I'm working in migrating a bunch of legacy projects in the organisation I'm currently working on from Python 2.7 to Python 3.7.

I've successfully used tox before for a few other personal projects, but this time I'm facing a situation where I have many dependencies which are available only via conda packages in an internal conda channel. Basically the deployable packages are all conda packages, since the company relies heavily in conda infrastructure to build/deploy/release software.

The "canonical" use case for tox is to be used in combination with PyPI. While many third-party dependencies packages are available externally on PyPI, all the dependencies which are built in-house are not (only via conda).

I tried using install_command option to invoke creating a conda environment based on the top-level package, then removing it and activating the environment to run the tests with pytest, but didn't get very far in succeeding with it.

I'd like to know if there are any guidelines, successful cases, hints, or anything that can point me in the right direction of integrating tox with conda.

In other words, my problem is that I need to install conda packages in the virtual environments created by tox before running the tests.

Can anyone please kindly shed some light here?

Many thanks.

Kind regards,

Gabriel