Detailed documentation to follow

1. Clone the repo
2. Run uv sync (for default setup with DuckDb support) or uv sync --extra mcp-postgres depending on where your OMOP data is.
3. Copy sample.env to .env and provide a SQLAlchemy compatible connection string to your database.
4. Setup tracing with Langfuse (See https://langfuse.com/self-hosting/docker-compose#get-started for local deployment using docker compose). Follow the docs to create public and secret keys and update the .env file using sample.env as a template.
5. Modify config.toml or preferably create a copy and name it to config.local.toml. This is already gitignored. Set the CONFIG_FILE_PATH variable in .env to ./config.local.toml. This file is read by src/fastomop/config.py and any structural changes to this file that require changes to config.py must also be reflected in the default config.toml.
6. Activate the virtual environment and run uv run fastomop or just fastomop to use the CLI. API and UI versions to be implemented.