# HDD environment: constructed model sensitivity

These are ten constructed histories, not measured disk failures. Interpretation:
https://hesela.com/analyses/smart-failure-signals/#environment-heading

Keep all six files in one directory: model.py, test_model.py, results.json,
sources.json, table.md and this README.md. Python 3.10 or newer; standard library only.

```sh
python3 -B model.py
python3 -B test_model.py
```

The verifier checks the existing capture, source hashes and generated table.
Do not run --write to verify the supplied capture: it replaces results and their
generation metadata. No original fleet data or third-party paper is included.
