Serial crystallography binning

[noemifrisina]

Add a cli for ssx fixed target collections

[noemifrisina]

As things stand, the current code will correctly bin the first image but return empty ones after that. The issue seems to come from the use of valid_events with arrays instead of a single value for start ans stop.

Going to try and find a solution.

[graeme-winter]

Having a very quick look at the output, rather than the change set. @noemifrisina you were concerned that the same frame was simply repeated 400 times: this is not teh case

Derived from:

import numpy
import hdf5plugin
import h5py
f = h5py.File(filename, "r")
d = f["/entry/data/data"]
for j in range(400):
  print(j, numpy.sum(d[j,:,:]))

[graeme-winter]

Looking at the very constant spot count per image output I am going to guess there are some unmasked bad pixels here.

[noemifrisina]

Thanks for checking.
About the almost constant spot count: unmasked pixels are entirely possible. But as this is a corner block, it might also be that there was nothing in most windows.
Btw, right now there's no mask in that dataset (as it's from last year, the file needs to be copied in the same directory as the binned images for it to work - newer collections won't have this issue).

[graeme-winter]

OK, lack of mask would probably do it, thank you

[keeble]

I've not tested it (It looked as though there's already testing afoot) so this is just a drive-by code review.

Generally looks pretty good to me, a couple of small inline comments added which may just be my misunderstanding.

Only overall thought (which applies to existing code, not just this PR) would be that a lot of the logic here sits within the cli function itself, meaning it's not necessarily that easy for someone else to use the code from elsewhere. I guess the purpose of this package is to provide cli only? Probably a separate PR in any case :-)