[GH-3182] GeoPandas: Implement distributed sjoin_nearest

[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: implement distributed sjoin_nearest #3182

What changes were proposed in this PR?

This PR implements distributed GeoPandas-compatible sjoin_nearest for the top-level API and GeoDataFrame.

The implementation uses Sedona's distributed KNN join rather than a Cartesian product, Python geometry UDF, or driver-side geometry collection. It supports inner, left, and right joins; equidistant matches; max_distance; distance_col; exclusive; suffixing; named and MultiIndex inputs; null/empty geometries; active geometry; CRS; and SRID preservation.

The SQL KNN predicate gains five- and six-argument forms for query-local include_ties and exclusive controls. The regular planar path now routes non-point queries by their full geometry and reconciles partition-local candidates with bounded per-query state and stable row identities. Legacy three/four-argument and geography tie behavior remain unchanged.

Candidate selection without ties is deterministic across regular and both broadcast plans. max_distance is currently an output post-filter and does not prune the distributed KNN search.

English and Chinese SQL and GeoPandas documentation are updated.

How was this patch tested?

• Direct KNN iterator and execution-option tests (12 tests)
• KnnJoinSuite (41 tests)
• GeoPandas nearest-join and Python ST_KNN plan/API/type tests
• Spark 3.4 / Scala 2.12 shaded package build
• Spotless and repository pre-commit hooks

The tests force regular and both broadcast KNN plans and cover non-point queries, duplicate rows, ties, exclusive ranking, outer reconstruction, indexes, column collisions, null/empty geometries, CRS/SRID metadata, and legacy/geography compatibility.

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.