Question/idea: Could a notebook-model supplant bespoke photographing processing software such as the "darkroom" mode of Lightroom (or darktable). The extant programs essentially take a lot of data (camera's raw output) and apply a configurable recipe to produce intelligible output (an image). Each recipe (stored as an XMP sidecar) is essentially a list of math operations (increase brightness, wavelet decompose, change color model, etc.) and their parameters.

Obviously a great part of why we use Lightroom/darktable is because of the speed with which the recipe-processing occurs. Plus a smooth UI, a catalog-viewing feature, and a well vetted choice of image operations. The appeal of moving this work to a notebook would be that an actively maintained Jupyter ecosystem could supplant lock-in to a specific software, and open up the underlying math magic.

At the very least, this could be an interesting platform for experimenting with image processing methods. And the reordering of cells could become a virtue, to run an image processing pipeline out of the standard order.

I'm curious if anyone has already worked along these lines. I find through a quick web search that people are doing some image processing, but more in the face detection or ML for medical imaging aspects. I see as a basic toolkit that http://scikit-image.org/docs/dev/auto_examples/ is something, though this isn't the whole range of operations needed for, say, fine art image tuning.