Tribunal.Assertions (Tribunal v1.4.0)

Copy Markdown View Source

Assertion evaluation engine.

Routes assertions to the appropriate implementation:

  • Deterministic: contains, regex, is_json, etc.
  • Judge (requires req_llm): faithful, relevant, etc.
  • Embedding (requires alike): similar

Summary

Functions

Checks if all assertions passed.

Returns list of available assertion types based on loaded dependencies.

Evaluates a single assertion against a test case.

Evaluates multiple assertions against a test case.

Functions

all_passed?(results)

Checks if all assertions passed.

available()

Returns list of available assertion types based on loaded dependencies.

evaluate(assertion_type, test_case, opts \\ [])

Evaluates a single assertion against a test case.

Returns {:pass, details} or {:fail, details}.

evaluate_all(assertions, test_case)

Evaluates multiple assertions against a test case.

Returns a map of %{assertion_type => result}.