getPhi for a DetectorGroup is not right in some circumstances

[Martyn Gigg]

DetectorGroup::getPhi() currently computes the individual phi values for each member and averages these. This does not give the same value as if you get the position of the group first and then compute the phi value like this

import math
def calculatePhi(detector)
    pos_xyz = detector.getPos()
    return math.atan2(pos_xyz[1],pos_xyz[0]) # radians

To test, use a workspace that has been grouped and then in python do

det = workspace.getDetector(index) # index of group
wrong_phi = det.getPhi()
phi = calculatePhi(det)
# The value will not match

[Martyn Gigg]

In [15160]:

Fixes #3923. Fix the phi averaging for DetectorGroups that was wrong in certain situations like on an edge.

[Martyn Gigg]

In [15166]:

Refs #3923. New reference file for a minor tweak to the cylinder absorption that is small and understood but above current test tolerance.

[Russell Taylor]

Looks good.

[Stuart Campbell]

This ticket has been transferred to github issue ​4770