feature #2253 tests for run_util, 'Usage' bugfix

[John-Sharples]

1. Note this PR includes a small change to source code in metplus/util/run_util.py. This change makes the check for "Usage" work as intended. get_process_list always returns a list of tuples, hence the current check 'Usage' in process_list can never evaluate to True. Change unpacks the tuples before checking. This

2. Adds tests for run_util.py

Pull Request Testing

• Describe testing already performed for these changes:
  

• Recommend testing for the reviewer(s) to perform, including the location of input datasets, and any additional instructions:
  
  Please review changes to source code.

• Do these changes include sufficient documentation updates, ensuring that no errors or warnings exist in the build of the documentation?
  Yes

• Do these changes include sufficient testing updates?
  Yes

• Will this PR result in changes to the test suite?
  Yes - more tests will run for run_util

• Please complete this pull request review by [Fill in date].
  

Pull Request Checklist

See the METplus Workflow for details.

• Add any new Python packages to the METplus Components Python Requirements table.
• Review the source issue metadata (required labels, projects, and milestone).
• Complete the PR definition above.
• Ensure the PR title matches the feature or bugfix branch name.
• Define the PR metadata, as permissions allow.
  Select: @georgemccabe
  Select:
  Select:
• After submitting the PR, select the ⚙️ icon in the Development section of the right hand sidebar. Search for the issue that this PR will close and select it, if it is not already selected.
• After the PR is approved, merge your changes. If permissions do not allow this, request that the reviewer do the merge.
• Close the linked issue and delete your feature or bugfix branch from GitHub.

[georgemccabe]

I think this assert always passes because a default value is specified. Should this be:

Suggested change

	assert actual.get('user_env_vars', 'GODS_OF_WEATHER', 'Indra_Thor_Zeus')
	assert actual.get('user_env_vars', 'GODS_OF_WEATHER') == 'Indra_Thor_Zeus'

[John-Sharples]

Good pick up, thanks. Fixed in latest commit.

[georgemccabe]

Is this supposed to include an assert? If so, would it make sense to change the 2nd check_err value to an empty string instead of None, then move the assert in the if statement after the if/else statement and change it to:

assert check_err == err

[John-Sharples]

Oops, a bit sloppy from me here, sorry. Fixed in latest commit.

[georgemccabe]

These changes look good besides a couple in-line suggestions for improvements.

I had not used the mock or capfd tools before but I am excited to use them now that I am familiar! Many of the existing tests rely on empty test files that are stored in the repo under internal/tests/data that could be removed if these tools were used instead. Thanks for bringing these tools to my attention!

I previously noticed that there are a few util functions that call sys.exit when they really should return a value so the entire calling script that uses these functions does not exit when something goes wrong with configuration. I will keep these tests in mind when making those changes since they expect that sys.exit is called.

Thank you for finding and fixing the bug checking for the Usage wrapper!

Please take a look at the in-line suggestions. I approve once these are resolved.

[John-Sharples]

Thanks for the review and for pickup up those assert bugs @georgemccabe

capfd is very useful, especially when a function has multiple paths that log an error and return None. It provides a way to check the correct message is logged for each error type. However, it relies on the message bubbling all the way up to stdout or stderr, which is fine for MET/METplus, but could (in theory) cause issues if these are suppressed by the application.

My other open PR #2320 provides a better solution for checking the logs when dealing with a config object, by replacing config.logger with a MagicMock object (I wish I'd thought of this sooner!). Some of the tests here could be updated to use that method instead of capfd, if/when that PR goes in.

I will keep these tests in mind when making those changes since they expect that sys.exit is called

One of the "features" of high test coverage is changes to source always breaks tests :)