[GH-1993] Geopandas __repr__() and to_geopandas() (#1990)

* Implement to_geopandas and __repr__ for GeoSeries

* Add check_geoseries_equal() to test_geoseries

* Use wkb.loads instead of from_wkb and fix lint

* Combine check_geoseries and check_geoseries_equal into one function

Author: petern48

python/sedona/geopandas/geoseries.py

@@ -28,6 +28,8 @@
 from pyspark.pandas.utils import scol_for
 from pyspark.sql.types import BinaryType
 
+import shapely
+
 from sedona.geopandas._typing import Label
 from sedona.geopandas.base import GeoFrame
 from sedona.geopandas.geodataframe import GeoDataFrame
@@ -43,6 +45,18 @@ def __getitem__(self, key: Any) -> Any:
         # Implementation of the abstract method
         raise NotImplementedError("This method is not implemented yet.")
 
+    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__()
+
+        except Exception as e:
+            # Fallback to parent's representation if conversion fails
+            return super().__repr__()
+
     def __init__(
         self,
         data=None,
@@ -206,13 +220,19 @@ def dtypes(self) -> Union[gpd.GeoSeries, pd.Series, Dtype]:
         # Implementation of the abstract method
         raise NotImplementedError("This method is not implemented yet.")
 
-    def to_geopandas(self) -> Union[gpd.GeoDataFrame, pd.Series]:
-        # Implementation of the abstract method
-        raise NotImplementedError("This method is not implemented yet.")
+    def to_geopandas(self) -> gpd.GeoSeries:
+        """
+        Convert the GeoSeries to a geopandas GeoSeries.
 
-    def _to_geopandas(self) -> Union[gpd.GeoDataFrame, pd.Series]:
-        # Implementation of the abstract method
-        raise NotImplementedError("This method is not implemented yet.")
+        Returns:
+        - geopandas.GeoSeries: A geopandas 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)))
+        )
 
     @property
     def geometry(self) -> "GeoSeries":

python/tests/geopandas/test_geoseries.py

@@ -17,6 +17,7 @@
 import os
 import shutil
 import tempfile
+from geopandas.testing import assert_geoseries_equal
 
 from shapely.geometry import (
     Point,
@@ -44,7 +45,7 @@ def teardown_method(self):
 
     def test_constructor(self):
         s = GeoSeries([Point(x, x) for x in range(3)])
-        check_geoseries(s)
+        check_geoseries_equal(s, s)
 
     def test_psdf(self):
         # this is to make sure the spark session works with pandas on spark api
@@ -109,6 +110,13 @@ def test_buffer_then_geoparquet(self):
 # -----------------------------------------------------------------------------
 
 
-def check_geoseries(s):
-    assert isinstance(s, GeoSeries)
-    assert isinstance(s.geometry, GeoSeries)
+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)