Translator Modulification

[crankynetman]

Right now the translator isn't setup properly as a python module. This makes doing stuff with pytest REALLY difficult, and it'd be nice to have this be setup as a proper module with init.py, and the likes.

During the solving of this, it would also be nice to come up with a better way to generate the gobgp python client. Right now all the files are generated when the container is built and then stored in the same directory as the translator.py and gobgp.py files. This has a few downsides:

• Breaks Intellisense/Code completion and other IDE tools that rely on being able to fully analyse the code (pylint/flake8 on save for example)
• You can't easily run the code outside of the container (because the deps are not copied back out in any way)
• You cant really run the code outside of translator.py even within the container because of dependencies and imports and other goofiness.

This work would, in my mind, involve:

• Taking the translator code and turn it into a proper python module using whatever the current Right Way® is in python that you're supposed to use for doing that these days (i.e. not setup.py anymore I think).
• Figuring out a way to actually package up the gobgp library into something that we can have available locally and inside the container.
  • Maybe we do this by copying the files around? I tried this a bit but with mounting them, it gets a bit crazy.
  • Maybe we generate these files by Some Process® and have them checked-into the repo. They Shouldn't Change® unless we bump the GoBGP version anyways, so re-building them every time the way we do isn't really necessary.
  • Figure out how other people fit this into their docker workflows. Surely someone has solved this.
• Maybe we even want to take this opportunity to separate the translator code/container into a separate repo and combine with with breaking apart the docker compose stuff from the actual code stuff?
• Potentially considering how a user might add translators to SCRAM without having to update the main repo and make a pull request, etc. (this might be early optimization though)