New Requirements, ability to run program locally

[Helithumper]

Solves #127

Initially was focused on getting requirements working with pip-compile, however had to add more changes to get the program to run due to alot of wrong variables and pathing.

make dev should begin the dev server

[jontyms]

Sorry to hear you had problems running the application, I just pushed new instructions to the readme. I also removed the ability to run the application directly from main.py.

[jontyms]

• Removing the app. from the imports breaks pytest.
• Why do we need pip-compile?
• Why do we need a Makefile?
• Can you split your config changes into a separate PR?

[Helithumper]

• Removing the app. from the imports breaks pytest.

This is because as currently configured the tests are in an entirely different module from the app. We can fix this by using relative modules and replacing each of the new imports to become .models for example.

• Why do we need pip-compile?

The piling up numbers of failing PRs from dependabot are due to no way to recompute the dependency chain. Dependabot tries to upgrade a version and ends up stuck due to circular or otherwise incorrect dependency versions. pip-compile allows us to actually fix the dependabot issues by just recomputing the latest version of requirements.txt from requirements.in

• Why do we need a Makefile?

Makes running the code easier. Provides macros for execution of the program as well as building the dependencies.

• Can you split your config changes into a separate PR?

I can do this, however it'll likely result in this PR not functioning as the current config isn't correct and was both named incorrectly as well as was missing parts. I began working on that in order to get anything running as the committed config wasn't up to date.

[Helithumper]

Test doesn't pass yet due to the test overall being bad and expecting a random file to be populated for openvpn without error handling and without gracefully handling the error.

[Helithumper]

Fixed test err

[jontyms]

This is because as currently configured the tests are in an entirely different module from the app. We can fix this by using relative modules and replacing each of the new imports to become .models for example.

I think the relative imports are messier, especially if you have to go multiple levels. I am not familiar with why you would do relative or absolute imports. Perhaps you could explain.

[jontyms]

I am open to trying pip-compile

[jontyms]

I am not a fan, if the config file it not found or misconfigured in prod, this could fail silently.

[Helithumper]

Then we can add the logic above to include a logging.warn like I did for keycloak

[Helithumper]

This is because as currently configured the tests are in an entirely different module from the app. We can fix this by using relative modules and replacing each of the new imports to become .models for example.

I think the relative imports are messier, especially if you have to go multiple levels. I am not familiar with why you would do relative or absolute imports. Perhaps you could explain.

We can do either. In this case I used relative imports due to pytest's dependency resolver. We can use absolute tests however we'll need to add config to include the package in Pytest's path or use python3 -m pytest for all tests.

Either way works

[jontyms]

Created #140 to address config file
working on other prs to address the other fixes in this pr