[GH-3196] GeoPandas: Add distributed sampling, point construction, and serialization

[jiayuasu]

Did you read the Contributor Guide?

• Yes, I have read the Contributor Rules and Contributor Development Guide

Is this PR related to a ticket?

• Yes, and the PR name follows the format [GH-XXX] my subject. Closes GeoPandas: add distributed sampling, point construction, and geometry serialization #3196

What changes were proposed in this PR?

• Add distributed GeoSeries.sample_points support for uniform sampling. Polygon and line sampling use native Sedona and Spark expressions without collecting geometry rows to the driver or using Python UDFs.
• Add top-level points_from_xy and GeoSeries.from_xy APIs backed by native ST_Point and ST_PointZ expressions, including scalar broadcasting, distributed index alignment, CRS metadata, and native SRID propagation.
• Add GeoDataFrame.to_wkt and GeoDataFrame.to_wkb, serializing every geometry-typed column in one lazy projection while preserving non-geometry columns and the existing index.
• Add direct and GeoPandas-parity coverage for scalar and distributed inputs, empty and null geometries, multipart lines, deterministic random seeds, MultiIndex inputs, duplicate labels, CRS, and SRID behavior.
• Document the new APIs in the English and Chinese GeoPandas tutorials. The WKT documentation also records that native Sedona serialization preserves stored coordinate precision rather than applying Shapely's default six-decimal rounding.

How was this patch tested?

• Ran the focused GeoPandas construction, sampling, and serialization test suite: 14 passed.
• Ran all pre-commit hooks against the changed files.
• Verified Python 3.8 and Python 3.9 syntax compatibility.
• Built the Spark shaded module with Spark 3.4, Scala 2.12, and GeoTools enabled.

Did this PR include necessary documentation updates?

• Yes, I am adding new APIs. I am using the current SNAPSHOT version number, v1.9.1.
• Yes, I have updated the documentation.