[Bug] GitHub does not return languages for some large repos

[tjwp]

The GitHub languages endpoint for a repo used here can return an empty response for very large repos:

https://github.com/Codium-ai/pr-agent/blob/279d45996fbc9ea70c1269ad1fca270782832b08/pr_agent/git_providers/github_provider.py#L230

An empty result for languages currently prevents some further processing of the pull request, e.g. no walkthrough for PR description.

This could be addressed by skipping file sorting if there are no languages, or allowing the languages weighting to be provided via configuration if it cannot be fetched via API.

[mrT23]

can you share an example for a public repo where we get an empty response ?

[tjwp]

can you share an example for a public repo where we get an empty response ?

I don't know of an example public repo. If I come across one, I'll post it. I think they should be easy to spot because Languages will not appear for the repository in the web UI. The symptom is that the get_languages() call above returns an empty Dictionary {}.

[mrT23]

i am not sure if this is a property of "big" repos, or repos that for some reason disabled usage of the language feature.

in any case, handled this In #306

[tjwp]

Thanks for fix!