Hello @dynamite-ready;

The input_video_paths argument should be a list of strings, not a string itself. If you refactor the call you have above to the following:

scenedetect.video_splitter.split_video_mkvmerge([str(video_path)], scene_list, "test", "tests", suppress_output=False)

That should resolve the error that you're seeing. The intention of the split_video_ functions is to allow multiple videos as input files, which will be concatenated together in the output.

Please let me know if passing a list of file paths (instead of a single path) works as intended so that I can resolve the issue.

Lol! Yes. That worked fine.

I'd perhaps argue that it could be made a little clearer in the docs - https://pyscenedetect-manual.readthedocs.io/en/latest/api/video_splitter.html - but your explanation helped.

Thank you.

@dynamite-ready very good point! I've modified the title to better reflect that and reopened the issue to take care of that. I'll try to include the proper documentation for those arguments in the next release, thanks for pointing that out.

Ping: Revisit this next week for v0.6. Top priorities right now are resolving ffmpeg split issues and camera flash issue resolution, but I don't want to lose track of quality-of-life items like this one.

For some reason, the documentation that now gets generated is completely blank for the video splitter. I've submitted a support request for ReadTheDocs (readthedocs/readthedocs.org#7245), but in the meantime I'll try to come up with a workaround.

Either way, I'll get this closed out this week as well, but just wanted to keep you posted.