Add typing utils module

[DrPyser]

• JSON typing alias

[wazo-community-zuul]

Build succeeded.
https://zuul.wazo.community/zuul/t/local/buildset/de11ab99e3fd4e48952b43990963fd9c

✔️ tox-linters SUCCESS in 4m 13s
✔️ wazo-tox-py39 SUCCESS in 4m 27s
✔️ wazo-tox-integration-py39 SUCCESS in 3m 24s
✔️ debian-packaging-bullseye SUCCESS in 1m 53s

[wazo-community-zuul]

Build failed.
https://zuul.wazo.community/zuul/t/local/buildset/619e30717d8a4d918f5e5cb4f47c4cae

✔️ tox-linters SUCCESS in 4m 42s
✔️ wazo-tox-py39 SUCCESS in 4m 36s
✔️ wazo-tox-integration-py39 SUCCESS in 3m 14s
❌ debian-packaging-bullseye FAILURE in 1m 15s

[wazo-community-zuul]

Build succeeded.
https://zuul.wazo.community/zuul/t/local/buildset/d047e77e86aa4496a4a3a27f239e7adc

✔️ tox-linters SUCCESS in 6m 24s
✔️ wazo-tox-py39 SUCCESS in 7m 02s
✔️ wazo-tox-integration-py39 SUCCESS in 3m 35s
✔️ debian-packaging-bullseye SUCCESS in 1m 56s

[fblackburn1]

why 2014?

[DrPyser]

Not sure. I think I copied it from somewhere? I'll change it.

[fblackburn1]

Added mergeit label in the release process
But I don't approve this PR
The usage (for now) is only in python client which should never depend of this library (even only for typing)

[DrPyser]

Added mergeit label in the release process But I don't approve this PR The usage (for now) is only in python client which should never depend of this library (even only for typing)

We already discussed this previously, sorry that wasn't understood then.

Why is it bad for the clients to depend on xivo-lib-python?

I suppose we should duplicate this module in wazo-lib-rest-client for usage in wazo python clients?

[fblackburn1]

Why is it bad for the clients to depend on xivo-lib-python?

xivo-lib-python is intended to be a server library, not to be installed outside of the server. We don't want to block ourself/packaging because we add a dependency on wazo python clients (ex: build debian based on requirements.txt, or move requirements.txt to setup.py/pyproject.toml)

Rule of thumb about lib-python, when it doesn't make sense or if the code is huge, make a separate module. IMHO we are in the first scenario

I suppose we should duplicate this module in wazo-lib-rest-client for usage in wazo python clients?

Yes it's the easier solution (and probably the one that is worth it)
Another solution for this specific library, is to make a generic typing library + debian packaging (if the python library already exist, then it's to make debian packaging). To be determined if it worth it ...