[HELP REQ] NightLift (low light enhancer) port to VS

[PatchWorKs]

Hi everyone,
I've just discovered this interesting OpenCV-based Python app to treat low-light images/videos:

https://github.com/LIUZHIYON/NightLift-

It would be interesting to have all those already-available algorithms (CLAHE, Gray-World, Retinex, etc.) in a script/plugin IMHO but, since I'm a complete noob, I've asked GPT-5.6 Sol Very high to generate a port:


https://raw.githubusercontent.com/MarcoRavich/VS-NightLift/refs/heads/main/enhancer.vpy (https://raw.githubusercontent.com/MarcoRavich/VS-NightLift/refs/heads/main/enhancer.vpy)

I've not tested it (yet) directly but - as always for LLM-outputs - at glance it seems a bit over-complicated (to me) even if some parts could be used as base.

Thanks in advance to anyone that can/will help.

[PatchWorKs]

OK, I've decided to proceed step by step so I first generated (using Claude Sonnet 5 Extra) a source low-light probing script:

https://raw.githubusercontent.com/MarcoRavich/VS-NightLift/refs/heads/main/analysis.vpy

It seems to work correctly (JSON file results example (https://raw.githubusercontent.com/MarcoRavich/VS-NightLift/refs/heads/main/_tests/CONCERT_analysis.json)) but, again, any feedback/suggestion - for example about the "right" number of frame to be sampled related to the whole video length - would be great.

Thanks in advance.