Make necessary updates to load HXN data from tiled

[psavery]

For this, I tried to take an approach such that the relevant functions, such as map_data2D_hxn(), will run properly whether db is a v0 DataBroker instance or a v2 instance (which uses tiled). This would enable db to just be swapped out for a v2 instance when the time is right.

There is a variable is_tiled that tries to determine whether v2 is being used or v0, and some of the logic, in just a few places, branches based upon whether tiled is being used or not.

Some of the metadata needed to be accessed like start_doc['dimensions'] instead of start_doc.dimensions, but this appeared to work for both v0 and v2.

I realize this does not take into account the changes in #339. But I think the changes here are simple enough to take into account once #339 is merged.

These changes appeared to work fine with DataBroker v0. I had an example using DataBroker v2 that appeared to have the metadata, but not the data itself. I think it loaded and formatted all the metadata correctly, but there is a little uncertainty as to whether loading the data in to the pandas frame at the end is correct - that is something we can determine once we have a real example.

[genematx]

Hi, @psavery , thank you putting this PR together -- I think this is a really great start! What we may need to do is to replace databroker completely and use the pure Tiled client instead (ideally pointing to the SQL-backed server), e.g. cat = from_uri("https://tiled.nsls2.bnl.gov")['hxn/migration'] ("migration" is the most modern catalog backed by SQL; it will become "raw" eventually once we fully migrate the beamline). With this catalog, we wouldn't need databroker at all, and we should aim to get rid of any db and hdr variables to make it unambiguous.

Ideally, any processing script/method would take an instance of the Tiled client as an argument and extract all necessary information (data and metadata) directly from it (instead of having several data-loading functions that rely on a global instance of db). I would like to better understand, which processing functions we need for this project and start unravelling this code from the end.

I am happy to help you with this migration to Tiled; it shouldn't be very hard, and it looks like you've already got most of the parts covered.