Separated Disordered Tracing into own module

[Max-Gamill]

This is the first part of the mega PR to refactor the 800-better-tracing branch.

This PR in particular does:

• Creates a smoke test for me to check the refactored workflow works (once it's finished)
• Adds to one of the tests to allow the comparison between np.nan and np.nan which previously failed
• Moves smoothing, skeletonisation and pruning workflow into it's own module. This works on the image but processes each grain separately, producing a dictionary of the grain's crop data (skeletonisation stages and bounding boxes). The individual crops are then compiled into a "full_image" to plot as the processing / diagnostic images. This differs from previously where the trace coordinates were stored / passed into dnatracing (this step will move into dnatracing as nodestats prefers skeletons)
• Moves the smoothing, skeletonisation and pruning parameters into it's own section in the config (and accommodates this in validation.py
• Adds the "disordered_trace" step in process_scan.
• Works along-side the current dnatracing pipeline and is NOT connected to anything on the output.

Some docstrings may not be right but I've ran (some) pre-commit checks so they at-least should be present (but I was going to check these together after the pipeline works as i/o's etc may change)

*when making changes to this, and other stacked branches, will need to rebase <branch_tip> --update-refs and that should update 800-better-traces-refactor too

[Max-Gamill]

nodestats tracing .topostats test fails in the comparison because previously the keys in the .topostats file for the new changes aren't saved out yet as the pipeline is not complete.

[Max-Gamill]

@llwiggins has moved out all the tests associated code from this PR into it's own branch (llwiggins/800-better-tracing-876_tests) and PR (#880) so this one can solely focus on the disordered traces step and I have reverted the commits associated with that.

[Max-Gamill]

@ns-rse Thank you for amending the pre-commit complaints, some of these I don't think can be helped and when I tried to add the ignoring statements it didn't like them but that's probably a user error :)

The general exception was from older code and might be useful to look into when writing the tests?

[Max-Gamill]

@ns-rse regarding the git-blame fixes, I've read your blog and it seems my commits were not atomic enough - the commit Moved disordered traces out from dnatracing includes: • The disordered tracing functions moved out from dnatracing into disordered_tracing.py • The updates to the config + validation • My own functions to run the disordered tracing pipeline With my reason being to ensure the pipeline worked so I wasn't making a load of tiny fix commits

So my question is, if I do put this commit in the .git-blame-ignore-revs file, how will it handle the new lines of code in the disordered_tracing.py file that wasn't there before? As it's a new filename, git wont associate those moved functions with their counterparts in the dnatracing file so there will be no blame references right, does this still make it worth doing?

[Max-Gamill]

This has been added in #889 (sorry)