Summary

This PR implements developer onboarding improvements requested in issue #43. It adds Docker Compose files for local development, onboarding documentation, example scripts demonstrating typical workflows (quickstart / RAG / hybrid search), and small developer convenience Makefile targets. These changes are intended to make it easier for new contributors to set up the environment and get started.

Closes

Closes #43

What changed

Added/new files:

• docker-compose.yml
• docker-compose.seekdb.yml
• CONTRIBUTING.md
• examples/README.md
• examples/quickstart.py
• examples/rag_demo.py
• examples/hybrid_search_demo.py
• README.md (updated quickstart + onboarding content)

Updated:

• Makefile (appended onboarding/dev targets: docker-up, docker-down, docker-logs, docker-up-seek, docker-down-seek, docker-logs-seek, fmt, lint, typecheck, test)

Why

• Provide reproducible local dev environment using OceanBase (and a lightweight SeekDB alternative).
• Document development setup and PR/test/lint expectations for new contributors.
• Provide runnable examples to demonstrate the library's intended usage patterns and to make manual testing easier.

How to test locally

1. Checkout branch:
   git checkout -b feat/onboarding-improvements-43

2. Bring up the database:
   make docker-up

   or for the lightweight alternative:

   make docker-up-seek

3. Create a .env file (or export env vars):
   OB_HOST=127.0.0.1
   OB_PORT=3306
   OB_USER=root
   OB_PASSWORD=
   OB_DB=langchain_ob_demo
   OPENAI_API_KEY=

4. Install example dependencies:
   pip install openai mysql-connector-python numpy

5. Run examples:
   python examples/quickstart.py
   python examples/rag_demo.py
   python examples/hybrid_search_demo.py

6. Run linters & tests:
   make fmt
   make lint
   make typecheck
   pytest -q

Notes

• The example scripts use OpenAI for embeddings/LLM; running them requires an API key. They are intentionally minimal and store embeddings as JSON in the DB and perform client-side ranking (portable approach).
• The Makefile changes were appended to avoid losing existing targets. If your repo's Makefile already has targets with the same names, please review and merge appropriately.
• The docker-compose healthchecks assume the images contain common utilities (nc/curl). If not, adjust healthchecks per image.
• If the repo has CI that enforces dependency files (pyproject.toml / requirements.txt), consider adding example dependencies or mark examples as optional.