feat: add asynchronous support to LLMRanker with run_async method

[GovindhKishore]

Related Issues

• fixes feat: Add asynchronous support (run_async) to LLMRanker feat: Add asynchronous support (run_async) to LLMRanker #11840

Proposed Changes:

Added native asynchronous support (run_async) to LLMRanker. This allows reranking pipelines to run concurrently inside asynchronous environments (like FastMCP or FastAPI) without stalling the main event loop during LLM network requests.

How it works:

• Runtime check for async support: Verifies whether the assigned chat generator natively supports async execution via hasattr(self._chat_generator, "run_async"). If it only supports synchronous execution, it automatically drops down to a safe thread-pool fallback via asyncio.to_thread instead of blocking the event loop.
• Identical logic to run: Deduplication, empty-query/empty-document guards, and the ranking/parsing post-processing (_get_reply_text, _rank_documents_from_reply) are unchanged and shared between the sync and async paths - only the chat generation call itself differs.
• Monitored failure tracking: Strictly matches the existing sync error-handling configuration (raise_on_failure), falling back to the original document order when ranking or generation fails and the flag is False, or re-raising when it's True.

How did you test it?

Added a complete async test suite (test_run_async_*) mirroring the existing sync test coverage in test_llm_ranker.py. The tests cover:

• run_async version for existing run sync tests: Implemented the corresponding test_run_async versions for test_run ones.
• Dual-path execution: Confirming run_async is used when the chat generator supports it, and that it is not used (with asyncio.to_thread calling sync run instead) when the generator lacks run_async.
• Live integration tests: Added async equivalents of the existing live integration tests against the real OpenAI API, confirming end-to-end ranking behavior.

Ran hatch run test:unit test/components/rankers/test_llm_ranker.py with a 100% clean pass, hatch run ruff format / hatch run ruff check --fix.

Notes for the reviewer

The implementation structure mirrors the synchronous run method identically to keep the component maintenance straightforward.

Checklist

• I have read the contributors guidelines and the code of conduct.
• I have updated the related issue with new insights and changes.
• I have added unit tests and updated the docstrings.
• I've used one of the conventional commit types for my PR title: feat: add run_async support to LLMRanker
• I have documented my code.
• I have added a release note file, following the contributors guidelines.
• I have run pre-commit hooks and fixed any issue.

[vercel]

@GovindhKishore is attempting to deploy a commit to the deepset Team on Vercel.

A member of the Team first needs to authorize it.