Open-FDD

This repository ships the open-fdd rules engine: YAML-defined fault detection on pandas DataFrames (open_fdd.engine). The published PyPI wheel contains only engine and schema modules.

Operator dashboards, HTTP bridges, ingest drivers, and deployment stacks are not bundled. Describe what you need in openfdd.toml, then use skills/ and the local agent shell to generate code under workspace/.

Install from PyPI

pip install "open-fdd[engine]"

Bare import with pandas only: pip install open-fdd (add [engine] for YAML rules and RuleRunner).

Rule authoring: Expression rule cookbook.

Build with skills + agent shell

1. Copy openfdd.toml.example to openfdd.toml and set [build] targets, drivers, auth, and deploy mode.
2. Install the shell (local only, not on the engine wheel):

cd packages/openfdd-agent-shell
pip install -e ".[dev]"

3. From the repo root:

openfdd-agent-shell --repo-root .

The shell loads AGENTS.md, selected skill recipes under skills/, and launches Codex CLI to scaffold only what the manifest requests. Generated apps live in workspace/.

Online documentation

• Open FDD fault detection engine — RuleRunner, YAML rules, pandas workflows. Documentation · GitHub · PyPI

Historical desktop/MCP how-tos under docs/howto/ describe the retired monolith; new integrations should follow skills/.

Develop and test

git clone https://github.com/bbartling/open-fdd.git
cd open-fdd
python3 -m venv .venv
source .venv/bin/activate   # Windows: .venv\Scripts\activate
pip install -U pip
pip install -e ".[dev]"
pytest open_fdd/tests/engine

Optional shim package:

cd packages/openfdd-engine && pip install -e .

Dependencies

• Python 3.10+ and pip — required: pandas; rule execution adds PyYAML and pydantic via the [engine] extra (NumPy via pandas).
• Codex CLI on PATH when using the agent shell (codex by default).
• Node.js only if a generated dashboard skill scaffolds a Vite/React app under workspace/.

License

MIT