merge

Author: xzrderek

development/RUNNING_EVALUATIONS.md

@@ -33,7 +33,7 @@ def load_jsonl(file_path: str) -> List[Dict[str, Any]]:
                 row_id_index = stripped.find("row_id")
                 if row_id_index != -1:
                     row_id = re.search(r'"row_id": (.*),', stripped[row_id_index:])
-                    raise ValueError(f"{e.msg} at line {line_number}: {stripped} ({row_id})")
+                    raise ValueError(f"{e.msg} at line {line_number}: {stripped} ({row_id})") from e
                 raise e
     else:
         with open(file_path, "r", encoding="utf-8") as f:
@@ -50,6 +50,6 @@ def load_jsonl(file_path: str) -> List[Dict[str, Any]]:
                     row_id_index = line.find("row_id")
                     if row_id_index != -1:
                         row_id = re.search(r'"row_id": (.*),', line[row_id_index:])
-                        raise ValueError(f"{e.msg} at line {line_number}: {line} ({row_id})")
+                        raise ValueError(f"{e.msg} at line {line_number}: {line} ({row_id})") from e
                     raise e
     return data

eval_protocol/benchmarks/suites/test_evaluation.py

@@ -213,7 +213,7 @@ async def default_mcp_gym_rollout_processor(
     """
     if config.server_script_path is None:
         raise ValueError("server_script_path is required for default_mcp_gym_rollout_processor")
-    server = MCPServerManager(config.server_script_path, port=9701)
+    server = MCPServerManager(config.server_script_path, port=9700)
 
     try:
         server.start()

eval_protocol/common_utils.py

@@ -232,7 +232,7 @@ def __init__(
             os.path.join(os.path.dirname(os.path.dirname(os.path.dirname(__file__))), "vite-app", "dist")
         ),
         host: str = "localhost",
-        port: Optional[int] = 8001,
+        port: Optional[int] = 8000,
         index_file: str = "index.html",
     ):
         # Initialize WebSocket manager