)])
-        self.sq = Polygon([(0, 0), (1, 0), (1, 1), (0, 1)])
-        self.g1 = GeoSeries([self.t1, self.t2])
-        self.g2 = GeoSeries([self.sq, self.t1])
-        self.g3 = GeoSeries([self.t1, self.t2], crs="epsg:4326")
-        self.g4 = GeoSeries([self.t2, self.t1])
-
-    def teardown_method(self):
-        shutil.rmtree(self.tempdir)
-
-    def test_constructor(self):
-        s = GeoSeries([Point(x, x) for x in range(3)])
-        check_geoseries_equal(s, s)
-
-    def test_psdf(self):
-        # this is to make sure the spark session works with pandas on spark api
-        psdf = ps.DataFrame(
-            {
-                "a": [1, 2, 3, 4, 5, 6],
-                "b": [100, 200, 300, 400, 500, 600],
-                "c": ["one", "two", "three", "four", "five", "six"],
-            },
-            index=[10, 20, 30, 40, 50, 60],
+        self.geoseries = sgpd.GeoSeries(
+            [
+                Point(2.3, -1),
+                LineString([(0.5, 0), (0, -3)]),
+                Polygon([(-1, -1), (-0.3, 5), (1, 1.2)]),
+                GeometryCollection(
+                    [
+                        Point(2.3, -1),
+                        LineString([(0.5, 0), (0, -3)]),
+                        Polygon([(-1, -1), (-0.3, 5), (1, 1.2)]),
+                    ]
+                ),
+            ]
         )
-        assert psdf.count().count() == 3
-
-    def test_internal_st_function(self):
-        # this is to make sure the spark session works with internal sedona udfs
-        baseDf = self.spark.sql(
-            "SELECT ST_GeomFromWKT('POLYGON ((50 50 1, 50 80 2, 80 80 3, 80 50 2, 50 50 1))') as geom"
-        )
-        actual = baseDf.selectExpr("ST_AsText(ST_Expand(geom, 10))").first()[0]
-        expected = "POLYGON Z((40 40 -9, 40 90 -9, 90 90 13, 90 40 13, 40 40 -9))"
-        assert expected == actual
-
-    def test_type(self):
-        assert type(self.g1) is GeoSeries
-        assert type(self.g2) is GeoSeries
-        assert type(self.g3) is GeoSeries
-        assert type(self.g4) is GeoSeries
 
-    def test_copy(self):
-        gc = self.g3.copy()
-        assert type(gc) is GeoSeries
-        assert self.g3.name == gc.name
+    def check_sgpd_equals_gpd(self, actual: sgpd.GeoSeries, expected: gpd.GeoSeries):
+        assert isinstance(actual, sgpd.GeoSeries)
+        assert isinstance(expected, gpd.GeoSeries)
+        assert len(actual) == len(expected)
+        sgpd_result = actual.to_geopandas()
+        for a, e in zip(sgpd_result, expected):
+            self.assert_geometry_almost_equal(a, e)
 
     def test_area(self):
-        area = self.g1.area
-        assert area is not None
-        assert type(area) is GeoSeries
-        assert area.count() == 2
+        result = self.geoseries.area.to_pandas()
+        expected = pd.Series([0.0, 0.0, 5.23, 5.23])
+        assert result.count() > 0
+        assert_series_equal(result, expected)
 
     def test_buffer(self):
-        buffer = self.g1.buffer(0.2)
-        assert buffer is not None
-        assert type(buffer) is GeoSeries
-        assert buffer.count() == 2
-
-    def test_buffer_then_area(self):
-        area = self.g1.buffer(0.2).area
-        assert area is not None
-        assert type(area) is GeoSeries
-        assert area.count() == 2
-
-    def test_buffer_then_geoparquet(self):
-        temp_file_path = os.path.join(
-            self.tempdir, next(tempfile._get_candidate_names()) + ".parquet"
-        )
-        self.g1.buffer(0.2).to_parquet(temp_file_path)
-        assert os.path.exists(temp_file_path)
-
-
-# -----------------------------------------------------------------------------
-# # Utils
-# -----------------------------------------------------------------------------
-
-
-def check_geoseries_equal(s1, s2):
-    assert isinstance(s1, GeoSeries)
-    assert isinstance(s1.geometry, GeoSeries)
-    assert isinstance(s2, GeoSeries)
-    assert isinstance(s2.geometry, GeoSeries)
-    if isinstance(s1, GeoSeries):
-        s1 = s1.to_geopandas()
-    if isinstance(s2, GeoSeries):
-        s2 = s2.to_geopandas()
-    assert_geoseries_equal(s1, s2)
+        result = self.geoseries.buffer(1)
+        expected = [
+            "POLYGON ((3.300000000000000 -1.000000000000000, 3.280785280403230 -1.195090322016128, 3.223879532511287 -1.382683432365090, 3.131469612302545 -1.555570233019602, 3.007106781186547 -1.707106781186547, 2.855570233019602 -1.831469612302545, 2.682683432365089 -1.923879532511287, 2.495090322016128 -1.980785280403230, 2.300000000000000 -2.000000000000000, 2.104909677983872 -1.980785280403230, 1.917316567634910 -1.923879532511287, 1.744429766980398 -1.831469612302545, 1.592893218813452 -1.707106781186547, 1.468530387697454 -1.555570233019602, 1.376120467488713 -1.382683432365090, 1.319214719596769 -1.195090322016129, 1.300000000000000 -1.000000000000000, 1.319214719596769 -0.804909677983872, 1.376120467488713 -0.617316567634910, 1.468530387697454 -0.444429766980398, 1.592893218813452 -0.292893218813453, 1.744429766980398 -0.168530387697455, 1.917316567634910 -0.076120467488713, 2.104909677983871 -0.019214719596770, 2.300000000000000 0.000000000000000, 2.495090322016128 -0.019214719596770, 2.682683432365090 -0.076120467488713, 2.855570233019602 -0.168530387697455, 3.007106781186547 -0.292893218813452, 3.131469612302545 -0.444429766980398, 3.223879532511286 -0.617316567634910, 3.280785280403230 -0.804909677983871, 3.300000000000000 -1.000000000000000))",
+            "POLYGON ((0.986393923832144 -3.164398987305357, 0.935367989801224 -3.353676015097457, 0.848396388482656 -3.529361471973156, 0.728821389740875 -3.684703864350261, 0.581238193719096 -3.813733471206735, 0.411318339874827 -3.911491757111723, 0.225591752899151 -3.974221925961374, 0.031195801372873 -3.999513292546280, -0.164398987305357 -3.986393923832144, -0.353676015097457 -3.935367989801224, -0.529361471973156 -3.848396388482656, -0.684703864350260 -3.728821389740875, -0.813733471206735 -3.581238193719097, -0.911491757111723 -3.411318339874827, -0.974221925961374 -3.225591752899151, -0.999513292546279 -3.031195801372874, -0.986393923832144 -2.835601012694643, -0.486393923832144 0.164398987305357, -0.435367989801224 0.353676015097458, -0.348396388482656 0.529361471973156, -0.228821389740875 0.684703864350260, -0.081238193719096 0.813733471206735, 0.088681660125173 0.911491757111723, 0.274408247100849 0.974221925961374, 0.468804198627127 0.999513292546279, 0.664398987305357 0.986393923832144, 0.853676015097457 0.935367989801224, 1.029361471973156 0.848396388482656, 1.184703864350260 0.728821389740875, 1.313733471206735 0.581238193719096, 1.411491757111723 0.411318339874827, 1.474221925961374 0.225591752899151, 1.499513292546280 0.031195801372873, 1.486393923832144 -0.164398987305357, 0.986393923832144 -3.164398987305357))",
+            "POLYGON ((-0.260059926604056 -1.672672793996312, -0.403493516968407 -1.802608257932399, -0.569270104475049 -1.902480890158382, -0.751180291696993 -1.968549819451744, -0.942410374326119 -1.998340340272165, -1.135797558140999 -1.990736606370705, -1.324098251632999 -1.946023426395157, -1.500259385009482 -1.865875595977814, -1.657682592935656 -1.753295165887471, -1.790471365675451 -1.612498995956065, -1.893651911234561 -1.448760806607280, -1.963359455800552 -1.268213644171327, -1.996983004332570 -1.077620158927971, -1.993263139087243 -0.884119300439822, -1.293263139087243 5.115880699560178, -1.252729137381052 5.303820984767603, -1.176977926029782 5.480530662139786, -1.068809614934931 5.639477736894415, -0.932222597700009 5.774786800970082, -0.772265752785876 5.881456214877171, -0.594851813959648 5.955542991081357, -0.406538808715662 5.994308544787506, -0.214287643700274 5.996319924510972, -0.025204797887634 5.961502780493132, 0.153720365261017 5.891144113007211, 0.315873956515097 5.787844698964485, 0.455262040354176 5.655422955350244, 0.566732198133767 5.498773793134933, 0.646163984953356 5.323687679062990, 1.946163984953356 1.523687679062990, 1.993263731568509 1.315875621036525, 1.995265095723606 1.102802318781350, 1.952077207005038 0.894142203137658, 1.865660978573300 0.699369327572194, 1.739940073395944 0.527327206003688, -0.260059926604056 -1.672672793996312))",
+            "POLYGON ((-0.844303230213814 -1.983056850984667, -0.942410374326119 -1.998340340272165, -1.135797558140999 -1.990736606370705, -1.324098251632999 -1.946023426395157, -1.500259385009482 -1.865875595977814, -1.657682592935656 -1.753295165887471, -1.790471365675451 -1.612498995956065, -1.893651911234561 -1.448760806607280, -1.963359455800552 -1.268213644171327, -1.996983004332570 -1.077620158927971, -1.993263139087243 -0.884119300439822, -1.293263139087243 5.115880699560178, -1.252729137381052 5.303820984767603, -1.176977926029782 5.480530662139786, -1.068809614934931 5.639477736894415, -0.932222597700009 5.774786800970082, -0.772265752785876 5.881456214877171, -0.594851813959648 5.955542991081357, -0.406538808715662 5.994308544787506, -0.214287643700274 5.996319924510972, -0.025204797887634 5.961502780493132, 0.153720365261017 5.891144113007211, 0.315873956515097 5.787844698964485, 0.455262040354176 5.655422955350244, 0.566732198133767 5.498773793134933, 0.646163984953356 5.323687679062990, 1.946163984953356 1.523687679062990, 1.993263731568509 1.315875621036525, 1.995265095723606 1.102802318781350, 1.952077207005038 0.894142203137658, 1.865660978573300 0.699369327572194, 1.739940073395944 0.527327206003688, 1.471895863976614 0.232478575642425, 1.474221925961374 0.225591752899151, 1.499513292546280 0.031195801372873, 1.486393923832144 -0.164398987305357, 1.426669391220515 -0.522746182975131, 1.468530387697454 -0.444429766980398, 1.592893218813452 -0.292893218813453, 1.744429766980398 -0.168530387697455, 1.917316567634910 -0.076120467488713, 2.104909677983871 -0.019214719596770, 2.300000000000000 0.000000000000000, 2.495090322016128 -0.019214719596770, 2.682683432365090 -0.076120467488713, 2.855570233019602 -0.168530387697455, 3.007106781186547 -0.292893218813452, 3.131469612302545 -0.444429766980398, 3.223879532511286 -0.617316567634910, 3.280785280403230 -0.804909677983871, 3.300000000000000 -1.000000000000000, 3.280785280403230 -1.195090322016128, 3.223879532511287 -1.382683432365090, 3.131469612302545 -1.555570233019602, 3.007106781186547 -1.707106781186547, 2.855570233019602 -1.831469612302545, 2.682683432365089 -1.923879532511287, 2.495090322016128 -1.980785280403230, 2.300000000000000 -2.000000000000000, 2.104909677983872 -1.980785280403230, 1.917316567634910 -1.923879532511287, 1.744429766980398 -1.831469612302545, 1.592893218813452 -1.707106781186547, 1.468530387697454 -1.555570233019602, 1.376120467488713 -1.382683432365090, 1.319214719596769 -1.195090322016129, 1.317505079406277 -1.177732053860557, 0.986393923832144 -3.164398987305357, 0.935367989801224 -3.353676015097457, 0.848396388482656 -3.529361471973156, 0.728821389740875 -3.684703864350261, 0.581238193719096 -3.813733471206735, 0.411318339874827 -3.911491757111723, 0.225591752899151 -3.974221925961374, 0.031195801372873 -3.999513292546280, -0.164398987305357 -3.986393923832144, -0.353676015097457 -3.935367989801224, -0.529361471973156 -3.848396388482656, -0.684703864350260 -3.728821389740875, -0.813733471206735 -3.581238193719097, -0.911491757111723 -3.411318339874827, -0.974221925961374 -3.225591752899151, -0.999513292546279 -3.031195801372874, -0.986393923832144 -2.835601012694643, -0.844303230213814 -1.983056850984667))",
+        ]
+        expected = gpd.GeoSeries([wkt.loads(wkt_str) for wkt_str in expected])
+        assert result.count() > 0
+        self.check_sgpd_equals_gpd(result, expected)