Autogen support

mas_arena/agents/autogen.py

@@ -0,0 +1,88 @@
+import os
+from typing import Dict, Any
+from openai import AsyncOpenAI
+from dotenv import load_dotenv
+from mas_arena.agents.base import AgentSystem, AgentSystemRegistry
+
+load_dotenv()
+
+
+class AutoGen(AgentSystem):
+
+    def __init__(self, name: str = "autogen", config: Dict[str, Any] = None):
+        """Initialize the AutoGen System"""
+        super().__init__(name, config)
+        self.config = config or {}
+
+        self.model_name = self.config.get("model_name") or os.getenv("MODEL_NAME", "qwen-plus")
+
+        self.num_rounds = self.config.get("num_rounds", 5)
+
+        self.client = AsyncOpenAI(api_key=os.getenv("OPENAI_API_KEY"), base_url=os.getenv("OPENAI_API_BASE"))
+
+        self.agents = [
+            {
+                "name": "primary",
+                "system_prompt": """You are a helpful AI assistant, skilled at generating creative and accurate content."""
+            },
+            {
+                "name": "critic",
+                "system_prompt": "Provide constructive feedback on the content provided. Respond with 'APPROVE' when the content meets high standards or your feedback has been addressed."
+            }
+        ]
+
+    async def run_agent(self, problem: Dict[str, Any], **kwargs) -> Dict[str, Any]:
+
+        problem_text = problem["problem"]
+        messages = [
+            {"role": "user", "content": f"Problem: {problem_text}"}
+        ]
+        conversation_history = messages.copy()
+
+        all_messages = []
+        final_answer = ""
+
+        for _ in range(self.num_rounds):
+            for n, agent in enumerate(self.agents):
+                agent_name = agent["name"]
+                agent_prompt = agent["system_prompt"]
+
+                agent_messages = [
+                    {"role": "system", "content": agent_prompt},
+                    *conversation_history
+                ]
+
+                response = await self.client.chat.completions.create(
+                    model=self.model_name,
+                    messages=agent_messages
+                )
+
+                response_content = response.choices[0].message.content
+
+                ai_message = {
+                    'content': response_content,
+                    'name': agent_name,
+                    'role': 'assistant',
+                    'message_type': 'ai_response',
+                    'usage_metadata': response.usage
+                }
+
+                conversation_history.append({"role": "assistant", "content": response_content, "name": agent_name})
+
+                if (agent_name == "primary"):
+                    final_answer = ai_message["content"]
+
+                if agent_name == "critic" and "approve" in response_content.lower():
+                    return {
+                        "messages": all_messages,
+                        "final_answer": final_answer
+                    }
+                all_messages.append(ai_message)
+
+        return {
+            "messages": all_messages,
+            "final_answer": final_answer
+        }
+
+
+AgentSystemRegistry.register("autogen", AutoGen)