Verification

Practical checks when authoring or shipping open-fdd rules.

---

1. Unit tests (CI)

From the repo root with dev dependencies:

pip install -e ".[dev]"
pytest

---

2. Rule YAML sanity

• Load each file with load_rule() or point RuleRunner(rules_path=...) at the directory.
• Confirm every inputs key has a matching entry in column_map (or manifest) before calling run().

---

3. Small DataFrame smoke test

from pathlib import Path
import pandas as pd
from open_fdd.engine.runner import RuleRunner

df = pd.read_csv("your_sample.csv", parse_dates=["timestamp"]).set_index("timestamp")
runner = RuleRunner(rules_path=Path("path/to/rules"))
out = runner.run(df, timestamp_col="timestamp", column_map={"Supply_Air_Temperature_Sensor": "SAT"})
assert out.filter(like="_flag").shape[1] >= 1

---

Platform / database checks

For HTTP APIs, databases, Brick / 223P graph workflows, and observability around a deployed stack, see open-fdd-afdd-stack documentation.