I did once implement a Filmulation module in darktable but it really runs against the way darktable’s pipeline runs.
The Filmulation step must run using the entire image together in order to accurately reflect the final result, whereas darktable’s modules try to constrain the calculated input region of interest when you zoom in to minimize computation.
Filmulator’s pipeline instead always calculates the full image extents and caches heavily to maintain interactivity.
So in the end it really kneecaps the performance and doesn’t fit with the way the way you work in dt. On top of that, while I remember some people liking the output, at least back then I couldn’t get it to look right (or anything like the Filmulator default output) even with the Filmulation module.