Theta calculation bug

[Owen Arnold]

ReflectometryReductionOne and ReflectometryReductionOneAuto output incorrect theta values.

The following script shows this problem in action:

(q,l,th)=ReflectometryReductionOneAuto('INTER00022699',OutputWorkspace='vsQ1',OutputWorkspaceWavelength='vsLam1')
print "incorrect theta reported as ", th

ws = mtd['INTER00022699']
inst = ws.getInstrument()
det = inst.getComponentByName('point-detector').getPos()
sample = inst.getComponentByName('some-surface-holder').getPos()

detsample  = det - sample

refFrame = inst.getReferenceFrame()
upoffset = refFrame.vecPointingUp().scalar_prod(detsample)
beamoffset = refFrame.vecPointingAlongBeam().scalar_prod(detsample)

import math
print "correct theta is:", math.atan(upoffset/beamoffset)/2 * 180 / math.pi

[Owen Arnold]

This is actually blocked by #9389, because in order to fix this in a general way, we need to be able to determine an averaged vertical detector location as well as the sample location. The code to do this is already in SpecularReflectionCorrectPositions (9389).

• A first step to getting this done will be to extract out an algorithm base class called SpecularReflectionAlgorithm from SpecularReflectionCorrectPositions. This will provide two protected method and will otherwise virtual. Two methods are getDetectorComponent() and getSampleComponent()
• Next step is to use the base class to create a new algorithm called SpecularReflectionCorrectTheta, which will use this base class.
• Next step is to wire this algorithm into ReflectometryReductionOne to replace to existing correction logic.

[Owen Arnold]

refs #9401. Create and Use base algorithm class

Changeset: 90f91bed9f6ae71c5460413e0de6015ac70b9e14

[Owen Arnold]

refs #9401. Add new algorithm.

Create a testing base class and move properties over to algorithm base class.

Changeset: 794bd10310192e0e494435a1a177720aa7268708

[Owen Arnold]

refs #9401. More shared tests.

Changeset: e835115b72826be03732327ec473750e122aa5f9

[Owen Arnold]

refs #9401. Implement theta calculation.

Changeset: 1c2558f7d544cf310407d07c604dfaef19f1a742

[Owen Arnold]

refs #9401. Documentation and logging.

Changeset: fb7976ce23be1507d2fde8b4085bb4c480ea8f28

[Owen Arnold]

refs #9401. Rename algorithm to better reflect job.

Changeset: 1eb228c7d1e92e2680f4302807c755f10a692d6c

[Owen Arnold]

refs #9401. Fixed in ReflectometryReductionOne.

Changeset: 57ac60c55017df965c0ed28442d1c67986789527

[Owen Arnold]

Tester: This issue has now been fixed. I have reused the code in SpecularReflectionPositionCorrect via an algorithm base class to achieve this. Unit tests have been added and adapted to handle this error.

Rerun the original python script (ask me for INTER data set if required). To see this fixed.

[Anders Markvardsen]

Merge remote-tracking branch 'origin/feature/9401_correct_positions'

Full changeset: ​48abd57ae32196550e2cb1aa06e28a38de801727

[Anders Markvardsen]

Unit test added to handle error. Output from running script is:

incorrect theta reported as  0.709694197533
correct theta is: 0.709694197533

[Stuart Campbell]

This ticket has been transferred to github issue ​10244