Bandwidth/Storage issue when rating a number of oss

[SimoMay]

Hi all, I'm using the YAML config approach to rate a list of GitHub projects.

My issue is every time the calculations run it clones each project to disk, using bandwidth and storage on my server.

Is there a way to stop the cloning part, even if it reduces the confidence level of the rating?

[sourabhsparkala]

@SimoMay

• The confidence level of the rating has nothing to do with the cloning.

The reason behind cloning was that we want to restrict the number of calls to GitHub API to as few as possible.

As part of the rating calculation, we try to collect features from a GitHub repo, and it would not make sense to exhaust GitHub token by making multiple API calls to the same project. Hence, to avoid this drawback, we decided to clone the specific repositories and then collect the necessary features directly from the cloned repo.

[SimoMay]

Hi @sourabhsparkala

Make sense then if the rating is built around getting the code and evaluating it locally. Thanks.

Although I remember I had issues at first compiling and running the tool (java versions mess ups) and one of the issue was the tool couldn't create directory for the repositories, it didn't break the process, got a message stating the error then saying it will continue rating but with a lower score, that's why I associated cloning the repo with a lower score.

The error was errors.JGitInternalException: Creating directories for /work/.fosstars/repositories/x/.git