is there a more robust way to get 3rd-party dependencies out of a project in Pants than parsing output of the

./pants dependencies project::

or

./pants peek project::

commands? I was relying on

./pants dependencies project:: --dependencies-type=3rdparty

but this option was deprecated in

2.9.0

. The help docs say that a more robust way of querying dependencies should be available in the later versions, but I struggle to find any.

We generally recommend using

./pants export ::

to create a virtualenv instead. But if you just want the strings, the script in https://www.pantsbuild.org/v2.10/docs/python-third-party-dependencies#manual-lockfile-generation-techniques has a snippet using

peek

+

jq

thanks, Eric! Sorry, I should have provided more context. What I am trying to do is to get a list of 3rd party dependencies that a project depends on (e.g.

projectA

depends on

numpy

and

pandas

— Pants should know this thanks to dependency inference). The

peek

+

jq

example gives me a list of requirements as per my

requirements.txt

file, not the actual dependencies of the projects. E.g. I’ve just added a few random PyPI packages (that no project is using) to the

requirements.txt

and the

export

command happily (totally expected though) created the venv with those deps. The use case is that the

requirements.txt

may contain arbitrary packages and I would like to make sure that all of them are actually being used in the projects. This is to find requirements that are present but are not used by code (some project depended on the package, but then code is refactored/removed and not this package is no longer required).

That sounds like a built-in lint everyone would appreciate for all languages Pants supports!