feat: add async and batch retrieval to VortexRAGRetriever

[vaishnavidesai09]

Extends the LangChain integration from #4 with async and batch support.

What's added

• aget_relevant_documents() / ainvoke() — async retrieval via ThreadPoolExecutor
• batch(queries) — concurrent sync batch retrieval
• abatch(queries) — concurrent async batch via asyncio.gather
• max_workers param for thread-pool tuning
• _executor initialised in model_post_init for Pydantic v2 compatibility

Tests

10 new tests in tests/test_async_batch_retriever.py. All 23 pass.

Closes #4

[vaishnavidesai09]

Hi @vignesh2027! Thanks for the detailed setup instructions and the reference implementation — it made it easy to understand the design.

I've extended the integration with async and batch support as you suggested:

• aget_relevant_documents() / ainvoke() — non-blocking async retrieval via ThreadPoolExecutor
• batch(queries) — concurrent sync batch retrieval
• abatch(queries) — concurrent async batch via asyncio.gather
• max_workers param for thread-pool tuning
• Fixed a Pydantic v2 compatibility issue with _executor in model_post_init

All 23 tests pass. Happy to make any changes based on your feedback!