feat: make fetch timeout for chunks configurable

[kworathur]

Summary

Addresses part of #874. The timeout for fetching chunks from the librarian is currently hardcoded. This PR makes the fetch timeout configurable through a new parameter called fetch_chunk_timeout.

• Bootstrapper initialiser timeouts - addressed by feat: make bootstrapper initialiser timeouts configurable (#874) #999
• Librarian RAG fetch timeout - addressed by this PR
• Config client timeout - TODO for a future PR
• Gateway timeout - TODO for a future PR

Changes

All changes are in the constructor of the Processor class in trustgraph-flow/trustgraph/retrieval/document_rag/rag.py. The parameter fetch_chunk_timeout is threaded through the constructor's existing params argument, which mirrors how fetch_limit and other parameters for document RAG are passed. The default value of 120 seconds has been preserved from before.

Configuring the parameter: An operator can configure fetch_chunk_timeout by setting its value in dev-tools/proc-group/groups/rag.yaml:

processors:
  - class: trustgraph.retrieval.document_rag.Processor
    params:
      <<: *defaults
      id: document-rag
      doc_limit: 20
      fetch_chunk_timeout: 180

Tests

Three unit tests have been added to tests/unit/test_retrieval/test_document_rag_service.py:

• processor picks up default of 120 seconds when fetch_chunk_timeout is not in params
• processor picks up override value when fetch_chunk_timeout is overridden.
• the configured value of fetch_chunk_timeout is passed as a parameter to librarian.fetch_document_text()

Notes

• No validation has been added for the new fetch_chunk_timeout, similar to the remaining parameters initialized in the same constructor.

[github-actions]

Contributor License Agreement ✅

All contributors have signed the CLA. Thank you!

[kworathur]

I have read the CLA Document and I hereby sign the CLA

[cybermaggedon]

Clean, small change. Two commits: tests first, then the feature — Ninja 🥷

[cybermaggedon]

@kworathur this is in draft state, is it ready to merge?

[kworathur]

@cybermaggedon all good to merge!