[GH-2800] Add ST_SharedPaths across engines and GeoPandas

[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 #2800.
Part of #2230.

What changes were proposed in this PR?

This PR adds a PostGIS-compatible ST_SharedPaths implementation for lineal geometries.

• Implements the shared-path algorithm in the Java common module using robust overlay and spatially indexed source-segment lookup.
• Returns a GeometryCollection containing:
  • same-direction shared paths as the first MultiLineString;
  • opposite-direction shared paths as the second MultiLineString.
• Matches PostGIS behavior for direction bucketing, nulls, typed empty geometries, non-lineal input validation, mixed SRIDs, and dimensional handling.
  • A null operand returns null.
  • Typed empty inputs or inputs without a shared path return two empty MultiLineString components.
  • A non-empty result retains Z only when at least one input has Z and every returned path coordinate resolves to a finite source Z. Otherwise, the whole result is two-dimensional. M is always dropped.
  • For non-simple inputs that traverse a path more than once, direction follows the first matching source traversal, so swapping operands can change bucket membership.
• Exposes ST_SharedPaths through:
  • Spark SQL and Spark Connect registration;
  • the Scala DataFrame API;
  • Python sedona.spark.sql.st_functions;
  • Flink SQL;
  • Snowflake legacy WKB and native GEOMETRY APIs.
• Prevents generation of a Snowflake GEOGRAPHY overload because this planar lineal operation is supported only for GEOMETRY. The native GeoJSON bridge's SRID limitation is documented.
• Adds distributed GeoSeries.shared_paths, including scalar-geometry handling, positional and index-based alignment, duplicate MultiIndex support, CRS warnings, and GeoDataFrame delegation.
• Adds Spark, Flink, and Snowflake API documentation, summary-table entries, and three detailed SVG examples covering direction buckets, multipart noding, and point-only intersections.

How was this patch tested?

• Passed all 1,271 tests in the common module, including 20 dedicated SharedPathsTest cases.
• Passed the Spark SQL suite: 236 tests.
• Passed the Spark DataFrame API suite: 260 tests.
• Passed targeted Flink and Snowflake DDL-generation tests.
• Passed focused GeoPandas regression tests on GeoPandas 1.0.1 and 1.1.3.
• Passed targeted Python DataFrame API tests.
• Verified direction and ordinate behavior across 520 deterministic lineal reference cases, including the large-coordinate near-parallel regression.
• Passed Spotless, Markdown lint, MkDocs static build, repository hooks, and git diff --check.

The credentialed Snowflake integration suites were not run locally.

Did this PR include necessary documentation updates?

• Yes, I am adding a new API and used the current SNAPSHOT version, v2.0.0.
• Yes, I updated the Spark, Flink, and Snowflake documentation and function summary tables.