Implement some 'integration' tests

[alexpovel]

• Create one reference, fully featured JSON resume
• Create derived resumes with all possible fields (all combinations thereof?) set to None where legal (given the schema)
• Compare all these to expected files (the files contain the ANSI escape characters literally)
• Load tests?

[alexpovel]

By now, we have integration tests (if they can be called that?) for both the CLI (which by extension tests a lot of underlying stuff) and both the API as well as single-file servers (through d645f80), which test the entire stack as well. Together, and with #7, c. 95% coverage is achieved. Some edge cases remain, like these two exceptions:

ancv/ancv/web/client.py

Lines 77 to 82 in 4d5a7ef

	except json.decoder.JSONDecodeError:
	raise ResumeLookupError("Got malformed JSON.")
	except ValidationError:
	raise ResumeLookupError(
	"Got legal JSON but wrong schema (cf. https://jsonresume.org/schema/)"
	)

These are hard to test without mocking the GitHub API or somehow providing actual gists of actual users that would trigger these code paths.

[alexpovel]

The above is now covered by 476bf58.

[alexpovel]

Not doing load testing for now. Automatic scaling takes care of that for us now. Plus, a single instance of the app is probably able to handle tons of requests, since it's async-first and templating/rendering is cheap.