ConvertToMD fails when it should not

[Andrei Savici]

Get a workspace that you can convert to |Q|. It is important not to have UB matrix or goniometer on it. From Mantidplot, open the ConvertToMD dialog, select QDimensions=Q3D, Q3DFrames=HKL, QConversionScales=HKL (this might be what were the values from when I previously run the algorithm). No need to execute before, just change QDimensions to |Q|, and try to run it. I get

 Error in execution of algorithm ConvertToMD:
 HKL frame needs UB matrix defined on the workspace

If I choose |Q|, it should not check for UB matrix. Works If I go back to Q3D, choose Q3DFrames=AutoSelect, then go back to |Q|

[Alex Buts]

Q mode uses rotation matrix to convert to Q and transform a coordinate system (e.g HKL or A^-1) but then calculates Q². This is also necessary for adding two images with different crystal orientation in Q-mode.

I suspect the bug is in GUI which hides Q-frame and Q-scale (or Q-scale only) setting from the algorithm but can not be entirely sure.

[Nick Draper]

Moved to backlog at the end of Release 3.1

[Nick Draper]

Bulk move to assigned at the introduction of the triage step

[Alex Buts]

Convert to MD has been changed in #8804 so the block to avoid merge conflicts

[Alex Buts]

refs #8026 Renamed ConvertToMDHelper and ConvertToMDHelper2

into ConvertToMDMinMaxGlobal, and ConvertToMDMinMaxLocal

Changeset: 9cb21ba9bc5e8b619f8de8a5e4615149bd13917d

[Alex Buts]

(In #9396) This ticket renames algorithms without changes in functionality.

It is easy to test as changes are trivial.

[Alex Buts]

refs #8026 Unit test illustrating the problem

and some spelling errors.

Changeset: 84dabdd7ca2f9cde32f29fc70912f50ba356072e

[Alex Buts]

refs #8026 Partial fix which breaks number of tests

and some spelling errors.

Changeset: 9f57e008d97f9dfd45065cea67b6000c7645db52

[Alex Buts]

refs #8026 Redefined isPowder function to check transformation mode

and more spelling errors.

Changeset: fe158e3dd0e005ca4e8e7336404f700af72d9a3d

[Alex Buts]

The sequence described in the ticket should work now.

Test only after #9396 is merged into master as I've started from this ticket to avoid merge conflicts.

[Alex Buts]

refs #8026 minor comments

Changeset: c4bf985c0831f57b383595508b4357f78becd22c

[Nick Draper]

This needs a section on how to test, ideally with a sample python script

[Alex Buts]

it is the bugfix and the problem was that wrong parameters were stuck in GUI. The ultimate test is that the problem described in the ticket does not appear any more.

Tester should try the sequence described in the ticket before applying the patch to confirm that the problem was existing and then try the same sequence after applying the patch. No failure should be observed.

[Alex Buts]

first commit to the ranch ticket was actually to ticket #9396. I mistyped the ticket number but the code changes are in the basis brunch, not in this one (and are already tested/merged).

[Alex Buts]

refs #8026 Merge branch 'bugfix/8026_ConvertToMDFails' into develop

Conflicts:

Code/Mantid/Framework/MDAlgorithms/src/ConvertToMD.cpp

Changeset: 74debd3b888682bf845c601e40049cf57e5f1868

[Alex Buts]

(In #9502) Should be tested after #8026 only as the branch originates from there.

The reason for error are incorrect or incorrectly arranged bin boundaries. I believe I've fixed the logic and the problem for a histogram workspace but for event workspace it is unclear what to do so, just programming errors are fixed and correct logic remains unclear.

To test one can run test script I supplied above to reproduce problem and check that the conversion runs smoothly after this ticket is applied. One can use smaller/different file from WISH00028146 -- I verified the script reproduces error using MARI11001.raw from test folder.

The error should disappear after the ticket has been applied.

[Nick Draper]

This brach has a merge conflict on ConvertToMD.cpp.

Please resolve and resubmit.

[Martyn Gigg]

Using kdfiff the conflict here seems to be entirely because of the wiki section removal - everything else merges just fine. I'm going to pick this back up for testing and check it.

[Martyn Gigg]

Merge remote-tracking branch 'origin/bugfix/8026_ConvertToMDFails'

Conflicts:

Code/Mantid/Framework/MDAlgorithms/src/ConvertToMD.cpp

Full changeset: ​266e0563595adc4266390d520d0a688d3742ac09

[Alex Buts]

refs #8026 Merge branch 'bugfix/8026_ConvertToMDFails' into develop

Conflicts:

Code/Mantid/Framework/MDAlgorithms/src/ConvertToMD.cpp

Changeset: 74debd3b888682bf845c601e40049cf57e5f1868

[Stuart Campbell]

This ticket has been transferred to github issue ​8871