#DI-95 implement tox for vestapol tests

[EvavW]

Ticket: #DI-95
Run tests in a reliable environment with tox.
Notes:

• some small changes were made in the codebase based on feedback from flake8 which now runs as part of tox testing
• please add comments with suggestions for which versions of Python we should be testing with tox. keep in mind all developers will have to have these installed
• please add comments with your opinion: should we add pylintor mypy to the tox tests? would require some overhead in terms of configuration and updates to the codebase

[waligob]

Something I was wondering: if poetry is creating a virtual env for you when you run poetry install, how can I select this environment in VSCode? This was apparently the most requested feature in VSCode's GitHub repo, and rolled out last May, but I couldn't get it to work. So instead I ran poetry config virtualenvs.in-project true, which then creates a .venv in my project's root directory, and VSCode automatically detects and activates it.

Not sure if you found a more straightforward way of getting VSCode and poetry environments to play nice.

[EvavW]

@waligob re:poetry virtualenv with vscode

will add this to the README. the solution here worked for me: microsoft/vscode-python#8372 (comment) - you can then select the poetry env from the list of python interpreters

[waligob]

the solution here worked for me: microsoft/vscode-python#8372 (comment) - you can then select the poetry env from the list of python interpreters

Nice! Worked for me too. Thanks.

[waligob]

please add comments with your opinion: should we add pylint or mypy to the tox tests? would require some overhead in terms of configuration and updates to the codebase

I think adding a linter would be really valuable for producing consistent, readable, and reliable code. I don't have strong opinions on our choice of linter—mypy's static type checking is intriguing given that there are at least a few implicit type constraints in the code (e.g. the json writers). We're using flake8 in our tox.ini—would that also be in contention for our linter solution?

At any rate, given the overhead in implementing the linter, I'd suggest that we create a new ticket for it.

[EvavW]

please add comments with your opinion: should we add pylint or mypy to the tox tests? would require some overhead in terms of configuration and updates to the codebase

I think adding a linter would be really valuable for producing consistent, readable, and reliable code. I don't have strong opinions on our choice of linter—mypy's static type checking is intriguing given that there are at least a few implicit type constraints in the code (e.g. the json writers). We're using flake8 in our tox.ini—would that also be in contention for our linter solution?

At any rate, given the overhead in implementing the linter, I'd suggest that we create a new ticket for it.

I'm leaning toward mypy as well. I'm most interested in the type checking so I wouldn't really consider flake8 as a contender, and I don't think it will compete too much with mypy. Agree that it should be a separate ticket (really 2 tickets for both vestapol and inq-data-resources).