Use the semver package to parse the log version in testing

[mcdonnnj]

🗣 Description

This pull request moves the semver package to a testing requirement and uses it to parse the versions in the test_log_version() test.

💭 Motivation and context

Currently the version comparison relies on explicit string matching. However, I think that this should be more contextually aware. If an underlying Python package is one version and the Docker package (which is in lockstep version-wise) has a build identifier then the versions will not match. However, if a build identifier is the only differentiation between the versions I believe they should be considered equivalent. An example is that 1.0.0 and 1.0.0+build.1 should be seen as equivalent versions instead of a mismatch. Using semver to parse the versions will result in these versions being seen as equivalent. Some example version comparisons:

>>> from semver import parse_version_info
>>> old = parse_version_info("1.0.0")
>>> new = parse_version_info("1.0.0+build.1")
>>> old == new
True
>>> new = parse_version_info("1.0.0-rc.1")
>>> old == new
False

Here we see that only versions with build identifiers are considered equivalent. I think this is a reasonable adjustment of how this logic compares versions when it comes to underlying dependencies.

🧪 Testing

Automated tests pass.

✅ Pre-approval checklist

• This PR has an informative and human-readable title.
• Changes are limited to a single goal - eschew scope creep!
• All relevant type-of-change labels have been added.
• I have read the CONTRIBUTING document.
• These code changes follow cisagov code standards.
• Tests have been added and/or modified to cover the changes in this PR.
• All new and existing tests pass.

[dav3r]

Makes sense to me. 👍