Custom GetEi algorithm for CNCS

[Stuart Campbell]

from Georg Ehlers:

I have played during the last few days a little bit with CNCS data looking
for a way to determine Ei from the data reliably to cross-check the entry
"energyrequest" stored in the nxs files. I have found that you need to use
the elastic line in the detectors and the spectrum of "monitor3". This is
the one mounted onto the backside of the high speed double disk chopper
(facing the sample). Even though the two locations are fairly close to
each other at the end of the spectrometer, one can seemingly determine Ei
accurately from the difference in tof. So, I am OK with you implementing
this for the general one-solve-all python reduction script for CNCS. I
suggest you restrict the detector counts to equator-near pixels (e.g.,
pixels 54-73) for banks (modules) 1-30. I suggest you read the module
distances from the sample (these locations change frequently) and the
monitor location from the nxs file. Further I do suggest that you don't
fit the peaks, rather use the center of mass. One can also get T0
accurately enough from this.

Do not use monitor2. I found that the spectrum it sees is "too white" and
the center of the peak is not close enough to the tof of the desired
energy at this location.

[Nick Draper]

Moved at end of release 2.1

[Andrei Savici]

Get Ei from a monitor and detectors. Refs #5105

Changeset: 07740d10fbcd79174fb5e30b67e62c79497b3227

[Andrei Savici]

Plan: rewrite in C++

[Andrei Savici]

Get Ei from a monitor and detectors. Refs #5105

Changeset: 07740d10fbcd79174fb5e30b67e62c79497b3227

[Andrei Savici]

The C++ version and test will be part of ticket #5686

[Michael Reuter]

This is working OK, but the C++ version should verify correctness. We're also OK for now since this is not is production code.

[Stuart Campbell]

This ticket has been transferred to github issue ​5951