Look for the closest declination line RF model to produce the DL2 files

[marialainez]

see issue #275

[codecov]

Codecov Report

Attention: Patch coverage is 85.18519% with 12 lines in your changes missing coverage. Please review.

Project coverage is 81.58%. Comparing base (e589deb) to head (cbdee3e).
Report is 14 commits behind head on main.

❗ Current head cbdee3e differs from pull request most recent head 9b4b441

Please upload reports for the commit 9b4b441 to get more accurate results.

Files	Patch %	Lines
src/osa/paths.py	81.81%	8 Missing ⚠️
src/osa/utils/utils.py	66.66%	4 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #280      +/-   ##
==========================================
+ Coverage   79.76%   81.58%   +1.81%     
==========================================
  Files          59       57       -2     
  Lines        5144     4984     -160     
==========================================
- Hits         4103     4066      -37     
+ Misses       1041      918     -123     

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

[morcuended]

@moralejo, I was going to comment in today's meeting. The way we look for the adequate RF model declination track is by using the pointing information from the TCU database (specifically the target coordinates), not checking the average pointing information inside the data files as you mentioned. Is this fine?

[moralejo]

Since the relevant pointing is the one in celestial coordinates, the target coordinates are perfectly fine for standard observations (if we ever do something like drift scans we will anyway have to do some custom work).

[morcuended]

Looks good! thanks @marialainez. I leave some comments. I'd like to test it a bit myself. How about the memory and time requirements of the jobs including the dl1 to dl2 stage?

[morcuended]

Let's try to push this forward. I left some comments. I think the tricky part is to readjust the job's requirements (time and memory) when including dl1 to dl2 step, although that is independent from this PR.

[morcuended]

source_dec = run["source_dec"]

The pointing info is also available directly at the RunCatalog files (from TCU db). So in principle no need to access TCU db again.

[marialainez]

I tested the dl2 production with this branch and it works correctly with MEMSIZE_DATA: 60 GB.