hercules

Processes large repositories like the Linux kernel in under two hours using efficient algorithms, such as a custom RB tree for incremental blaming, making it significantly faster than alternatives like git-of-theseus.

Includes a wide range of analyses out-of-the-box, such as burndown statistics, couples detection with Tensorflow embeddings, and sentiment analysis, demonstrated with examples like torvalds/linux and tensorflow/tensorflow in the README.

Features a Directed Acyclic Graph (DAG) of analysis tasks and a plugin system, allowing users to extend or customize analyses, as detailed in the PLUGINS.md documentation.

Supports combining results from multiple repositories via 'hercules combine' and offers caching mechanisms to speed up repeated analyses on the same repository, enhancing organizational insights and efficiency.