# DAMICORE DAMICORE groups objects by how well they compress together. It materializes objects from a dataset or a directory of files, computes an exact normalized compression distance between every pair, builds a deterministic Neighbor Joining tree from that matrix, and partitions the tree with FastGreedy community detection. The result is verified on load, and every run directory is self-contained. The library ships as five independently installable distributions. `damicore` orchestrates the pipeline and is what most readers want; the four stage distributions are there when one stage is wanted on its own. ```{toctree} :caption: Start here :maxdepth: 1 quickstart input-contract ``` ```{toctree} :caption: Use :maxdepth: 1 artifacts dataset-partitioning scalability ``` ```{toctree} :caption: The method :maxdepth: 1 method/index method/normalization method/ncd method/neighbor-joining method/community-detection method/evidence ``` ```{toctree} :caption: Where this is going :maxdepth: 1 roadmap ``` ```{toctree} :caption: Reference :maxdepth: 2 reference/index references ``` ```{toctree} :caption: Internals and contributing :maxdepth: 1 internals/documentation quality-drivers releasing internals/decisions ```