[JUNGFRAU] [Correct] Allow selecting ROIs to save 1D projections

Description

FXE project spectra onto JUNGFRAU, so analysis requires selecting the relevant region and taking a 1D projection. They have asked for this to be integrated into calibration, to speed up later processing.

Closes calibration/planning#136

To be determined:

• Input parameters format?
• Output format - including source name (currently like FXE_XAD_JF500K/ROIPROC/JNGFR03:output)
• Should mask be applied before doing the reduction?
• sum, mean (or nanmean)? Or make it another parameter?
• Are 2 ROIs enough?

And probably more, this is just off the top of my head.

How Has This Been Tested?

Running the notebook manually.

Relevant Documents (optional)

Types of changes

• New feature (non-breaking change which adds functionality)

Checklist:

• My code follows the code style of this project.

Reviewers

@schmidtp @ahmedk

changed milestone to %3.6.2

added 1 commit

• 5b4c46ea - Allow for arbitrary number of ROIs

Compare with previous version

marked the checklist item Are 2 ROIs enough? as completed

marked the checklist item Input parameters format? as completed

Yohei was suggesting up to 10 ROIs, which is far more than I had thought, so I've made it accept an arbitrary number of ROI definitions, although this means the config for it is decidedly cryptic - a list of integers which will be interpreted as groups of 6 - module_no, a1, a2, b1, b2, sum_axis.

I've also made it so the output will have ROIs numbered within each module - so if you define 2 ROIs on 2 different modules, they will both have roi1 in the key (but different source names, e.g. FXE_XAD_JF500K/ROIPROC/JNGFR03:output vs .../JNGFR02:output).

marked this merge request as ready

The test failure is already fixed in master (!706 (merged)).

added 2 commits

• 6d3922d6 - Apply mask before reducing ROI data
• 8106c8f0 - No ROIs defined by default

Compare with previous version

added 1 commit

• 4225c85a - Mean selected values rather than summing

Compare with previous version

Thank you for this nice MR, I have added some comments/questions. And I see that there are some unchecked points in the MR description are they still undecided? Who should answer these questions? Us or FXE?

marked the checklist item Should mask be applied before doing the reduction? as completed

marked the checklist item sum, mean (or nanmean)? Or make it another parameter? as completed

added 1 commit

• af71c501 - Add example in parameter comment

Compare with previous version

added 1 commit

• 2c61cefe - Add check for ROI reduction axis

Compare with previous version

resolved all threads

LGTM, thank you