feat(python): add a Ray Data VortexDatasink - #11478

Summary

• Tracking Issue: [Python] Add a Ray Data VortexDatasink #11414

Ray Data could read Vortex files but had no distributed write path. This adds VortexDatasink, so ds.write_datasink(VortexDatasink(path)) streams each write task to one Vortex file instead of collecting data on the driver or switching formats.

Changes

• Converts pandas and Arrow blocks to a streaming RecordBatchReader and writes through vx.io.write.
• Reuses Ray's file-sink path handling, save modes, unique task filenames, and local-filesystem scheduling; includes the Ray 2.48 filename-hook fallback.
• Adds round-trip, null, append, empty-input, and S3 URL tests plus user documentation.

Verified: 701 passed, 2 skipped for the Python suite; Ray 2.48 and 2.59 datasource suites passed; 317 Sphinx doctests passed. Ruff and ty passed.

API Changes

Adds vortex.ray.datasink.VortexDatasink.

AI assistance

Written with agentic AI assistance; the implementation, tests, and compatibility behavior were reviewed and executed locally.