Workspace Environment

[photosssa]

I've read workspace enviroment codes, some questions:

1. What's different between workspace's operation and enviroment's function; that says if I can splite work into operation list, and operation list will be executed with codes like:

   async def exec_op_list(self,oplist:List,agent_id:str)->tuple[List[str],bool]:
       result_str = "op list is none"
       if oplist is None:
           return None,False
   
       result_str = []
       have_error = False
       for op in oplist:
           if op["op"] == "create":
               await self.create(op["path"],op["content"])
           elif op["op"] == "write_file":
               is_append = op.get("is_append")
               if is_append is None:
                   is_append = False
               error_str = await self.write(op["path"],op["content"],is_append)
           elif op["op"] == "delete":
               error_str = await self.delete(op["path"])
           elif op["op"] == "rename":
               error_str = await self.rename(op["path"],op["new_name"])
           elif op["op"] == "mkdir":
               error_str = await self.mkdir(op["path"])

   it's totally same as what function call does: dumps function name and paramters into todo list,

2. And more, who determines what operation a agent can do, if with enviroment's function list, enviroments defines what functions agent can call; but in workspace, it's hardcode in workspace implement, and i can't find how to make agent know what operations supported by workspace do.

[waterflier]

Thanks for your suggestions, this part of the new implementation has been basically completely refactored.

From the implement side, there is no need to distinguish between function and action. For the Agent, function means at least two LLMs, while action can be completed with one LLM in many scenarios.