feat: refactor run_multienv_qwen25vl.py and qwen25vl_agent.py for improved logging and task management

- Introduced signal handling for graceful shutdown of environments and processes.
- Enhanced logging configuration to support dynamic log levels and structured output.
- Updated argument parsing to include new parameters for model selection and task execution.
- Refactored task distribution logic to streamline environment task management.
- Improved error handling during task execution and environment cleanup.
- Adjusted Qwen25VLAgent initialization to support new model and thought prefix options.
- Reduced max tries for LLM calls to optimize performance.

mm_agents/qwen25vl_agent.py

@@ -66,25 +66,24 @@ class Qwen25VLAgent:
     def __init__(
         self,
         platform="ubuntu",
-        planner_model="gpt-4o",
-        executor_model="qwen2.5vl",
+        model="qwen2.5-vl-72b-instruct",
         max_tokens=1500,
         top_p=0.9,
         temperature=0.5,
         action_space="pyautogui",
         observation_type="screenshot",
         history_n=4,  # Number of previous interactions to include in full detail
+        add_thought_prefix=False,
     ):
         self.platform = platform
-        self.planner_model = planner_model
-        self.executor_model = executor_model
-        assert self.executor_model is not None, "Executor model cannot be None"
+        self.model = model
         self.max_tokens = max_tokens
         self.top_p = top_p
         self.temperature = temperature
         self.action_space = action_space
         self.observation_type = observation_type
         self.history_n = history_n  # Control how many previous interactions to include
+        self.add_thought_prefix = add_thought_prefix
         assert action_space in ["pyautogui"], "Invalid action space"
         assert observation_type in ["screenshot"], "Invalid observation type"
         self.thoughts = []
@@ -277,19 +276,20 @@ Previous actions:
             })
 
         # append_text = f"""Step {current_step+1}: Thought:"""
-        append_text = f"""Thought:"""
-        messages.append({"role": "assistant", "content": [{"type": "text", "text": append_text}]})
+        if self.add_thought_prefix:
+            append_text = f"""Thought:"""
+            messages.append({"role": "assistant", "content": [{"type": "text", "text": append_text}]})
 
         # Call the LLM
         response = self.call_llm(
             {
-                "model": self.executor_model,
+                "model": self.model,
                 "messages": messages,
                 "max_tokens": self.max_tokens,
                 "top_p": self.top_p,
                 "temperature": self.temperature,
             },
-            self.executor_model,
+            self.model,
         )
 
         logger.info(f"Qwen25VL Output: {response}")
@@ -483,10 +483,10 @@ Previous actions:
                 continue
             
             # Handle lines inside tool call markers
-            if line.startswith("<tool_call>"):
+            if line.startswith("<tool_call>") or line.startswith("⚗") or line.startswith("📐"): # Yeah, it's a bug during data processing
                 inside_tool_call = True
                 continue
-            elif line.startswith("</tool_call>"):
+            elif line.startswith("</tool_call>") or line.startswith("⚗") or line.startswith("📐"): # Yeah, it's a bug during data processing
                 if current_tool_call:
                     # Process the collected tool call
                     process_tool_call("\n".join(current_tool_call))
@@ -540,12 +540,13 @@ Previous actions:
             # todo: check
         ),
         interval=30,
-        max_tries=10,
+        max_tries=5,
     )
     def call_llm(self, payload, model):
         messages = payload["messages"]
-        base_url = "your_base_url"
-        api_key = "your_api_key"
+
+        base_url = os.getenv('DASHSCOPE_BASE_URL', "https://dashscope.aliyuncs.com/compatible-mode/v1")
+        api_key = os.getenv('DASHSCOPE_API_KEY', "sk-123")
         
         client = openai.OpenAI(
             base_url=base_url,