change the NIKA option to read the nika attributes directly

[EliotGann]

on a loadrun, the beamcenter x and y and detector distance are read once into the beamcenter, but if you change these attributes before runbning the integration, they are not reread, and have to be entered manually ie
geomethod='nika',
NIdistance=raw.attrs['sdd'],
NIbcx=raw.attrs['beamcenter_x'],
NIbcy=raw.attrs['beamcenter_y'],
I think it makes more sense, when using the nika geomethod, to reread the attrs each time, incase the user has tweaked them, that is unless there is an easier (and faster?) way to interactively tweak the beamcenter in a integrator.

[pbeaucage]

I had a brainwave on this. I should define NIbcx and NIbcy as properties of the integrator - so that they can be tweaked just as integrator.NIbcx = 123.45; integrator.NIbcy = 234.56

[pbeaucage]

I'm of two minds on actually defining where in the attribute tree a beam center should live - there are so many units, centers, etc. The 'right way' (TM) would be for a loader to convert the instrument's units to PyFAI poni1/poni2/dist (all in m), which iirc is what it does when you initialize with template_xr.

If we then have the ability to tweak in either Nika or pyFAI space with the underlying pyFAI data store updating automatically, would that basically fulfill this feature?

[EliotGann]

Of course having a way to visualize and drag around a beamcenter as a cursor would be the nicest way to solve all of this (so I don't have to do it in igor, and pull the new coordinates over)

[EliotGann]

but yeah, I think having a way to tweak an integrator on the fly either way would solve the problem. I should store the pyFAI center in the metadata perhaps...

[pbeaucage]

Preliminary fix for this in [2e589de], did not tag properly in commit message.

[pbeaucage]

Folks brought up on Slack today that this is really useful, but completely undocumented. Needs a tutorial update to show tweaking the beam center on-the-fly.