Matched cuts shift plot
Created by: RosalynLP
This supersedes #314, and has separated out the actions for making the shift plots, which should be nearly ready to merge. The more recent testing actions remain on the branch covmat_cuts2
.
To do: @Zaharid's previous comments:
So there are several things I don't like about this. I'll list a several but not all
- The tests do not pass for obvious reasons.
- The process_lookup thing shouldn't be a provider but rather a normal function that gets called inside other providers. It should be dataset -> lookup string. As it is, it introduces some ugly dependencies.
- New stuff does not have docstrings in general.
- The previous interface was better in that it ensured that the dataspecs come from matched cuts and so the things had always the right dimensions. We should see how to have something similar here.