Is the recommended pattern for writing repo specific plugins still “add path to python_path” and then building a regular python module on that path? For example in

pants-plugins/src/python

I create a module called

awesome_plugin

. This module would need to have

awesome_plugin.register

defined. Is there a pattern that takes advantage of “auto discovery” of plugins that are in the monorepo as opposed to being loaded via pip via an external repository?

Hey Raúl! Pardon the delay. Yes, this is still the case. FYI it’s because Pants loads plugins before doing everything else, so we add to the PYTHONPATH for it be discoverable. (I believe this is how it works, might be wrong) Also, no need to use

src/python

.

pants-plugins/awesome_plugin/register.py

works too. Plugins must be written in Python, so

src/python

is redundant

Good to know.

For instance if I create a databricks target for my repo and want to share the plugin, seems the best way is to publish it to pypi.

Yes, that’s the plan. Like how people share Flake8 plugins by releasing a library like

flake8-2020

. Then, users add your plugin to the

plugins

option in their

pants.toml

. I’m not sure how well it works yet for a 3rd party plugin author, though, tbh. Pants releases a couple plugins like

pantsbuild.pants.contrib.node

, but I don’t think people have done 3rd party plugins in a while. That’s something we want to change, in part by making V2 plugins much easier to write than V1 plugins and making them more powerful. We’d love to get to a place like Flake8, where you have dozens of plugins you could choose from.