Added better test messages for failures

app.py

@@ -97,16 +97,19 @@ def add_new_eval(
     )
 
     # Compute score
-    file_path = path_to_file.name
+    file_path = path_to_file.name        
     scores = {"all": 0, 1: 0, 2: 0, 3: 0}
     num_questions = {"all": 0, 1: 0, 2: 0, 3: 0}
     with open(f"scored/{organisation}_{model}.jsonl", "w") as scored_file:
         with open(file_path, 'r') as f:
-            for line in f:
-                task = json.loads(line)
+            for ix, line in enumerate(f):
+                try:
+                    task = json.loads(line)
+                except Exception:
+                    return format_error(f"Line {ix} is incorrectly formatted. Please fix it and resubmit your file.")
 
                 if "model_answer" not in task:
-                    raise Exception("No model_answer key in the file provided")
+                    raise format_error(f"Line {ix} contains no model_answer key. Please fix it and resubmit your file.")
                 answer = task["model_answer"]
                 task_id = task["task_id"]
                 try: