Feat/reproducibility

@karnem, @amunnich, @samartse, @qingtian, and FYI: @fangohr

This adds basic consistency testing for the correction notebooks of AGIPD and LPD, with reports also building into the documentation, see e.g:

https://in.xfel.eu/readthedocs/docs/european-xfel-offline-calibration/en/feat-reproducibility/test_results.html

and

https://in.xfel.eu/readthedocs/docs/european-xfel-offline-calibration/en/feat-reproducibility/testing.html

for a summary of how things work.

Note that in the documentation example, I've included quite a few failure cases, which are not part of this commit into the artefacts directory, these are removed for the latest commit.

Next steps after this is review and merged:

• add tests for Jungfrau and FastCCD correction notebooks
• add tests for characterisation notebooks. We should likely think about mock calibration database handling there, which resolves requests to always the same data, and provides verification for submissions to it.

Setting to WIP until the two update MRs are merged