Vitallens

What's Changed

* Distinguish between `Mode.BATCH` for processing independent videos and `Mode.BURST` for processing subsequent segments of a video stream
* Add real-time webcam example script at `examples/live.py`
* Update prpy and add Dockerfile

**Full Changelog**: https://github.com/Rouast-Labs/vitallens-python/compare/v0.3.4...v0.4

What's Changed

- Updated `prpy` dependency to 0.2.12

**Full Changelog**: https://github.com/Rouast-Labs/vitallens-python/compare/v0.3.3...v0.3.4

What's Changed

- When reading videos with issues that return unexpected number of frames, automatically pad or trim the result
- If video read is not possible without missing many frames, raise error
- Support both global and non-global video parsing
- Check videos for issues before parsing
- Check if parsed videos will fit in memory if globally parsing
- Check if face movement is an issue for global parsing
- Choose whether to parse globally based on the above factors

**Full Changelog**: https://github.com/Rouast-Labs/vitallens-python/compare/v0.3...v0.3.3

What's Changed

- Added option to generate running vitals from predicted waveforms - e.g., `running_heart_rate` which is the heart rate computed from the predicted PPG waveform using sliding windows. Useful for longer videos.
- Added JSON export which writes all estimation results to JSON.
- Lighter build by not including sample data.
- Some internal improvements.

**Full Changelog**: https://github.com/Rouast-Labs/vitallens-python/compare/v0.2.2...v0.3