Why Use Conda?

I recently wrote a post guiding users through the process of publishing their python package to PyPI to be installed via pip. However, many users prefer to use conda. I like to have my packages available on conda since it plays nicely with packages that have external dependencies. Sometimes a package will also only be available on a conda channel, and it is best to avoid mixing pip and conda installations in a single environment in order to steer clear of dependency issues.

Note: If you have not already published your package on PyPI, make sure to see the initial steps I layout in my other post. You need to have the correct file hierarchy, ___init_.py_ files, and a License. See my example Github repo for an example of proper file hierarchy.

Conda Channels

Channels are simply the locations where packages are stored. Some very popular packages are stored in the default conda channel, many smaller packages are stored in individual user’s own channel, or there are community channels like conda-forge.

You can specify the channel you would like to use when installing a package in the command line. Or you can create a conda configuration file (.condarc ), which allow you to list channels in order of priority.

conda install swepy --channel conda-forge

#conda #anaconda #open-source #python