[NEXT] Pick a todo

[henryruhs]

Todos are ordered by priority:

• [ ] Output video is either too fast or too slow (fps mismatch)

• [ ] Migrate to tempfile using dir=temp to keep temp directory in root

• [ ] Exit threading gracefully via signal.SIGINT without tons of read/write errors

• [ ] Resolve SSL related download issue on MacOS / remove the monkey patch

• [ ] Try to get the TensorrtExecutionProvider running

• [ ] Figure out why import torch is needed before a onnxruntime usage in order to detect CUDA

• [ ] Submit a PR to CustomTkinker to support Drag and Drop - Related comment

• [x] Fix color spaces for create video

• [x] mypy

  • [x] Add a mypi.ini to ignore existing validations (certain rules per file)
  • [x] Integrate mypi to the CI pipeline

• [x] Use thread pool executor

• [x] Override fails if output file exists and output path is a directory while using deprecated -f or --face

• [x] Only one frame processor is applied

  • [x] preview
  • [x] image to image

• [x] Add NSFW filter to preview

• [x] Code of face enhancer needs cleanup as it still refers to source_path and throws error on standalone run

• [x] Override of existing output file not working with --keep-audio

• [x] Resolve multi-threading for the face enhancer

• [x] Evaluate the performance of multi-processing vs. multi-threading for CPU-only - Related comment

• [x] Fix CTkImage but <class 'PIL.ImageTk.PhotoImage'> in UI

• [x] Implementation check for frame processors that raise NotImplemented if method does not exist

• [x] Model download in our own Hugging Face space