👷 Run Qdrant in CI for integration tests

[ebigunso]

Motivation and Context

• Integration tests need to run against a live vector database in CI to validate Qdrant-backed functionality end-to-end.
• This adds a GitHub Actions workflow that provisions a Qdrant service container during CI so tests exercise real DB behavior instead of mocks.
• Fixes: N/A

Description

High-level overview:

• Adds .github/workflows/ci.yml.
• Starts a qdrant/qdrant:v1 service container, exposes port 6333, and blocks until the /readyz endpoint is healthy before running tests.
• Sets environment variables consumed by Settings::load():

  QDRANT_CONNECTION_STRING: ${{ vars.QDRANT_CONNECTION_STRING }}
  OXIGRAPH_CONNECTION_STRING: ${{ vars.OXIGRAPH_CONNECTION_STRING }}
  OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
  EMBEDDING_MODEL: ${{ vars.EMBEDDING_MODEL }}
  

• Installs the Rust toolchain, and runs cargo check and cargo test and cargo clippy.

If helpful:

• CI validates integration with the Qdrant-backed VectorMemoryRepository and the OpenAI embedding path when OPENAI_API_KEY is available.

Checklist

• Code builds with no errors or warnings
  cargo check --all-targets
• Unit tests pass
  cargo test --verbose
• Clippy passes with warnings denied
  cargo clippy --all-targets -- -D warnings
• cargo fmt has been run
• Documentation updated where needed
• BREAKING CHANGE? No
• I didn’t break anyone 😊

Additional Notes (optional)

• Tests generate unique collection names and clean them up; a single ephemeral Qdrant instance is safe even with parallel tests.
• Ensure OPENAI_API_KEY is configured in repository secrets so CI can exercise OpenAIEmbeddingRepository. For forked PRs without secrets, consider a follow-up to provide a mock embedding repository under a CI feature flag.