Cool looking graphs analyzing repo history with git-of-theseus

[echoix]

I learned last week about the cool little tool that analyzes the git history to produce graphs, git-of-theseus: https://github.com/erikbern/git-of-theseus. There's also a blog post: https://erikbern.com/2016/12/05/the-half-life-of-code.html

So I tried it on our repo and on grass-legacy. It took a while, maybe 30-45 min or so to analyze all commits and lines changed. It created various summary JSON files, and they can be plotted with other tools of the project (Matplotlib scripts).

Here are some figures for the main GRASS repo:

Stack plots (some are normalized to 100%)

Line plots (one is normalized to 100%)

Survival plots

Here are the raw JSON files for the main repo
git-of-theseus.zip
I also saved them as svg locally

See next comment for the grass-legacy, for upload limits, here: #4941 (comment)

In summary, we see that the locales folder is really the biggest part of our repo, and the author committing changes to these have the largest part of the contributions.

[echoix]

For the grass-legacy repo:

JSON summary data+all figures (png+svg):
git-of-theseus-grass-legacy.zip

Stack plots (normalized to 100% and non-normalized)

Line plots (normalized to 100% and non-normalized)

Survival plots

SVG version of figures + one PDF

stack_plot_dirs.pdf