Installation#
There are two basic options for installing tskit
: either through
pre-built binary packages using Conda or
by compiling locally using Pip. We recommend using conda
for most users, although pip
can be more convenient in certain cases.
Tskit is installed to provide succinct tree sequence functionality
to other software (such as msprime),
so it may already be installed if you use such software.
Requirements#
Tskit requires Python 3.8+. There are no external C library dependencies. Python
dependencies are installed automatically by pip
or conda
.
Conda#
Pre-built binary packages for tskit
are available through
conda, and built using conda-forge.
Packages for recent version of Python are available for Linux, OSX and Windows. Install
using:
$ conda install -c conda-forge tskit
Quick Start#
Install
conda
using miniconda . Make sure you follow the instructions to fully activate yourconda
installation!Set up the conda-forge channel using
conda config --add channels conda-forge
.Install tskit:
conda install tskit
.Try it out:
tskit --version
.
There are several different ways to obtain conda
. Please see the
anaconda installation documentation
for full details.
Pip#
Installing using pip
is somewhat more flexible than conda
and
may result in code that is (slightly) faster on your specific hardware.
Pip
is the recommended method when using the system provided Python
installations. Installation is straightforward:
$ python3 -m pip install tskit
Development versions#
For general use, we do not recommend installing development versions.
Occasionally pre-release versions are made available, which can be
installed using python3 -m pip install --pre tskit
. If you really need to install a
bleeding-edge version, see Installing development versions.