Generalize planemo test to cover testing more kinds of artifacts - in particular Galaxy workflows and CWL tools and workflows.

• Develop a generalize test format external to the artifact under test (e.g. a <artifact>-test.yml file for instance).
• Refactor much of Galaxy's testing comparison stuff out into galaxy-lib or a new testing package for reuse in planemo.
• Generalize the concept of importing a thing, running a test on thing, collecting named outputs, and inspecting the named outputs against tests and move logic out of TwillTestCase in Galaxy.
• Extend to workflows.
• Extend to other non-Galaxy backends (cwltool for sure and possibly Toil or Arvados).
• Rework testing backend to consume only structured json produced by newer testing engine (instead of json + xunit). f8b6ee1
• Implement XUnit reports for newer testing stuff.