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

Manual installation
-------------------

``PROSSTT`` was developed and tested in Python 3.5 and 3.6. For an optimal experience, we suggest Python 3.6. In order to run ``PROSSTT``, the following Python libraries have to be installed:

- scipy_ (tested: 1.0.0)
- numpy_ (tested: 1.14)
- pandas_ (tested: 0.22.0)
- newick_ (tested: 0.8.0)

In order to reproduce the example notebooks and produce plots with `generate_simN.py` the following libraries are additionally required:

- matplotlib_ (tested: 2.0.0, 2.1.2)
- jupyter_ (tested: 1.0.0)
- scanpy_ (tested: 0.4.2+9.ged7130f)

After installing all libraries: 

* download a ``PROSSTT`` release_  or clone the git_ repository.
* (optional) unpack the archive
* ``cd prosstt/directory/``
* ``pip install .``

.. _git: https://github.com/soedinglab/prosstt
.. _release: https://github.com/soedinglab/prosstt/releases
.. _newick: https://pypi.python.org/pypi/newick
.. _scipy: https://www.scipy.org/
.. _numpy: https://www.numpy.org
.. _pandas: https://pandas.pydata.org/
.. _matplotlib: https://matplotlib.org/
.. _scanpy: https://github.com/theislab/scanpy
.. _jupyter: http://jupyter.readthedocs.io/en/latest/index.html

Install via Bioconda
--------------------

(work in progress)