[Nexus] NXsas with multi-frame image set

Benjamin Watts benjamin.watts at psi.ch
Mon Jun 27 10:03:23 BST 2016


Hi Pete,
    I would suggest your first choice unless you know that it will mess 
up the analysis pipeline for most users. A 3D array is much tidier and 
easier to write code for, though you may need to set some standard axes 
labels if people are going to start adding dimensions for other 
variables (e.g. photon energy, X-ray polarisation). So I think the 1st 
choice should be obvious unless you have the situation where a large 
fraction of the users use a particular piece of analysis software that 
will choke on the 3D array (and it can't be easily fixed), but one of 
the other choices for storing frames can be handled by this analysis 
software.

Cheers,
Ben

On 24/06/16 19:20, Pete R Jemian wrote:
>
> I want to store a set of images according to the NXsas application 
> definition.  In the current case, these images form a time series. 
> NXsas defines /entry/instrument/detector/data as a 2-D image frame.
>
> What's my storage strategy?  Any of these or something different?
>
> first choice: 3-D frame set
>     /entry/instrument/detector/data: (frames, rows, columns)
>
> second choice: single frames in same NXdetector
>     /entry/instrument/detector/data_frame_1: (rows, columns)
>     ...
>     /entry/instrument/detector/data_frame_n: (rows, columns)
>
> third choice: one group for each frame
>     /entry_frame_1/instrument/detector/data: (rows, columns)
>     ...
>     /entry_frame_n/instrument/detector/data: (rows, columns)
>     and link other objects (NXsample, ...) from /entry_frame_1/*
>
> Pete
>
> _______________________________________________
> NeXus mailing list
> NeXus at nexusformat.org
> http://lists.nexusformat.org/mailman/listinfo/nexus



More information about the NeXus mailing list