tox-conda

tox-conda is a plugin that provides integration with the conda package and environment manager for the tox automation tool. It's like having your cake and eating it, too!

By default, tox creates isolated environments using virtualenv and installs dependencies from pip.

In contrast, when using the tox-conda plugin tox will use conda to create environments, and will install specified dependencies from conda. This is useful for developers who rely on conda for environment management and package distribution but want to take advantage of the features provided by tox for test automation.

tox-conda has not been tested with conda version below 4.5.

Getting Started

tox-conda can be used in one of two ways: by installing it globally and by enabling it on a per-project basis. When the plugin is installed globally, the default behavior of tox will be to use conda to create environments. To use it on a per-project basis instead, use tox's auto-provisioning feature to selectively enable the plugin.

To enable the use of tox-conda by default, follow the Installation instructions. To use the plugin selectively, do not manually install it, but instead enable it by adding tox-conda as a provisioning requirement to a project's tox config:

[tox]
requires = tox-conda

More information on auto-provisioning can be found in the tox documentation.

Installation

The tox-conda package is available on pypi. To install, simply use the following command:

$ pip install tox-conda

To install from source, first clone the project from github:

$ git clone https://github.com/tox-dev/tox-conda

Then install it in your environment:

$ cd tox-conda
$ pip install .

To install in development mode:

$ pip install -e .

The tox-conda plugin expects that tox and conda are already installed and available in your working environment.

Usage

Details on tox usage can be found in the tox documentation.

With the plugin enabled and no other changes, the tox-conda plugin will use conda to create environments and use pip to install dependencies that are given in the tox.ini configuration file.

tox-conda adds four additional (and optional) settings to the [testenv] section of configuration files:

conda_deps, which is used to configure which dependencies are installed from conda instead of from pip. All dependencies in conda_deps are installed before all dependencies in deps. If not given, no dependencies will be installed using conda.
conda_channels, which specifies which channel(s) should be used for resolving conda dependencies. If not given, only the default channel will be used.
conda_spec, which specifies a conda-spec.txt file that lists conda dependencies to install and will be combined with conda_deps (if given). These dependencies can be in a general from (e.g., numpy>=1.17.5) or an explicit form (eg., https://conda.anaconda.org/conda-forge/linux-64/numpy-1.17.5-py38h95a1406_0.tar.bz2), however, if the @EXPLICIT header is in conda-spec.txt, all general dependencies will be ignored, including those given in conda_deps.
conda_env, which specifies a conda-env.yml file to create a base conda environment for the test. The conda-env.yml file is self-contained and if the desired conda channels to use are not given, the default channels will be used. If the conda-env.yml specifies a python version it must be compatible with the basepython set for the tox env. A conda-env.yml specifying python>=3.8 could for example be used with basepython set to py38, py39 or py310. The above conda_deps, conda_channels, and conda_spec arguments, if used in conjunction with a conda-env.yml file, will be used to update the environment after the initial environment creation.
conda_create_args, which is used to pass arguments to the command conda create. The passed arguments are inserted in the command line before the python package. For instance, passing --override-channels will create more reproducible environments because the channels defined in the user's .condarc will not interfer.
conda_install_args, which is used to pass arguments to the command conda install. The passed arguments are inserted in the command line before the dependencies. For instance, passing --override-channels will create more reproducible environments because the channels defined in the user's .condarc will not interfer.
If mamba is installed in the same environment as tox, you may use it instead of the conda executable by setting the environment variable CONDA_EXE=mamba in the shell where tox is called.

An example configuration file is given below:

[tox]
envlist =
    {py35,py36,py37}-{stable,dev}

[testenv]
deps=
    pytest-sugar
    py35,py36: importlib_resources
    dev: git+git://github.com/numpy/numpy
conda_deps=
    pytest<=3.8
    stable: numpy=1.15
conda_channels=
    conda-forge
conda_install_args=
    --override-channels
commands=
    pytest {posargs}

More information on tox configuration files can be found in the documentation.

Contributing

Contributions are very welcome. Tests can be run with tox, please ensure the coverage at least stays the same before you submit a pull request.

License

Distributed under the terms of the MIT license, "tox-conda" is free and open source software

Issues

If you encounter any problems, please file an issue along with a detailed description.