Committing codelists.json to the repo

[alexwalkerepi]

Perhaps for @inglesp:
It doesn't seem to make a difference if I do or don't, but is it ever necessary or useful to commit the codelists.json to the repo? Is it intended as useful metadata for when the codelists were downloaded etc, or is it just used under the hood and not useful later? If the latter we could add it to the .gitignore in the research template.

[alexwalkerepi]

Ah, it seems it does make a difference to the github tests. It failed before I committed the changes to codelists.json:
https://github.com/opensafely/covid-vaccine-effectiveness-research/runs/2022648885?check_suite_focus=true

then passed after I committed the changes:
https://github.com/opensafely/covid-vaccine-effectiveness-research/runs/2022879219?check_suite_focus=true

[inglesp]

One for @evansd?

[evansd]

Yes, this should be committed alongside all the other files in the codelists directory.

At the moment you can get away with not including it at all because we wanted to avoid breaking existing studies so we put a workaround in place. The test output contains a warning about this but it gets lost in all the other output:

==> WARNING
    Using temporary workaround for Github Actions tests.
    You should run: opensafely codelists update

The reason we create this file is to make sure that the contents of the codelists directory exactly matches what was downloaded from OpenCodelists, but without having to contact OpenCodelists each time we run the tests which has the potential to make tests unreliable.