NeXus/HDF5 data input allowing selection of data paths

[toqduj]

I'm seeing a lot of cases where SasView doesn't read NeXus files, likely due to inconsistencies between the various files. As a back-up plan, would it be an idea to allow the user to configure which data paths of the input file map to (1D or 2D) Q, I, ISigma (and maybe QSigma)? This dialog could come up if the standard HDF load methods fail..

[smk78]

Do you mean NeXus or NXcanSAS?

[toqduj]

NXcanSAS specifically. I've got someone's GRASP-processed datafiles which are not being read by SasView. Ideally, I would just point SasView at the locations where qx, qy, I ISigma and the mask are stored, instead of writing an HDF5-to-HDF5 converter.

[smk78]

So the likely possibilities here are that i) GRASP is not writing compliant NXcanSAS, or ii) something we did recently has broken NXcanSAS loading for some files (though I just tried generating and loading some 1D and 2D NXcanSAS in Mantid and it reads into SasView fine).

Can you upload an example file? Is it 1D or 2D reduced?

[butlerpd]

I would be very surprised if GRASP is outputting NXcanSAS. I strongly suspect it is putting out standard NeXus with some "extensions" as defined by Charles Dewhurst. ILL converted to NeXus slightly before NXcanSAS was approved I am guessing?We should ask @gonzalezma to look into that.

[butlerpd]

or perhaps better yet Lionel Porcar?

[gonzalezma]

I'm out of the ILL, so Lionel could check this. But I think that Grasp just writes its own hdf format (not sure how much NeXus compliant), so I'm not terribly surprised that it's not read by SasView.

…

On 2020-07-27 18:49, Paul Butler wrote: or perhaps better yet Lionel Porcar? -- You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub [1], or unsubscribe [2]. Links: ------ [1] #1639 (comment) [2] https://github.com/notifications/unsubscribe-auth/ACTXM5QVCCG4XJP7NJRYSELR5WV25ANCNFSM4PF4Z5GA

-- Dr. Miguel A. Gonzalez Institut Laue Langevin (ILL) 71 avenue des Martyrs 38042 Grenoble Cedex 9, FRANCE Tel: +33 (0)4.76.20.71.66 Fax: +33 (0)4.76.48.39.06 E-mail: gonzalezm@ill.fr

[toqduj]

well, I suspect that with every flavor of instrument, there will be yet another slight variant on how the HDF5 files are written in the end. Having the (backup) option of setting the HDF5 data paths to the matrices would future-proof this a bit...