Installation

PYPI

We have posted wgd v2 on PYPI so the easiest way of installing is via pip:

$ pip install wgd

The version can be specified by replaceing “wgd” with, for instance “wgd==2.0.19”.

Source

Another recommended way is to install from the source. The example code is below:

$ git clone https://github.com/heche-psb/wgd
$ cd wgd
$ virtualenv -p=python3 ENV (or python3 -m venv ENV)
$ source ENV/bin/activate
(ENV) $ python setup.py install (or python -m pip install -r requirements.txt)
(ENV) $ pip install .

Note

We suggest setting a separate virtual environment in case that the dependency of wgd v2 messes up the original package environment. When met with permission problem in the pip install . step, please try adding the option -e. If multiply versions of wgd were installed in the system, please make sure setting the one (either wgd v1 or wgd v2) desired ahead in the environment variables. The version of numpy is important and often the dependency of other packages, for instance fastcluster. We found that some build errors come from the incompatibility of numpy version required by other packages. So far, the numpy version 1.19.0 works in our test. Since the pip can not always decide the best compatible package and the existing packages that match the requirement will not be redownloaded again. If users met some installation issues, it’s suggested to pre-install the numpy version 1.19.0 before installing the wgd v2 package and try again.

Conda

We also posted wgd v2 on bioconda. With an activated Bioconda channel, install with:

$ conda install wgd

Docker

The docker container is another possible option for installation.

$ docker pull quay.io/biocontainers/wgd:<tag>