Celux

- **Pre-Release Update:**
- (somewhat) Fixed `VideoWriter` class. Working on `cuda` for now, but `cpu` still has incorrect output.
- Added `VideoWriter`, and `LogLevel` definitions to `.pyi` stub file.
- Adjusted github actions to publish to `pypi`.

- **Pre-Release Update:**
- Added logging utility for debugging purposes.
py
import celux
celux.set_log_level(celux.LogLevel.debug)

- **Pre-Release Update:**
- Added `buffer_size` and `stream` arguments.
- Choose Pre-Decoded Frame buffer size, and pass your own cuda stream.
- Some random cleanup and small refactorings.

- **Pre-Release Update:**
- Adjusted Frame Range End in `VideoReader` to be exclusive to match `cv2` behavior.
- Removed unnecessary error throws.
- **Encoder Functionality:** Now fully operational for both CPU and CUDA.

- **Pre-Release Update:**
- Renamed from `ffmpy` to `CeLux`.
- Created official `pypi` release.
- Refactored to split `cpu` and `cuda` backends.

- **Pre-Release Update:**
- Removed `Numpy` support in favor of `PyTorch` tensors with GPU/CPU support.
- Added `NV12ToBGR`, `BGRToNV12`, and `NV12ToNV12` conversion modules.
- Fixed several minor issues.
- Updated documentation and examples.