Installation
============

PhyloGNN is a Python package requiring Python 3.12 or newer. Install from a
local checkout with the smallest extra set that matches your task.

Core package
------------

Use the core install when you need `ete3.Tree` objects, feature engineering,
graph conversion, models, and local training. This includes the public
`PyTorch`, `PyTorch Geometric`, `torch-scatter`, `ETE3`, `NumPy`, and `tqdm`
runtime dependencies.

.. code-block:: bash

   python -m pip install -e .

Development tools
-----------------

Use the development extra for tests, linting, and formatting.

.. code-block:: bash

   python -m pip install -e ".[dev]"

Documentation tools
-------------------

Use the documentation extra to build this local Sphinx site. It adds Sphinx and
the Read the Docs HTML theme, but does not make either package a runtime
dependency.

.. code-block:: bash

   python -m pip install -e ".[docs]"
   python -m sphinx -b html -n -W --keep-going docs/source docs/_build/html

Optional tree I/O
-----------------

The core tree workflow works with `ete3.Tree` objects. File workflows that read
`NEXUS` or `BEAST`-style trees through `DendroPy` require the `beast` extra.

.. code-block:: bash

   python -m pip install -e ".[beast]"

Optional experiment tracking
----------------------------

Weights & Biases tracking is isolated behind the `wandb` extra and is only
needed when tracking is enabled.

.. code-block:: bash

   python -m pip install -e ".[wandb]"

Everything optional
-------------------

Use the `all` extra for optional tree I/O, workflow helpers, and tracking. It
does not include development or documentation tooling.

.. code-block:: bash

   python -m pip install -e ".[all]"

Next steps
----------

Continue with the :doc:`quickstart` as the first tutorial. Use
:doc:`user_guide/graph_conversion` when you need to interpret the graph fields
produced by conversion.