Improve support for short data type in xml/hdf5

[tischi]

See discussion here: https://imagesc.zulipchat.com/#narrow/stream/327326-BigDataViewer/topic/XML.2FHDF5.20specification

Here is a dataset that was stored in python as unsigned short (Corrected). The data was rescaled such that the min and max values should be 0 and 65535. https://oc.embl.de/index.php/s/BvxrscO9bxPf4No

[mkitti]

tracking

[mkitti]

Do we want a signed or unsigned short, int16 or uint16?

[constantinpape]

Here is a dataset that was stored in python as short.

The dataset you have provided here as stored as unsigned short. (I think this is actually what you want, but it's confusing that you write short).

Here's the details for one of the datasets:

/t00000/s00/0/cells      Dataset {750/750, 551/551, 699/699}
    Location:  1:3896
    Links:     1
    Chunks:    {64, 64, 64} 524288 bytes
    Storage:   577723500 logical bytes, 372274970 allocated bytes, 155.19% utilization
    Filter-0:  deflate-1 OPT {4}
    Type:      native unsigned short

[mkitti]

Have you tried to load this via n5-viewer?

https://www.youtube.com/watch?v=nYzYiXi0sYs&t=165s

[tischi]

The dataset you have provided here as stored as unsigned short. (I think this is actually what you want, but it's confusing that you write short).

Sorry, yes, I corrected the description.

[tpietzsch]

157 addresses this issue. Could you have a look?

@tischi @StephanPreibisch