[Nexus] NeXus update

"V. Armando Solé" sole at esrf.fr
Wed Jan 7 09:45:05 GMT 2015


On 06/01/2015 14:30, Tobias Richter wrote:
> Dear all,
>
> For more complete information of what went on at the NIAC, Here is a 
> link to the minutes:
> http://wiki.nexusformat.org/NIAC2014_Meeting#Minutes
>
Happy New Year and thanks for the minutes.

I can easily understand Ray's disagreement:

"""
Vote on proposal for NXdata axes with indices attribute only required 
when required to resolve ambiguity. All in favour. Vote on proposal with 
indices required as in 2014_axes_and_uncertainties 
<http://wiki.nexusformat.org/2014_axes_and_uncertainties>. Accepted: 9 
in favour. Strong disagreement by Ray Osborn as adoption is noted: 
"Adoption of NeXus will be hampered by additional complexity being 
required unnecessarily". Ben Watts would like to add that the default 
plot in NXdata should be simple and not contain more data than required. 
He may make a proposal at that end in future.
"""

The indices attribute is not necessary. To get rid of ambiguities on 
multidimensional data we already had the dataset attribute 
"interpretation" (agreed upon values "scalar",  "spectrum", "image", 
"vertex" - Oak Ridge.Meeting - ) On the meeting in the vicinity of 
Diamond, I already showed to you how to get rid of plotting ambiguities 
by using that attribute.

Below you will find how that attribute could solve most, if not all, 
your cases.

If you really want to visualize multidimensional data, sooner or later 
you will end up with the generic solution: you will need to specify 
vertices and what to be visualize at those vertices (scalaras, spectra, 
images, vectors, ...). It does not help that you avoid to write it in 
the file because the visualization software *has* to generate those 
positions to visualize. I think you should take into consideration the 
advice of people actually writing data visualization software (Ray, 
myself and even Eugen was thinking about whether the generic solution 
specifying points would not be the best).

I doubt I will ever add support to the new indices stuff. I can default 
to the current situation where the user can specify what he actually 
want to visualize.

Best regards,

Armando


CASE 1: Indices not needed

NXroot
    NXentry
      NXdata
        @signal="data"  --> *names* the default data to be visualized
        @axes="x"  --> *names* the default independent data
        @x_indices=0
        data: float[100]  --> the default dependent data
        x: float[100]  --> the default independent data

CASE 2: Indices not needed, order is self-explanatory or example is broken.

NXroot
    NXentry
      NXdata
        @signal="data"
        @axes="time","pressure"
        @pressure_indices=1
        @temperature_indices=1
        @time_indices=0
        data: float[1000,20]
        pressure: float[20]
        temperature: float[20]
        time: float[1000]


CASE 3: Indices not needed. Order is enough. In case of doubt one can 
always set @interpretation"="scalar" for the signal, but it is not 
necessary.

NXroot
    NXentry
      NXdata
        @signal="det"
        @axes="pressure","tof"
        @pressure_indices=0
        @tof_indices=1
        det: float[100,100000]
        pressure: float[100]
        tof: float[100000]

CASE 4:A bit weird but can be handled with attribute interpretation set 
to "scalar" to "x" and "y" and "spectrum" to "tof" and "det".

NXroot
    NXentry
      NXdata
        @signal="det"
        @axes="x","y","tof"
        @tof_indices=2
        @x_indices=0,1
        @y_indices=0,1
        det: float[100,512,100000]
        tof: float[100000]
        x: float[100,512]
        y: float[100,512]

CASE 5: det1 must have attribute interpretation set to spectrum and time set to scalar.

NXroot
    NXentry
      NXdata
        @signal="det1"
        @axes="polar_angle_demand","frame_number","."
        @frame_number_indices=1
        @polar_angle_rbv_indices=0,1
        @time_indices=0,1
        @polar_angle_demand_indices=0
        polar_angle_rbv: [50,5]
        det1: [50,5,1024]
        polar_angle_demand: [50]
        frame_number: [5]
        time: [50,5]

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.nexusformat.org/pipermail/nexus/attachments/20150107/5bfcec45/attachment.html>


More information about the NeXus mailing list