.

code_eval.py

@@ -194,6 +194,7 @@ class CodeEval(evaluate.Metric):
                         test_program = body + "\n" + test
                         args = (test_program, timeout, tid, pid)
                         future = executor.submit(check_correctness, *args)
+                        result.test_program.append(test_program)
                         result.add(future)
                     results[tid].append(result)
 
@@ -262,7 +263,8 @@ def estimate_pass_at_k(num_samples, num_correct, k):
 class Result(BaseModel):
     task_id: str
     prediction_id: int
-
+    
+    test_program: List[str] = []
     passed: Optional[bool] = None
     result: List[str] = []
     futures: List[object] = []