PhyloGNN ======== .. container:: hero Practical graph learning tools for phylogenetic trees. PhyloGNN converts `ete3.Tree` objects into PyTorch Geometric `Data` graphs, adds deterministic node features, and supports local GNN training workflows on those graph samples. .. container:: hero-actions :doc:`Start with the quickstart ` :doc:`Install the package ` Fastest path ------------ .. container:: entry-grid .. container:: entry-card **Install** Start with the core package and add optional extras only for docs, `BEAST` file input, or experiment tracking. :doc:`installation` .. container:: entry-card **Quickstart** Build a small `ete3.Tree`, attach features, convert it to graph data, run a tiny training smoke test, and print a prediction. :doc:`quickstart` .. container:: entry-card **User Guide** Follow the full workflow from tree input through feature engineering, graph conversion, datasets, training, configuration, and tracking. :doc:`user_guide/index` .. container:: entry-card **Examples** Launch six runnable scripts that cover feature engineering, graph conversion, tree I/O, training, TOML configuration, and prediction. :doc:`examples/index` .. container:: entry-card **Reference** Look up curated public APIs for data conversion, models, training, I/O, and utilities. :doc:`reference/index` .. container:: workflow-strip **Typical workflow:** tree input → feature engineering → graph conversion → training → metrics tracking when needed → API lookup for implementation details. .. Maintainers: add new user-facing task pages under user_guide/ and link them from user_guide/index.rst. Add lookup pages under reference/, and recovery advice in troubleshooting.rst. Keep legacy docs/issues, docs/myprompt, and specs out of official toctrees unless the content is rewritten for ordinary users. .. toctree:: :maxdepth: 2 :caption: User Documentation installation quickstart user_guide/index examples/index reference/index troubleshooting