Codecov Report

Base: 62.26% // Head: 62.18% // Decreases project coverage by -0.07% ⚠️

Coverage data is based on head (c757529) compared to base (31c3865).
Patch coverage: 52.00% of modified lines in pull request are covered.

❗ Current head c757529 differs from pull request most recent head 82f6691. Consider uploading reports for the commit 82f6691 to get more accurate results

@@            Coverage Diff             @@
##           master    #1374      +/-   ##
==========================================
- Coverage   62.26%   62.18%   -0.08%     
==========================================
  Files          89       89              
  Lines       20260    20239      -21     
  Branches     3652     3657       +5     
==========================================
- Hits        12614    12585      -29     
- Misses       6856     6868      +12     
+ Partials      790      786       -4     

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

@JPGlaser I think that adjusting your script to call pint.utils.preload_cache should reliably ensure that all files are loaded to the Astropy cache even starting from nothing. You could call pint.utils.set_no_internet() at the beginning of all scripts, but almost certainly you are better off creating an Astropy config file with the same options in it. (You'll need to set them back in the updater script.)

This one is a bit of a headache because I'm trying to get the "don't access the Internet, be okay with old data" to work with everything from Astropy 4.0 to 5.1, and the code to deal with out-of-range data has changed several times over the course of that time. I might need to create some specialized tox tasks and test all the versions. We particularly need 4.3 to work because that's what NANOGrav used for 15yr and they want to keep using it; 5.1 works fine and quite gracefully, but so far 4.3, 5.0, and 5.1 need different solutions.

@aarchiba We are okay with advancing to Astropy 5.1+ with the 15yr. This is because we have to advance our versions of other software for the wideband results. I think this feature would be a fantastic add as we get more folks in the IPTA using pint. ^_^