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

From PyPI
---------

.. code-block:: bash

   pip install densitree

From source (development)
-------------------------

.. code-block:: bash

   git clone https://github.com/fuzue/densitree.git
   cd densitree
   pip install -e ".[dev]"

Dependencies
------------

densitree requires Python 3.10+ and the following packages (installed automatically):

.. list-table::
   :header-rows: 1
   :widths: auto

   * - Package
     - Minimum version
     - Purpose
   * - numpy
     - 1.24
     - Array operations
   * - scipy
     - 1.10
     - MST computation, distance metrics
   * - scikit-learn
     - 1.3
     - k-NN, agglomerative clustering
   * - networkx
     - 3.0
     - Tree data structure
   * - matplotlib
     - 3.7
     - Static visualization
   * - plotly
     - 5.15
     - Interactive visualization
   * - pandas
     - 2.0
     - DataFrame support, cluster statistics

Optional dependencies
~~~~~~~~~~~~~~~~~~~~~

For running benchmarks:

.. code-block:: bash

   pip install densitree[bench]

This adds ``flowio`` (FCS file reading), ``leidenalg`` and ``igraph`` (PhenoGraph-style clustering for comparisons).

For building documentation:

.. code-block:: bash

   pip install densitree[docs]