Reuse master dark, bias & flat when it was already stacked

[TomyCesaille]

Problem

As of now, each light set-sequence will run a full src/pipeline/Mono_Preprocessing/Mono_Preprocessing.ssf scripting.

Biases are re-computed each time, when they are fundamentally the same over and other. High chances that dark and flat are being re-stacked as well.

This is quite concerning for projects with 10+ hours around LRGBSHO layers. The computation time and space on disk can greatly be reduced.

Side note

As of now, we rely on a full .ssf script that is isolated for each light set-sequence. Each {{masters}}/bias_stacked are unique paths.

Solution

Solution A - general approach

We could change {{masters}}/bias_stacked (where {{master}} is light-centric, like [...]/poto__barnard-347/O/Light_60.0s_Bin1_O_gain100__20240626-232133_masters/) to {{biases}}/masters and have Siril skip re-stacking if file is already generated (TBD: Siril would already skip on its own? conditions in script?)

Solution B - optimization made from users

Have poto preprocess stack-biases.ssf and then poto preprocess Mono_Preprocessing.ssf designed to read bias_stacked on a predetermined location.

Solution C - read stacked files from bank

Could be a generic command to build stacks in the bank file directly for Poto-Siril to read it back.
The major concerns with this approach is that it kills A/B testing (#9) design by considering bank stacks as static.