Export of Signal/nEvents/Error array to python looks incorrect

[Alex Buts]

e.g. mixes array dimensions.

If I try to extract slice (see attached picture 1) from Mantid to Python using script below, the row and column dimensions of this array look mixed (see attached picture 2)

from numpy import *

def write_slice(eMin,eMax):	
	EnRange='DeltaE,'+str(eMin)+','+str(eMax)+',1'
	if eMin<0:
		eTag = 'm'+str(abs(eMin));
	else:
		eTag = str(eMin)
	slice_name = 'sliceE'+eTag
	hw=BinMD(InputWorkspace='mdws4D',AlignedDim0='[H,0,0],-1,5,121',AlignedDim1='[0,K,0],-3,5,161',AlignedDim2='[0,0,L],-0.2,0.2,1',AlignedDim3=EnRange,OutputWorkspace=slice_name)
        signal=hw.getSignalArray();
	nEvents=hw.getNumEventsArray();

	f=open(slice_name+'.txt','w')
	ranges = signal.shape;
	print ranges;

	for j in range(0,ranges[1]):
		for i in range(0,ranges[0]):			
			if (nEvents[i,j,0,0]>0):
				rez = signal[i,j,0,0]/nEvents[i,j,0,0]
				f.write('%10g' % rez);
			else:
				rez= 0
				f.write('%10g' % rez);			
		f.write('\n')		
	f.close();		

write_slice(30,40)

[Alex Buts]

fig1

[Alex Buts]

fig2

[Alex Buts]

refs #5986 unit tests for suspicious behavior.

and some code rearrangement.

Changeset: 177a3c94902090861ea125163507665489acd28b

[Alex Buts]

refs #5986 inverted dimensions in numpy converter

to make Histo-workspace accessible from Numpy in a C-way.

Changeset: 42ea268b0b472f7c3149bd49fc2cc468c11189f5

[Alex Buts]

refs #5986 some tests to changed python import

not all yet run correctly

Changeset: 04e5fed762656fb72e5f610dce4a571a3da2eb14

[Alex Buts]

refs #5986 Modified FakeMDEventData to generate desired data kind

which needed for testing python classes and heterogeneous data arrays

Changeset: 0f0ecc82e16cfe8a15a1180976cf0362558d40f2

[Alex Buts]

refs #5986 more tests (failing)

Changeset: d057149f47643f339716fb1b6e12d18e83e6b26d

[Alex Buts]

refs #5986 temporary disabling number of failing tests

(to be able to commit on another ticket)

Changeset: c48b82a50896cf9b32b46635ca3b3806251e2257

[Alex Buts]

refs #5986 fixing Unix warnings

Changeset: 8ed3744dbdf2be2b7cccddf6c4d4c872b5a8f6b7

[Alex Buts]

refs #5986 Added getNEvents to IMDWorkspace.h

Changeset: d48d70d13373e0f587cce914cc4db4236b409b5b

[Alex Buts]

refs #5986 test for BinMD fixed

and python uncommented

Changeset: 836e57db4c448a73209b63c5f66d9be6b2604685

[Alex Buts]

refs #5986 Rumbling around

Changeset: 98db3508a5117ea7a5bf6fbb064c1f62c1c2d064

[Alex Buts]

refs #5986 Tests for C++ and MDHistoWorkspaceTest.py

seems all work as expected

Changeset: 25460d7b73b88a14ed061e037a491af280e8e6cd

[Alex Buts]

Internal structure of MDHisto data array looks rather like fortran array If it right or wrong -- is different question.

Meanwhile, export behaves correctly

[Andrei Savici]

The file I save, when read in IDL, looks like my original data

[Alex Buts]

refs #5986 unit tests for suspicious behavior.

and some code rearrangement.

Changeset: 177a3c94902090861ea125163507665489acd28b

[Alex Buts]

refs #5986 inverted dimensions in numpy converter

to make Histo-workspace accessible from Numpy in a C-way.

Changeset: 42ea268b0b472f7c3149bd49fc2cc468c11189f5

[Alex Buts]

refs #5986 some tests to changed python import

not all yet run correctly

Changeset: 04e5fed762656fb72e5f610dce4a571a3da2eb14

[Alex Buts]

refs #5986 Modified FakeMDEventData to generate desired data kind

which needed for testing python classes and heterogeneous data arrays

Changeset: 0f0ecc82e16cfe8a15a1180976cf0362558d40f2

[Alex Buts]

refs #5986 more tests (failing)

Changeset: d057149f47643f339716fb1b6e12d18e83e6b26d

[Alex Buts]

refs #5986 temporary disabling number of failing tests

(to be able to commit on another ticket)

Changeset: c48b82a50896cf9b32b46635ca3b3806251e2257

[Alex Buts]

refs #5986 fixing Unix warnings

Changeset: 8ed3744dbdf2be2b7cccddf6c4d4c872b5a8f6b7

[Alex Buts]

refs #5986 Added getNEvents to IMDWorkspace.h

Changeset: d48d70d13373e0f587cce914cc4db4236b409b5b

[Alex Buts]

refs #5986 test for BinMD fixed

and python uncommented

Changeset: 836e57db4c448a73209b63c5f66d9be6b2604685

[Alex Buts]

refs #5986 Rumbling around

Changeset: 98db3508a5117ea7a5bf6fbb064c1f62c1c2d064

[Alex Buts]

refs #5986 Tests for C++ and MDHistoWorkspaceTest.py

seems all work as expected

Changeset: 25460d7b73b88a14ed061e037a491af280e8e6cd

[Stuart Campbell]

This ticket has been transferred to github issue ​6832