Epic: improve testing framework

[forsyth2]

Request criteria

• I searched the zppy GitHub Discussions to find a similar question and didn't find it.
• I searched the zppy documentation.
• This issue does not match the other templates (i.e., it is not a bug report, documentation request, feature request, or a question.)

Issue description

zppy testing could benefit from a number of improvements.

Currently open "Testing" issues, ranked by priority:

• Make shorter tests #517
• Avoid testing duplication for packages zppy uses #537
• Address "flaky" ts subtask in complete_run test #549. Supposedly simple fix -- will need to test.
• Ignore unique ID in bundles test #530. This is annoying but not a huge problem. I just have to expect that the test will always fail because the unique_id is different for each test run. In theory, simple fix.
• Replace image diff tests with metrics tests #536
• Convert tests to use v3 #552. I anticipate this being a challenge. All the expected files/images will need to be changed. Perhaps this could be paired with Replace image diff tests with metrics tests #536.
• Further automate tests #516. This is a nice-to-have. The testing-for-release process is tedious, but there are now directions (https://e3sm-project.github.io/zppy/_build/html/main/dev_guide/release_testing.html). That said, it is easy to miss something when going through the steps manually.
• test_bash_generation.py doesn't inspect input.txt #104. This isn't strictly necessary, but would somewhat improve test quality.
• Save previous expected files for testing #93. This may no longer be relevant since we have directions to generate official expected results for each release. E.g., https://github.com/E3SM-Project/zppy/blob/main/tests/integration/generated/directions_chrysalis.md#commands-to-generate-official-expected-results-for-a-release

The following would be made irrelevant by #536, since they involve changing how the image testing is done:

• Fix test update commands #540. Simple fix.
• Clarify simple_image_name #358