Metadata-Version: 2.5
Name: roboscope
Version: 0.0.1a0
Summary: Small utilities for embodied policy evaluation — early development release
Classifier: Development Status :: 2 - Pre-Alpha
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.10
Description-Content-Type: text/markdown

# RoboScope

Small, dependency-free utilities for embodied policy evaluation.

## Status

This is an early development release with a deliberately small API. The project
name and API are provisional. This release provides episode success statistics;
it does not include model inference, benchmark runners, or simulator integrations.

This is an independent project, unaffiliated with MIT Media Lab's RoboScope,
vexide's RoboScope simulator, or viadee's RoboScope test management platform.

## Install

```sh
pip install --pre roboscope
```

## Summarize episode outcomes

```python
from roboscope import summarize_successes

summary = summarize_successes([True, False, True, True])
print(summary.episodes)      # 4
print(summary.successes)     # 3
print(summary.failures)      # 1
print(summary.success_rate)  # 0.75
```

The function consumes any iterable of booleans in one pass. Non-boolean values
raise `TypeError`, avoiding accidental truthiness of strings such as `"false"`.
For zero episodes, `success_rate` is `None`, distinguishing missing observations
from a measured zero success rate.
