simplified the logic

Author: SYED-M-HUSSAIN

deepeval/benchmarks/gsm8k/gsm8k.py

@@ -132,21 +132,6 @@ def _extract_prediction_from_response(self, res) -> str:
         elif isinstance(res, tuple):
             return self._extract_from_tuple(res)
 
-        # Case 3: Response is a string
-        elif isinstance(res, str):
-            return res
-        
-        # Case 4: Response has .text or .content attribute
-        elif hasattr(res, 'text'):
-            return str(res.text)
-        elif hasattr(res, 'content'):
-            return str(res.content)
-        
-        # Case 5: Response is a dict
-        elif isinstance(res, dict):
-            return self._extract_from_dict(res)
-        
-        # Case 6: Fallback - convert to string
         else:
             return str(res)
 
@@ -155,37 +140,8 @@ def _extract_from_tuple(self, res: tuple) -> str:
         if len(res) == 0:
             return ""
         first_elem = res[0]
-        
         if hasattr(first_elem, 'answer'):
             return str(first_elem.answer)
-        elif hasattr(first_elem, 'text'):
-            return str(first_elem.text)
-        elif hasattr(first_elem, 'content'):
-            return str(first_elem.content)
-        elif len(res) > 1 and res[1] is not None:
-            second_elem = res[1]
-            if hasattr(second_elem, 'answer'):
-                return str(second_elem.answer)
-            else:
-                return str(second_elem)
-        
-        else:
-            return str(first_elem)
-    
-    def _extract_from_dict(self, res: dict) -> str:
-        """Extract prediction from dictionary response."""
-        # Common keys that might contain the answer
-        possible_keys = ['answer', 'text', 'content', 'response', 'output', 'result']
-        
-        for key in possible_keys:
-            if key in res:
-                return str(res[key])
-
-        for value in res.values():
-            if value is not None and str(value).strip():
-                return str(value)
-        
-        return str(res)
 
     def load_benchmark_dataset(self) -> List[Golden]:
         from datasets import load_dataset