Closed taldcroft closed 4 years ago
@jeanconn - the actual fix here is trivial. Testing to prove that it worked was rather more work, but I think this is good to go.
I think the 'live' testing we did at the time was sufficient.
As I have no recollection or record of that, the notebooks were necessary. My bad to not make note of this in the original PR description. That's part of why I'm working to not let PR's linger.
My recollection was we both ran independently-hacked-but-basically-identical versions of this code to get the non load commands in after the NSM (and both reviewed outputs). I didn't realize the PR was lingering.
This is a simple change to the ingest of non-load commands. It makes the applicable date range for commands go out to the ingest
stop
date instead of the end of the last timeline. The previous logic was incorrect since during an anomaly the last timeline is clipped so you don't get non-load commands in the kadi cmds archive.Testing
This works, as demonstrated in these two notebooks: