LineCuts units

[Owen Arnold]

In most cases, MD line cuts should have axes and units reported w.r.t their origin workspace. I generated a work around for this issue here #8104, but there should be a good way to access this information on the generated MDWorkspaces.

The following script illustrates the problem. The generated plot lines should completely overlap.

CreateMDWorkspace(Dimensions='2',EventType='MDEvent',Extents='-10,10,-10,10',Names='Q_lab_x,Q_lab_y',Units='A,B',OutputWorkspace='test')
FakeMDEventData(InputWorkspace='test',UniformParams='1000',PeakParams='100000,0,0,1')

line_md = BinMD(InputWorkspace='test',AxisAligned='0',BasisVector0='X,units,0.888154,0.459545',BasisVector1='Y,units,-0.459545,0.888154',Translation='-4.93562,-2.3918',OutputExtents='0,9.61632,-0.1,0.1',OutputBins='100,1')
save_path = "/Users/spu92482/Desktop/line_md.nxs"
SaveMD(Filename=save_path, InputWorkspace=line_md)
reloaded_line_md = LoadMD(Filename=save_path)
plotMD([line_md, reloaded_line_md]) # Should be the same!!!!

[Nick Draper]

Bulk move of tickets out of triage (new) to assigned at the introduction of the triage state

[Owen Arnold]

After some initial investigation, there seems to be two problems here:

1) The is no original workspace if reloaded, so picking up the transformations from the original workspace, (even though they do exist) is skipped.

2) The MD nexus format does not save multiple transformations (treat under a separate ticket).

[Owen Arnold]

Tester:

Run the script in the description, and check that the plots line-up.

Then run this at the end of the script above and check that the outputs match the comments.

print "These should be different"
print line_md.numOriginalWorkspaces(), reloaded_line_md.numOriginalWorkspaces()
print "But these should be the same"
print line_md.getNumberTransformsToOriginal(), reloaded_line_md.getNumberTransformsToOriginal()

[Owen Arnold]

refs #8992. Utilise coord trans to original.

This has been prevented thus far, because use of coord transforms has been based on the presence of original workspaces. Original workspaces will not be present upon a reload.

I've also exposed the new getters on MDGeometry to python.

Changeset: e4253729ec04765d7930be6b5fbe90a72824cb80

[Owen Arnold]

The next steps here are to merge in master after #9009 is complete, and change ConvertToMDHistoWorkspace so that uses the new getters added to MDGeometry rather than relying on the number of workspaces to be set.

[Owen Arnold]

This is a core change owing to the fact that MDGeometry has been extended.

[Owen Arnold]

refs #8992. Use number of transformations to original.

ConvertMDHistoToMatrixWorkspace should not count original workspaces. I've also refactored this code so that I'm not having to work with raw pointers.

Changeset: b65e6140d90ae3a9e40e3250717f235fe216c778

[Owen Arnold]

Replying to Owen Arnold:

The next steps here are to merge in master after #9009 is complete, and change ConvertToMDHistoWorkspace so that uses the new getters added to MDGeometry rather than relying on the number of workspaces to be set.

Done

[Owen Arnold]

Pending checkbuild develop branch is broken.

[Karl Palmen]

Merge remote-tracking branch 'origin/feature/8992_coord_trans_loading'

Full changeset: ​62d7cf70e9160c90a29f7eac36b259cde0f71420

[Owen Arnold]

I think this is possibly a patch candidate, and I'm going to mark it up as such.

[Nick Draper]

Moved to patch release 3.1.1

[Owen Arnold]

refs #8992. Problem fixed.

Although, I'd like to do some more refactoring to neaten things up before hand.

Changeset: e6789be8c3fa014b13bf43f5c321dadcf34fd4f7

[Russell Taylor]

What's the story with the commit in comment:20? The ticket has been tested and merged to master, but now this new commit has popped up on a branch with the same name as the merged one (but not the same branch).

[Owen Arnold]

Replying to Russell Taylor:

What's the story with the commit in comment:20? The ticket has been tested and merged to master, but now this new commit has popped up on a branch with the same name as the merged one (but not the same branch).

My mistake. Entered the wrong ticket number. That commit should have gone against #8376. This commit has nothing to do with this ticket.

[Owen Arnold]

refs #8992. Utilise coord trans to original.

This has been prevented thus far, because use of coord transforms has been based on the presence of original workspaces. Original workspaces will not be present upon a reload.

I've also exposed the new getters on MDGeometry to python.

Changeset: 892641698fbb0348afe3df85ab766af2a399ca00

[Owen Arnold]

refs #8992. Use number of transformations to original.

ConvertMDHistoToMatrixWorkspace should not count original workspaces. I've also refactored this code so that I'm not having to work with raw pointers.

Changeset: 7e538b154c98d583e3e8899adc66299e5659d79f

[Stuart Campbell]

This ticket has been transferred to github issue ​9835