ENH: handler for hdf5 files that have variable frames per point

[tacaswell]

Work with @mrakitin to enable variable size images in streams

[danielballan]

Nice to see! I recall this limitation came up in our recent internal-DAMA ophyd tutorial. Funny that it showed up in the wild so soon after.

[mrakitin]

I've done the py2->3 conversion.

[mrakitin]

Also may need https://github.com/bluesky/ophyd/blob/f20c004430dab6dc8d00f9f018f570574928893c/ophyd/ophydobj.py#L297-L307.

[danielballan]

Can you explain more about the connection to dotted_name?

[mrakitin]

Can you explain more about the connection to dotted_name?

It's used here: https://github.com/NSLS-II-FXI/profile_collection/pull/3/files#diff-87e76dd09be85349d2409d8fdd9695e9R111-R113.

[danielballan]

Sorry, still not up to speed on how that relates to the handler code.

[mrakitin]

My #11 (comment) is not specifically relevant here, but rather is for NSLS2/fxi-profile-collection#3 in terms of the integration testing on the beamline. Sorry about the confusion.

[danielballan]

Ohhh right. No, I’m sorry for being dense. :-)

[danielballan]

Rebased, satisfied flake8, added entrypoints.

[danielballan]

This still needs a unit test, but it's ready for testing in the wild at FXI.