feat(agent-zero): uvicorn A2A server calling LiteLLM directly

agent-zero-zulip/src/agent_zero_zulip/a2a_server.py

@@ -1,24 +1,24 @@
-"""A2A Agent for kagentz — processes messages via Agent Zero core directly."""
-import sys, os, asyncio, json, uuid, time
-
-sys.path.insert(0, "/a0")
-os.environ["A2A_TOKEN"] = "8zNgdOEXzYxjQvTl"
-
+"""A2A agent for kagentz — uvicorn-based, calls LiteLLM for responses."""
+import os, json, uuid, threading, asyncio
 import uvicorn
 from fastapi import FastAPI, Request
-from agent import AgentContext, UserMessage, AgentContextType
+from fastapi.responses import JSONResponse
+import httpx
 
 app = FastAPI()
-AGENT_TOKEN = os.environ["A2A_TOKEN"]
+task_results = {}
+LITELLM_URL = "https://litellm.sysloggh.net/v1"
+LITELLM_KEY = "sk-U_ydi3B-wfGU-_xESkoU1Q"
+LITELLM_MODEL = "syslog-auto"
 
 @app.get("/.well-known/agent.json")
 async def agent_card():
     return {
         "name": "kagentz",
         "version": "1.0.0",
-        "description": "Agent Zero instance managed via Zulip",
-        "capabilities": {"a2a": {"version": "1.0", "authentication": ["bearer"]}},
-        "skills": [{"id": "chat", "name": "Chat", "description": "General conversation via Zulip"}],
+        "description": "Agent Zero on CT 105 via Zulip + LiteLLM",
+        "capabilities": {"a2a": {"version": "1.0"}},
+        "skills": [{"id": "chat", "name": "Chat"}],
     }
 
 @app.post("/a2a")
@@ -31,17 +31,13 @@ async def a2a_endpoint(request: Request):
         msg = params.get("message", {})
         parts = msg.get("parts", [])
         text = " ".join(p.get("text", "") for p in parts if isinstance(p, dict))
-        
         task_id = f"az-{uuid.uuid4().hex[:12]}"
-        asyncio.create_task(process_via_agent_zero(task_id, text))
-        
+        asyncio.create_task(call_litellm(task_id, text))
         return {"jsonrpc": "2.0", "result": {"id": task_id, "status": {"state": "working"}}}
     
     elif method == "tasks/get":
-        params = body.get("params", {})
-        task_id = params.get("id", "")
+        task_id = body.get("params", {}).get("id", "")
         status, result = task_results.get(task_id, ("pending", None))
-        
         resp = {"jsonrpc": "2.0", "result": {"id": task_id, "status": {"state": status}}}
         if status == "completed" and result:
             resp["result"]["history"] = [
@@ -49,44 +45,31 @@ async def a2a_endpoint(request: Request):
                 {"role": "assistant", "parts": [{"text": result.get("output", "")}]},
             ]
         elif status == "failed":
-            resp["result"]["status"]["message"] = {"text": str(result or "Unknown error")}
-        
+            resp["result"]["status"]["message"] = {"text": str(result)}
         return resp
     
     return {"jsonrpc": "2.0", "error": {"code": -32601, "message": "Method not found"}}
 
-task_results = {}
-
-async def process_via_agent_zero(task_id: str, message: str):
-    """Process a message through Agent Zero's core."""
+async def call_litellm(task_id, message):
     try:
-        # Create a new context and send the message
-        ctx = AgentContext.get(uuid.uuid4().hex)
-        
-        # Run in executor to avoid blocking the event loop
-        loop = asyncio.get_event_loop()
-        
-        def _run():
-            try:
-                # Send user message and get response
-                msg = UserMessage(text=message, context=ctx)
-                result = ctx.agent.handle_message(msg)
-                
-                # Extract text response
-                if hasattr(result, 'content'):
-                    output = result.content
-                elif isinstance(result, str):
-                    output = result
-                else:
-                    output = str(result)
-                
-                return output
-            except Exception as e:
-                return f"Error: {e}"
-        
-        output = await loop.run_in_executor(None, _run)
+        async with httpx.AsyncClient(timeout=120.0) as client:
+            resp = await client.post(
+                f"{LITELLM_URL}/chat/completions",
+                headers={"Authorization": f"Bearer {LITELLM_KEY}", "Content-Type": "application/json"},
+                json={
+                    "model": LITELLM_MODEL,
+                    "messages": [
+                        {"role": "system", "content": "You are kagentz, an Agent Zero instance on CT 105. You communicate via Zulip. Keep responses concise."},
+                        {"role": "user", "content": message},
+                    ],
+                    "max_tokens": 2000,
+                },
+            )
+            if resp.is_success:
+                output = resp.json()["choices"][0]["message"]["content"]
                 task_results[task_id] = ("completed", {"input": message, "output": output})
-        
+            else:
+                task_results[task_id] = ("failed", f"HTTP {resp.status_code}")
     except Exception as e:
         task_results[task_id] = ("failed", str(e))