CI/CD
Package
Meta

This provides a plugin for Hatch that allows the use of conda environments.

This project is a copied and modified version of the hatch-containers plugin by Ofek Lev.

Table of Contents

• Installation
• Configuration
  • Python
  • Command
  • Conda-forge
  • Environment file
• Notes
• License

pip install hatch-conda

The environment plugin name is conda.

• pyproject.toml

  [tool.hatch.envs.<ENV_NAME>]
  type = "conda"

• hatch.toml

  [envs.<ENV_NAME>]
  type = "conda"

If the Python version is set to a multi-character integer like 310 then it will be interpreted as its <MAJOR>.<MINOR> form e.g. 3.10.

If not set, then the <MAJOR>.<MINOR> version of the first python found along your PATH will be used, defaulting to the Python executable Hatch is running on.

The command option specifies the command that will be used to setup the environment. The possible options are conda, mamba and micromamba.

Default:

[envs.<ENV_NAME>]
command = "conda"

Indicates if the conda-forge index should be used.

Default:

[envs.<ENV_NAME>]
conda-forge = true

By default packages will be installed using pip. However, to install packages using conda, conda-forge, or any other channel, you can specify a conda environment file:

[envs.<ENV_NAME>]
environment-file = "environment.yml"

When using an environment file, the channel and python version specified in the environment file will be used. After installing the environment, any extra packages specified in the dependencies will be installed, as well as the local package.

It's possible to identify environment py prefix (the path in the filesytem) rather than by name.

[envs.<ENV_NAME>]
prefix = .venv/

Will create a conda environment in the .venv/ folder. This is useful because a text editor might check for this location to use the virtual environment.

• There must be a conda, mamba, or micromamba executable along your PATH.
• The env-exclude environment variable filter has no effect.

hatch-conda is distributed under the terms of the MIT license.