Context Navigation

← Previous Ticket
Next Ticket →

Ticket #10944 (closed: fixed)

Opened 6 years ago

Last modified 5 years ago

Plotting Bug from CutMD

Reported by:	Owen Arnold	Owned by:	Owen Arnold
Priority:	blocker	Milestone:	Release 3.4
Component:	Framework	Keywords:
Cc:		Blocked By:
Blocking:		Tester:	Harry Jeffery

Description

Here's the issue as reported by Russell

#Set things up
hkl = LoadMD(Filename='C:\Russell\Software\Mantid\VATES\VATESdemoSSC15\Templates\FeDataset.nxs' ,FileBackEnd=1)
 
#Create projections (equivalent of proj in Horace):
projection = CreateEmptyTableWorkspace()
projection.addColumn("double", "u")
projection.addColumn("double", "v")
projection.addColumn("double", "w")
projection.addColumn("double", "offsets")
projection.addColumn("str", "type")
projection.addRow([1, 1, 0, 0, "r"])
projection.addRow([1, -1, 0, 0, "r"]) 
projection.addRow([0, 0, 1, 0, "r"]) 
 
 
#QE slice and look at it in the sliceviewer
horace_style = CutMD(hkl,projection, [-3,0.05,3], [0.9,1.1], [-0.1,0.1], [0,4,300], NoPix=True, CheckAxes=False)
sv=plotSlice(horace_style)
sv.setNormalization(2)#set norm to number of events
sv.setXYDim(0,3)#Choose the correct plot axes (0th and 3rd in Python!)
sv.setColorScaleMin(0)#set colour scale limits
sv.setColorScaleMax(1)
sv.setXYLimits(-2,2,0,200)
 
#If you try to make a 1d cut along (1,1,0), integrating in energy, direct in line mode in the sliceviewer, something weird happens to the integration box – it becomes a trapezium… Nevertheless, if you manage to make the 1d cut this way, and then plotMD, the result has the correct x-axis of (H,H,0).
 
#1d Q cut direct
horace_1d=CutMD(hkl,projection, [-2,0.05,2], [0.9,1.1], [-0.1,0.1], [33,40], NoPix=True, CheckAxes=True)
pm=plotMD(horace_1d,error_bars=1,normalization=mantid.api.MDNormalization.NumEventsNormalization)
 
#The plotMD here has axes of (H,0,0), with the “correct” (H,H,0) not available as an option.

Change History

comment:1 Changed 6 years ago by Nick Draper

Status changed from new to assigned

comment:2 Changed 6 years ago by Owen Arnold

Status changed from assigned to inprogress

refs #10944. Enhance QueryMDWorkspace.

To help diagnose the problem it's useful to look at coordinates of boxes without the transform applied according to any original workspace.

Changeset: 462a7ef0a0ee9d7ae398cbbed126f45c1b0d26f5

comment:3 Changed 6 years ago by Owen Arnold

refs #10944. Allow MDGeometry to clear transforms and workspaces.

The main change here is to what CutMD can do. It Clears BOTH the transforms and the original workspaces off the output workspace.

1) This fixes the issue with the labeling when plotMD is run, because there is no original workspace to try to take dimension information from.

2) However, the plots generated from CutMD from the SliceViewer are now garbage. This could because the line cuts (BinMD) are operating over a Histogram workspace already (CutMD created with NoPix=True), but is something I will need to get to the root of.

Changeset: 48f6daaf0af7f0eeb89b1b7fc83cd5df39928dc8