[tim777z]

DataFactor Assessment

Score: 32.4/100 (grade F)

Summary

Seaborn (as captured in this repository snapshot, 2015 vintage, v0.7.0.dev) is a mature, well-tested statistical plotting library, not really an "ai_research" project despite the classifier's label. It has a real test suite (10 spec files, 1:7 ratio) exercising public plotting API behavior, decent docstrings/API docs embedded in source, and Travis CI, but no coverage enforcement, no lint/type enforcement in CI, several 900-2000+ LOC god files, no lockfile, no Dockerfile or reproducibility tooling, and no experiment-tracking or ML-specific reproducibility apparatus since this is a visualization library rather than an ML/research codebase. Onboarding is functional (README covers install/test) but thin on architecture, changelog, and environment reproducibility.

Recommendations Addressed

• Make install, build, and test work from a fresh clone [high]
• Add a docker compose that brings the whole app up [medium]
• Add an enforced test job to CI that actually runs the suite [high]
• Introduce a lockfile and pin dependency versions [medium]
• Split god files and add lint enforcement [medium]
• Document architecture and add reproducible dev environment [medium]
• Sustain small, test-paired commits going forward [high]

[tim777z]

Closing duplicate - replaced by newer PR