ghcontribs: A tool for looking at user contributions to GitHub

As I’ve been preparing to get a “real job,” I realized that one of my selling points – especially to organizations that focus on open source development – is that I have a decent bit of experience in the open source community, particularly around computational chemistry.

But although I tend to remember the highlights of my contributions, especially contributions to big projects (millions of downloads per month), I was curious what the actual statistics on my contributions looked like.

So I wrote a little tool, ghcontribs, that would help me look at that. In the end, there are some other tools that might be more useful for some users, in particular show-me. I think my tool does a little more to facilitate a deep-dive into a user’s contributions, especially if you’re working from an interactive environment like a Jupyter notebook.

From a software development standpoint this project has two kind of interesting aspects. The first is that it gave me an excuse to start to learn the GitHub v4 GraphQL API. The GraphQL API is, for many reasons, an improvement on the older v3 REST API. However, it doesn’t lend itself to writing a library as easily as the older API does, so one tends to need to write custom code to meet the goals of a given tool.

The second interesting software development idea I played with here has to do with Python’s Enum class, and how it differs from an enum in C. For the Python Enum, the values can be any object, which means that we can carry additional information (or methods) in the values of that Enum, unlike the C enum which must take integer values. This allows some simplified design choices that I tried experimenting with in this package: for example, the specific URL for an issue/pull request (PR) in GitHub can be determined based on the Enum value identifying whether a contribution was an issue or a PR.

In any case, this tool is a bit of a one-off to scratch an itch. It has enough of a test suite that it should be maintainable, if there is interest, but I don’t expect to be adding significant new features in the future. It has been released on PyPI, so you can install it with python -m pip install ghcontribs.