Make sure DAVE MSlice can read NXSPE files with NaNs for masked detectors

[Stuart Campbell]

To make sure that if we replace the -1e30 for masked detectors in NXSPE files with a NaN.

Alex: Could you please create a test file with NaNs in it so Andrei can test DAVE or let him know when you change the algorithm.

[Alex Buts]

(In [12267]) refs #3127 A Test file nxspe file with NaN-s instead of -1e+30 committed

[Alex Buts]

(In [12273]) refs #3127 A Test file nxspe file for MARI with NaN-s instead of -1e+30 committed -- proven to work with mslice (MAPS is unclear at the moment)

[Andrei Savici]

By default the nxspe reader transforms -1e30 to NaNs. Will need to check that if people add multiple files together in DAVE, the total masks are correct. I will do this when the SaveNXSPE algorithm is updated.

[Alex Buts]

(In [12321]) refs #2745 final tests and conclusion; it should be it if everything works fine on real systems algorithm to calculate azimuthal width in cylindrical system has changed, which may be wrong for square detectors (and there should have 365% anyway) but may be reasonable for maps. refs #3127 SaveNXSPE modified to write NaN instead of -1e+30; Questionable but can be easy reverted

[Alex Buts]

(In [12347]) refs #2745 the discrepancy between ISIS and SNS par files in nxspe has been fixed. refs #3127 erroneous nxspe test file with correct NaNs but wrong angular values deleted.

[Nick Draper]

Accepted and assigned tickets moved at iteration 29 code freeze

[Andrei Savici]

DAVE fixed, starting with development version 2011-07-05.

; $Id: dm_load_nxspe.pro,v 1.8 2011/07/05 15:27:16 ymqiu Exp $

; HISTORY:
; 04/2010 Added flag to keep track of ki/kf correction
; 07/2011 Add finite value check for the mask
;

line 257 of dm_load_nxspe.pro

 mask = where(((data[*,0] le masklimit) or (finite(data[*,0]) eq 0)),mskcount,complement=tmpnomask)

[Stuart Campbell]

This ticket has been transferred to github issue ​3974