Skip to content

Worker plan database

Subclass of the worker_plan service that runs the PlanExe pipeline with a Postgres database.

  • Polls TaskItem rows, marks them processing, and runs the pipeline.
  • Reports state/progress back to the DB and posts confirmations to MachAI.
  • Uses the same worker_plan_internal code as worker_plan, plus the shared database_api models.
  • Configure MachAI confirmation endpoints with PLANEXE_IFRAME_GENERATOR_CONFIRMATION_PRODUCTION_URL and PLANEXE_IFRAME_GENERATOR_CONFIRMATION_DEVELOPMENT_URL (both are required; the worker fails fast if missing).

Docker usage

  • Build/run single worker: docker compose up --build worker_plan_database
  • Run three workers (each with PLANEXE_WORKER_ID=1/2/3): docker compose up -d worker_plan_database_1 worker_plan_database_2 worker_plan_database_3
  • Reads SQLALCHEMY_DATABASE_URI when provided, otherwise builds one from:
  • PLANEXE_POSTGRES_HOST|PORT|DB|USER|PASSWORD
  • falls back to the database_postgres service defaults (planexe/planexe on port 5432)
  • Logs stream to stdout with 12-factor style logging. Configure with PLANEXE_LOG_LEVEL (defaults to INFO).
  • Volumes mounted in compose: ./run (pipeline output), .env, llm_config.json
  • Entrypoint: python -m worker_plan_database.app

Run locally with a venv

For a faster edit/run loop without Docker. Work from inside worker_plan_database so its dependencies stay isolated:

cd worker_plan_database
python3.13 -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install -e ../worker_plan
pip install -r requirements.txt
export PYTHONPATH=$PWD/..:$PYTHONPATH
python -m worker_plan_database.app

Run deactivate when you are done with the venv.

The PYTHONPATH addition allows imports of database_api and worker_plan_database modules. The pyrightconfig.json and .vscode/settings.json configure the same paths for editor/IDE support. In Cursor/VS Code, select the interpreter from .venv/bin/python via Cmd+Shift+P → "Python: Select Interpreter".