Metadata-Version: 2.4
Name: robot-recovery-bench
Version: 0.1.0
Summary: Review metrics and schema for robot intervention and recovery segments.
Author: AuraOne
License-Expression: MIT
Project-URL: Homepage, https://auraone.ai/open
Project-URL: Source, https://github.com/auraoneai/robot-recovery-bench
Classifier: Development Status :: 3 - Alpha
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.10
Description-Content-Type: text/markdown
License-File: LICENSE
Dynamic: license-file

# robot-recovery-bench

`robot-recovery-bench` validates and summarizes human intervention and recovery segments in robot episodes. It helps robotics teams understand whether recovery data is reviewable and training-ready without sharing private videos.

This is a HIL/recovery data-quality diagnostic, not a robot safety benchmark.

## Quickstart

```bash
python -m venv .venv
. .venv/bin/activate
pip install robot-recovery-bench
robot-recovery-bench validate examples/mock_recovery_segments.jsonl
robot-recovery-bench report examples/mock_recovery_segments.jsonl --out report.md
```

## Metrics

- intervention rate;
- recovery success rate;
- time-to-intervention;
- average recovery duration;
- repeated failure clusters by reason;
- incomplete segment count.

All bundled examples are mock/tutorial data.
