tickets/PREOPS-4647: update night in prenight when visits loaded, if needed

[ehneilsen]

No description provided.

[codecov]

Codecov Report

Attention: 15 lines in your changes are missing coverage. Please review.

Comparison is base (2eb034e) 12.24% compared to head (fdbd545) 12.53%.
Report is 4 commits behind head on main.

❗ Current head fdbd545 differs from pull request most recent head 490f425. Consider uploading reports for the commit 490f425 to get more accurate results

Files	Patch %	Lines
schedview/app/prenight/prenight.py	0.00%	8 Missing ⚠️
schedview/compute/astro.py	0.00%	7 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #58      +/-   ##
==========================================
+ Coverage   12.24%   12.53%   +0.29%     
==========================================
  Files          39       40       +1     
  Lines        3243     3271      +28     
  Branches      492      495       +3     
==========================================
+ Hits          397      410      +13     
- Misses       2837     2852      +15     
  Partials        9        9              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[ehneilsen]

I think something is wrong with codecov. It's complaining about a bunch of lines in schedview.compute.astro.compute_central_night not being tested, but I'm sure they are. Not only to I see the test case I expect to be covering them being run in the github action, I can double check that these lines are in fact covered by the test by setting a breakpoint when running locally, and walking through the lines of code when they are run with pytest.

[rhiannonlynne]

Yeah, the codecov report seems odd in other repos as well. I'm not entirely sure what to do about it.
The report seems not necessarily in line with what seems reasonable to me either.
I suppose a minimum thing to do would be to take the badges off the front pages (of our other repos, not added yet for schedview).

[rhiannonlynne]

Does this mean the dashboard will start up on the middle night out of a set of visits? I still think the first night is probably where you'd like to start, but I suppose further testing will help understand better what's most useful.

[ehneilsen]

Yes, currently it starts in the middle. I chose that rather than the first because I've seen sims that just have one or two exposures on one night, followed by a full night on the next, in which case it's clearly the next night that is wanted.

A better approach might be use the date of the first night with more than some number of exposures, but really, when faced with a database with multiple nights, anything the code might do is a just a wild guess anyway.

I can imagine that in the future that the code that automatically generates sims for the pre-night briefing will do exactly some number of nights, and reliably not start in the middle of a night, in which case it will either not matter (if it does one), or will be easy to change to do the first night.