chore(geopandas): Add more edge cases to the `test_match_geopandas_series.py` suite

`test_match_geopandas_series.py` is where we compare that our results match the results of geopandas. We reuse the same lists of geometries (e.g `self.linestrings`, `self.geoms`, etc) in nearly every test. I wrote these cases very early and failed to consider various edge cases. In this issue, I'd like to add some test cases to these lists (e.g `self.linestrings`). I wouldn't be surprised if we fail some of these.

Here are the edge cases I think we should add. I think each of these should be their own PR, since it's possible that they cause test failures and require us to modify the implementation of existing sedona geopandas functions.
- empty geometries (e.g `POINT EMPTY`)
- non-valid polygons (e.g `POLYGON((0 0, 0 1, 2 1, 2 2, 1 2, 1 0, 0 0))`, see [here](https://postgis.net/workshops/postgis-intro/validity.html) to learn about validity)
- GeometryCollections cases (apparently, there's only [one case](https://github.com/apache/sedona/blob/6dc39e753b9d2087d27d2a6072b95d3eda91c53f/python/tests/geopandas/test_match_geopandas_series.py#L92) at the moment...)
- geom with only empty geometries
- geom collection w/ only a single geom (e.g is_closed only applies to linestrings, should it return true for this GeomCol (< a single closed linestring >)?
- Z and M dimensions for everything
- CRS
- Other geometry types that Sedona doesn't support (e.g [LinearRing](https://shapely.readthedocs.io/en/2.1.1/reference/shapely.LinearRing.html#shapely.LinearRing), [box](https://shapely.readthedocs.io/en/2.1.1/manual.html#shapely.geometry.box))

Here's an example of the current cases (and hence where we would need to add more cases).
https://github.com/apache/sedona/blob/6dc39e753b9d2087d27d2a6072b95d3eda91c53f/python/tests/geopandas/test_match_geopandas_series.py#L59-L67

Start with python/tests/geopandas/test_match_geopandas_series.py, especially the shared geometry lists around lines 59-67 and the existing GeometryCollection case near line 92. Run this suite first, then choose one listed edge-case group for its own change. Done means the relevant cases are covered and the comparison tests pass, with implementation changes only if the new cases expose failures.

Written by the indexing model from the issue text.