// the find
yihong0618/GitHubPoster
Make everything a GitHub svg poster and Skyline!
A CLI that turns activity data from ~30 different services (Strava, Duolingo, WakaTime, Bilibili, GPX files, even Apple Health) into GitHub-contribution-graph-style SVGs and 3D skylines. It's for developers who want a novel way to visualize non-coding habits — running, reading, language learning, gaming — in the same visual language as a commit graph, and who don't mind scraping their own session tokens to get there.
The loader abstraction (base_loader.py) is clean enough that adding a new data source is a well-defined, repeatable task, which is how it got to 30+ integrations without the core renderer changing. The actual rendering options are more than a gimmick — animated SVGs, circular layout, and the 3D skyline mode produce genuinely good-looking output, not just a reskinned heatmap. It ships a GitHub Actions workflow so you can auto-regenerate and commit the poster to your profile README on a schedule, which is the actual use case most people want.
Over half the loaders (Bilibili, Kindle, Shanbay, Jike, WeChat Read, Nike) depend on manually extracted cookies or session tokens from undocumented/private XHR endpoints — these are the first things to silently break when a site changes its auth flow, and there's no fallback or clear error path described for when that happens. Test coverage is two files (utils and Strava) covering a codebase with 30+ independent loaders, so regressions in any of the other integrations likely go unnoticed until a user reports them. Documentation is primarily Chinese with a separate English README that lags behind — most of the per-loader setup instructions (the fiddly cookie-extraction steps) are only in Chinese, which is a real barrier for non-Chinese-reading contributors or users. 'Refactor some code' has sat on the TODO list for years, suggesting the maintainer already knows there's debt in the core that outsiders will hit before it gets addressed.