Mantid reading multiperiod nexus and multiperiod data with SANS GUI

[Nick Draper]

We discovered a number a errors last night on SANS2D.

Reading a multiperiod nexus file results in an incorrect grouping of the workspaces. (raw file reading is fine) This is and example

loadnexus(".....SANS2D00005508.nxs","w5508")

the file contains 13 periods, results in

	w5508 
	   |-w5508_2 
	   |-w5508_3 
	   |-w5508_4 
	   |-w5508_5 
	   |-w5508_6 
	   |-w5508_7 
	   |-w5508_8 
	   |-w5508_9 
	   |-w5508_10 
	   |-w5508_11 
	   |-w5508_12 
	   |-w5508_13 

	w5508_1

More importantly for SANS2D the SANS gui cannot deal with multi-period data. (this is undoubtedly because it has not been tested as no files have existed)

There are two problems. On windows the files are to big for a 4GB system and everything falls over -?? On the Linux box it is happier but produces log file reading faults. - ??

SANS2d runs 5511 to 5519 should contain multi period data

[Roman Tolchenov]

(In [4708]) WorkspaceGroup returned by LoadISISNexus confused MantidPlot because it didn't contain the groups own name. Monitors were read for all periods but only the first of them was used. re #1285

[Steve Williams]

(In [4736]) SANS and LOQ scripts can now deal with multi-period input files, just append the required period to the end of the AssignSample, AssignCan, Assign... commands. This process is slow as these commands load the whole file before removing the unwanted spectra. The required periods are removed from the group and appended with _px where x is the period number, processing after that proceeds as for ungrouped workspaces. re #1285

[Steve Williams]

(In [4757]) Getting the SANS GUI to deal with multi-period workspaces in Single file mode and a bug so that the monitor defaults to 2 again re #1285

[Steve Williams]

(In [4782]) The SANS GUI can now process a sample workspace contained in a multi-period workspace. Extensions to this will follow re #1285

[Steve Williams]

(In [4816]) The SANS GUI was writing a non-keyword argument after keyword argument error, this stops that re #1285

[Martyn Gigg]

(In [4819]) Alleviate problems with WorkspaceGroup crashes. The Fit browser was causing a crash while cheking a group member that was being deleted. Problem not solved, WorkspaceGroups just temporarily disabled within the Fit browser. Re #1285

[Martyn Gigg]

(In [4825]) Fix for SANS centre-finder crashes along with a problem with the residue graph if it was closed prematurely. Re #1285

[Steve Williams]

Centre finder was giving me an error

[Steve Williams]

(In [4832]) In the centre finder the sample workspace was being deleted before it was needed by the can processing re #1285

[Steve Williams]

My change only changed centre finding Python code. Perhaps just check that.

[Stuart Campbell]

This ticket has been transferred to github issue ​2132