AD_HDF5_SINGLE: return sliceable object

[prjemian]

• Is AD_HDF5_SINGLE for general use? #31

[prjemian]

The CI failure is because the conf.py file is out of date with current Sphinx. (need to comment the language= line now, the default is acceptable).

[danielballan]

I am slightly worried this API change could break other uses of AD_HDF5_SINGLE. Would it be better to make a new handler (and new spec name) with the new behavior? Or is this actually a bug, @tacaswell?

[tacaswell]

I think this is actually a bug and I think this change is mostly backwards compatible (what would you be doing to a list that a N-D stack of numpy arrays would not handle?).

I am in favor of merging this but making sure the SXSS AST is very aware of this.

[tacaswell]

This may also fix a latent bug that we have not hit at SIX.

[prjemian]

Here's my notebook that relies on this change: https://github.com/BCDA-APS/apstools/blob/691-AD-HDF5-Single-mode/docs/source/examples/de_2_adsim_hdf5_single_mode.ipynb

[prjemian]

@tacaswell - Thanks for the review, and the fix code (that's yours). @danielballan - Thanks for the merge.

[danielballan]

Released https://pypi.org/project/area-detector-handlers/0.0.10/