From 39691e0e3ce1ea96bcde39735d9428389325f79d Mon Sep 17 00:00:00 2001
From: Alex Ward <alxwrd@googlemail.com>
Date: Tue, 14 Apr 2026 15:15:27 +0100
Subject: [PATCH 1/2] allow evaluate() to accept multiple cases to allow
 result.evaluate(*case.evaluators) pattern

Co-Authored-By: Claude <claude[bot]@users.noreply.github.com>
---
 src/pyeval/_core.py                  | 65 ++++++++++++++--------------
 tests/evals/eval_extra_evaluators.py | 20 +++++++++
 2 files changed, 53 insertions(+), 32 deletions(-)
 create mode 100644 tests/evals/eval_extra_evaluators.py

diff --git a/src/pyeval/_core.py b/src/pyeval/_core.py
index c878693..06dfa89 100644
--- a/src/pyeval/_core.py
+++ b/src/pyeval/_core.py
@@ -69,48 +69,49 @@ def duration(self) -> float:
         """How long the task took to run, in seconds."""
         return self.ctx.duration
 
-    def evaluate(self, evaluator: Evaluator) -> None:
+    def evaluate(self, *evaluators: Evaluator) -> None:
         results = _CURRENT_EVAL_RESULTS.get()
         if results is None:
             raise RuntimeError(
                 "result.evaluate() called outside of an eval context. "
                 "Did you call this eval function directly instead of running it via pytest?"
             )
-        try:
-            raw = evaluator.evaluate_sync(self.ctx)
-            normalized = _EVALUATOR_OUTPUT_ADAPTER.validate_python(raw)
-            default_name = evaluator.get_default_evaluation_name()
-            mapping = (
-                normalized
-                if isinstance(normalized, Mapping)
-                else {default_name: normalized}
-            )
-            for name, result in mapping.items():
-                if not isinstance(name, str):
-                    name = str(name)
-                if not isinstance(result, EvaluationReason):
-                    result = EvaluationReason(
-                        value=result
-                        if isinstance(result, bool | int | float | str)
-                        else str(result)
+        for evaluator in evaluators:
+            try:
+                raw = evaluator.evaluate_sync(self.ctx)
+                normalized = _EVALUATOR_OUTPUT_ADAPTER.validate_python(raw)
+                default_name = evaluator.get_default_evaluation_name()
+                mapping = (
+                    normalized
+                    if isinstance(normalized, Mapping)
+                    else {default_name: normalized}
+                )
+                for name, result in mapping.items():
+                    if not isinstance(name, str):
+                        name = str(name)
+                    if not isinstance(result, EvaluationReason):
+                        result = EvaluationReason(
+                            value=result
+                            if isinstance(result, bool | int | float | str)
+                            else str(result)
+                        )
+                    results.append(
+                        EvaluationResult(
+                            name=name,
+                            value=result.value,
+                            reason=result.reason,
+                            source=evaluator.as_spec(),
+                        )
                     )
-                results.append(
-                    EvaluationResult(
-                        name=name,
-                        value=result.value,
-                        reason=result.reason,
+            except Exception as e:
+                self.failures.append(
+                    EvaluatorFailure(
+                        name=evaluator.get_default_evaluation_name(),
+                        error_message=f"{type(e).__name__}: {e}",
+                        error_stacktrace=traceback.format_exc(),
                         source=evaluator.as_spec(),
                     )
                 )
-        except Exception as e:
-            self.failures.append(
-                EvaluatorFailure(
-                    name=evaluator.get_default_evaluation_name(),
-                    error_message=f"{type(e).__name__}: {e}",
-                    error_stacktrace=traceback.format_exc(),
-                    source=evaluator.as_spec(),
-                )
-            )
 
 
 def _group_by_type(
diff --git a/tests/evals/eval_extra_evaluators.py b/tests/evals/eval_extra_evaluators.py
new file mode 100644
index 0000000..7bb3b72
--- /dev/null
+++ b/tests/evals/eval_extra_evaluators.py
@@ -0,0 +1,20 @@
+"""Regression test: case.evaluators can be unpacked into result.evaluate()."""
+
+from pyeval import Case, dataset, execute
+from pyeval.evaluators import EqualsExpected, IsInstance, MaxDuration
+
+
+@dataset(
+    Case(name="basic", inputs="hello", expected_output="hello"),
+    Case(
+        name="with_extra_evaluators",
+        inputs="hello",
+        expected_output="hello",
+        evaluators=[MaxDuration(seconds=1)],
+    ),
+)
+def eval_case_evaluators(case: Case):
+    result = execute(lambda x: x, case)
+    result.evaluate(EqualsExpected())
+    result.evaluate(IsInstance(type_name="str"))
+    result.evaluate(*case.evaluators)

From a4b05c8af6c6ee0c618b3721083eb07f38361a13 Mon Sep 17 00:00:00 2001
From: Alex Ward <alxwrd@googlemail.com>
Date: Tue, 14 Apr 2026 15:47:10 +0100
Subject: [PATCH 2/2] fix: use tuple for extra evaluators

Co-Authored-By: Claude <claude[bot]@users.noreply.github.com>
---
 tests/evals/eval_extra_evaluators.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tests/evals/eval_extra_evaluators.py b/tests/evals/eval_extra_evaluators.py
index 7bb3b72..a8f2c3e 100644
--- a/tests/evals/eval_extra_evaluators.py
+++ b/tests/evals/eval_extra_evaluators.py
@@ -10,7 +10,7 @@
         name="with_extra_evaluators",
         inputs="hello",
         expected_output="hello",
-        evaluators=[MaxDuration(seconds=1)],
+        evaluators=(MaxDuration(seconds=1),),
     ),
 )
 def eval_case_evaluators(case: Case):