`SubtitlesWriter` long-term vision?

Hi. I had the need to get a transcript of some audio clips, and whisperX did a decent-enough job. However, i have some thoughts/mixed feelings about the actual subtitle writing:

Bugs:

• line_len does not consider joining whitespaces, and thus there is no way to enforce Characters Per Line. Is this the intended behaviour?
• There is a hardcoded (3s) duration which hard-splits segments, which is a not-unreasonable default. Should it be configurable? (Also, any reason why why > and not >=?) EDIT: Oooh, but that's not what the current code is doing! It counts from the beginning of the previous word to the beginning of the new word...

Features:

• There is no way to configure maximal duration of a single segment.
• It might be good to be able to configure what the shortest segment duration must be. (Would conflict with other rules, i guess this would probably be solved by extending the segment duration)

I'm also wondering if there should be an official script that would do JSON->subtitles processing.

Thank you!