Tidy up the two workflow method configs

[rcannood]

Describe your changes

Stacked on #37 -- please merge that one first, the diff here is the last two commits.

Two small inconsistencies between our two multi-component methods.

novel merges the wrong api spec. simple_mlp merges wf_method.yaml, which exists precisely for "workflows that combine multiple components"; novel merges comp_method.yaml. The practical difference is that comp_method.yaml pulls in run_and_check_output.py and the test resource directory, neither of which means anything for a component with only a nextflow runner and no script of its own.

simple_mlp passes a key that doesn't exist.

fromState: ["input_train_mod2", "input_test_mod1", "input_model", "input_transform"],

input_transform is not an argument of simple_mlp_predict (nor of comp_method_predict.yaml) and is never put in the state. _processFromState silently drops keys that aren't required inputs, so it's been harmless -- but it reads like something is being threaded through when nothing is.

After the change novel resolves to test_resources: ['check_config.py'] and keeps the label added in #37.

Part of a series of PRs coming out of a pre-run review of the benchmark.

Checklist before requesting a review

• I have performed a self-review of my code

• Check the correct box. Does this PR contain:

  • Breaking changes
  • New functionality
  • Major changes
  • Minor changes
  • Bug fixes

• Proposed changes are described in the CHANGELOG.md

• CI Tests succeed and look good!