| # Quick Recommendations for Working with the Repo | |
| | Goal | Suggested Starting Point | | |
| |------|--------------------------| | |
| | **Run a simple Elizabeth chat** | `cd elizabeth/e-1-first_session && python elizabeth_chat` (or `elizabeth_full.py`). | | |
| | **Inspect memory calls** | Open `elizabeth_memory_integration.py` and follow calls to bloom_memory_api modules in `bloom-memory/`. | | |
| | **Run the full autonomous stack** | `cd mlops && python deploy_autonomous.py` (ensure required env vars for DBs and vLLM are set). | | |
| | **Track an experiment** | After running, open `mlflow.db` via the MLflow UI (`mlflow ui --backend-store-uri sqlite:///mlflow.db`). | | |
| | **Add a new tool for the LLM** | 1. Add a JSON entry in `mlops/agents/tool_registry.json`. 2. Implement the function in `mlops/elizabeth_mlops_tools.py`. 3. Update `elizabeth_tool_demo.py` to call it. | | |
| | **Scale memory services** | Look at `bloom-memory/deployment/` scripts (`deploy.sh`, `DEPLOYMENT_GUIDE_212_NOVAS.md`) to launch on a Kubernetes‑like environment. | | |