Custom providers: trigger evals even if yaml config file hasn't changed

[tleyden]

I am using a custom provider as follows:

prompts:
  - "dummy_prompt_ignored"  # Dummy prompts, not even used for anything

providers:
  - "file://my_custom_provider.py"

but when I change code in my_custom_provider.py or dependent code, the github action doesn't re-run the evals. Instead it gives this message:

No LLM prompt or config files were modified.

I can trigger it by manually updating the config file with a dummy change. Is there a config option to have it run no matter what?

PS: promptfoo is an awesome tool!