Add random_labels negative control

[rcannood]

Describe your changes

• Add a random_labels negative control (predicts labels by sampling from the training label distribution) so the template demonstrates both sides of the control bracket, and register it in the run_benchmark workflow.
• process_dataset: --seed only seeded random, not np.random, so the random split method was not reproducible. Also treat seed 0 as a valid seed.
• accuracy: write metric_values as a list to match metric_ids, as the comment above it already suggested.

All three components pass viash test.

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!