Riskfolio-Lib

Riskfolio-Lib turns a returns DataFrame into position weights. It is built on CVXPY and integrates with pandas data structures, exposing allocation through two main classes — Portfolio for mean-risk and risk-parity problems and HCPortfolio for hierarchical-clustering allocation — alongside modules for risk functions, parameter estimation, constraint construction, OWA weights, the Gerber statistic, DBHT clustering, plotting and reports. The README’s own counts describe mean-risk and logarithmic mean-risk (Kelly criterion) optimization over 26 convex risk measures with four objective functions, and risk-parity optimization over 22.

There is no peer-reviewed paper behind it: the project’s own Citing section supplies a @misc BibTeX entry crediting Dany Cajas and pointing at the GitHub repository. Maintenance is current — 36 commits in the trailing twelve months and five open issues as of August 2026, with the most recent commit moving the documentation to a new domain — but the bus factor is one: Cajas is the sole author and effectively the only committer.

Funding is worth knowing before reading the README, which opens with affiliate-tracked links to the author’s book and a paid Python portfolio-optimization course, and offers paid consulting in its Contributing section. None of this touches the library itself, which is fully BSD-3-licensed, sells no signals and makes no returns claims.