Tox: https://tox.wiki/en/latest/
pre-commit: https://pre-commit.com/
I would like to understand the borders for both choices.
I know that pre-commit creates a py environment - same as tox. To me, their architecture looks a bit the same.
Some people use them in combination... what pre-commit can't do, that tox can? I saw examples where with pre-commit during CI pipeline you can run unit tests, etc.
Which one is the best to integrate within the CI build?
In short, pre-commit is a linter/formatter runner, tox is a generic virtual env management and test command line tool.
While tox could run linters too, it is tedious to manage the versions of the linters. In pre-commit you can just run pre-commit autoupdate, and all linters get updated.
On the other hand tox can run e.g. a test suite or a coverage report for many different Python versions. This is not only helpful for a library, but also for an app - so you can already test the upcoming Python versions.
tox is also used to create documentation, and sometimes also to make a release - you can't (shouldn't) do this with pre-commit.
And tox is certainly not obsolete, although GitHub actions can test against different Python versions, as you can run tox both local and in CI.
I gave a lightning talk on this topic: https://www.youtube.com/watch?v=OnM3KuE7MQM
Which one is the best to integrate within the CI build?
I like to run pre-commit via tox, both in CI and locally.