At present, only the first sequence file in a dark run for each module is used to create darks for LPD-1M. This appears to be typically 256 trains. The notebook skips the first 10 trains by default, leaving 246. Hazem tells me this is not enough for good quality constants.
This adapts the notebook to use EXtra-data and select a given number of trains from the dark runs, regardless of how many files they are spread over.
I'd suggest reviewing the two commits separately:
I removed some unnecessary complexity for looping over capacitor settings, even though the notebook only allowed for a single setting. This is not strictly necessary, but it makes the code easier to follow.
The second commit introduces EXtra-data.
How Has This Been Tested?
Ran the notebook interactively with data from p3338, for a single module, to check that it works and the plots look similar to ones in an injected report for the same dark runs.
Types of changes
New feature (non-breaking change which adds functionality)
