A teachable moment using Intake & catalog

[bruceravel]

Following a suggestion from @danielballan, I am leaving this issue here. I have recently completed a crude scan telemetry tool for BMM which could make a decent catalog tutorial. The basic idea was to determine the amount of overhead for an XAS scan above the measurement time (i.e. overhead due to mono motion, network latency, gremlins, etc). My solution was to search for all the XAS scans I had made since some time in 2019 and to do some simple statistics on the integration time column of the data tables a well as the timestamps in the start and stop documents. While this is admittedly a blunt tool for determining scan telemetry, it's not bad. "Good enough for government work."

Among the things I've done that could make a good teaching moment:

1. Searching catalogs by date and by content of start document and making new catalogs from the results
2. Accessing and making calculations from 1000s of database records
3. Developing heuristics for rejecting problematic records (e.g. scans that span a beam dump, scans that did not run to completion)
4. Saving the results in a reusable form -- in this case, to make a guess (with uncertainty) for how long a new XAS scan will take

I would be happy to develop this into a Jupyter thingie suitable for a Users' Meeting presentation.

See https://github.com/NSLS-II-BMM/profile_collection/blob/master/startup/BMM/telemetry.py for all my crappy code.