Utilize yamllint severity levels for rule severity

[FloSchwalm]

yamllint allows to categorize found problems as "error" or "warning"
ansible-lint treats them all as having VERY_LOW severity anyway

I think that the rule categorization of yamllint shouldn't be lost, after all it can be configured with a .yamllint config file to fit the user's needs.

This PR keeps the VERY_LOW severity categorization for problems that yamllint reports as "warning" just as before, but uses MEDIUM severity for problems reported as "error".
Syntax errors reported by yamllint are raised to VERY_HIGH severity, though they won't be used as long as ansible-lint skips further rule checking after its own syntax check.

[ssbarnea]

Ideally we should have included some tests.

[FloSchwalm]

Ideally we should have included some tests.

I guess I can try to add some. I'm no python developer, but I think I can figure it out from the existing tests.
What kind of test were you thinking of? Hardcoded ansible-lint output lines which are then tested if categorized correctly? Or would you rather include a test playbook and actually run ansible-lint in the test?

[ssbarnea]

Create one extra file inside examples/playbooks that would raise all 3 types of yamllint severity issues when you lint it. After this you can add one extra test that verifies that the output is the expected one.

lookat def test_example_custom_module and copy it to test_example_yamllint_severity, test for number of failures reported and their content (especially severity levels). Print the result to see what it contains, or add a breakpoint() and run pytest -k test_example_yamllint_severity once you add it. Once inside the debugger you will be able to inspect what result contains.

To save your time: result is a List[MatchError], so you will know how to assert each entry has the expected severity level.