Bug with negative xstep in IDF

[Owen Arnold]

Message from Dennis below:

"NEW BUG: There was a side-effect to the new SXD definition file. If a new calibration file is not loaded, when an SXD peaks workspace is written to an ISAW peaks file, any detector where xstep is negative in the definition file gets written with a NEGATIVE WIDTH. Similarly, if ystep is negative in the definition file, the detector gets a NEGATIVE HEIGHT. We could fix this symptom by making SaveIsawPeaks() just write the absolute value of the width and height to the file, but fixing the symptom may hide worse problems that will be hard to find later. That is, however the instrument definition information gets placed in the rectangular detector information, it is now getting set with a negative width and/or height for some of the SXD detectors. Perhaps someone who is more familiar with the system than I could check this out and fix it in a way that should not break other things. The negative width and height "broke" the detector calibration in Mantid, so I just used an editor to make them positive and used that fixed peaks file to do the calibration in ISAW."

The offending IDF is attached.

[Owen Arnold]

refs #5667. Add two failing test cases

Changeset: 419fc2f041434ca51bc2b709623c3d75a7f93101

[Owen Arnold]

refs #5667. Fix failing test by fixing code

Changeset: eb72adae009e466193761d4de093fda913712449

[Owen Arnold]

Tester:

Run the following and notice that the Height and Width entries in the output *.peaks file are not negative for all 11 Banks.

LoadRaw(Filename=r'SXD23767.raw',OutputWorkspace='SXD23767',LoadLogFiles='0',LoadMonitors='Exclude')

LoadInstrument(Workspace='SXD23767',Filename=r'SXD_Definition4.xml')

ConvertToDiffractionMDWorkspace(InputWorkspace='SXD23767',
                                OneEventPerBin='1',LorentzCorrection='1',
                                SplitInto='2,2,2',SplitThreshold='25',MaxRecursionDepth='12',Extents='-12,12,-12,12,-12,12',
                                OutputWorkspace='SXD23767_MD')

FindPeaksMD(InputWorkspace='SXD23767_MD',
            PeakDistanceThreshold='0.5',MaxPeaks='325',DensityThresholdFactor='50',
            OutputWorkspace='SXD_peaks_0pt3_300_100')
            
SaveIsawPeaks(InputWorkspace='SXD_peaks_0pt3_300_100', Filename="demo.peaks")

[Owen Arnold]

I'm going to undo these changes and move the ticket to the next release because the issue is more complicated that we thought at first.

[Owen Arnold]

refs #5667. Revert changes.

Changeset: 2397d0d869b65b7bf2092c5811c7d6e2f56510d6

[Owen Arnold]

refs #5667. Revert changes.

Changeset: 2397d0d869b65b7bf2092c5811c7d6e2f56510d6

[Owen Arnold]

refs #5667. Add two failing test cases

Changeset: 419fc2f041434ca51bc2b709623c3d75a7f93101

[Owen Arnold]

refs #5667. Fix failing test by fixing code

Changeset: eb72adae009e466193761d4de093fda913712449

[Owen Arnold]

refs #5667. Revert changes.

Changeset: 2397d0d869b65b7bf2092c5811c7d6e2f56510d6

[Nick Draper]

Moved to milestone 2.4

[Dennis Mikkelson]

Work on this ticket could possibly break lots of things for lots of instruments. It should be addressed by someone with a thorough understanding of how detector geometry is treated in Mantid.

[Nick Draper]

Moved at the code freeze for release 2.4

[Nick Draper]

Moved to r2.6 at the end of r2.5

[Karl Palmen]

I had a similar problem when running my tube calibration scripts on MAPS for tubes that had a 'step' of -1 and subsequently fixed it.

[Nick Draper]

Moved to backlog at the code freeze for R2.6

[Nick Draper]

Poker Estimate: 6.5d

[Nick Draper]

Bulk move to assigned at the introduction of the triage step

[Stuart Campbell]

This ticket has been transferred to github issue ​6513