Update backend/model_router.py

backend/model_router.py

@@ -1,19 +1,20 @@
-import os, asyncio, random
+import os, random, aiohttp
 from typing import List, Dict, AsyncGenerator
-from groq import Groq
+
+# Correct imports for groq 0.3.0+
+from groq import Client as Groq
 from openai import AsyncOpenAI
 
 groq_keys = os.environ.get("GROQ_API_KEYS", "").split(",")
 openrouter_keys = os.environ.get("OPENROUTER_KEYS", "").split(",")
 
-# Free models via Pollinations (no key needed)
 POLLINATIONS_URL = "https://text.pollinations.ai/"
 
 async def route_chat(model: str, messages: List[Dict]) -> AsyncGenerator[str, None]:
-    if model.startswith("groq"):
+    if model.startswith("groq-"):
         key = random.choice(groq_keys)
         client = Groq(api_key=key)
-        model_id = model.split("-",1)[1]  # e.g., "groq-mixtral-8x7b-32768"
+        model_id = model[5:]   # strip "groq-"
         stream = client.chat.completions.create(
             messages=messages,
             model=model_id,
@@ -22,13 +23,13 @@ async def route_chat(model: str, messages: List[Dict]) -> AsyncGenerator[str, No
         for chunk in stream:
             if chunk.choices[0].delta.content:
                 yield chunk.choices[0].delta.content
-    elif model.startswith("openrouter"):
+    elif model.startswith("openrouter-"):
         key = random.choice(openrouter_keys)
         client = AsyncOpenAI(
             base_url="https://openrouter.ai/api/v1",
             api_key=key,
         )
-        model_id = model.split("-",1)[1]  # e.g., "openrouter-anthropic/claude-3-opus"
+        model_id = model[11:]  # strip "openrouter-"
         stream = await client.chat.completions.create(
             messages=messages,
             model=model_id,
@@ -38,9 +39,7 @@ async def route_chat(model: str, messages: List[Dict]) -> AsyncGenerator[str, No
             if chunk.choices[0].delta.content:
                 yield chunk.choices[0].delta.content
     elif model == "free-pollinations":
-        # Free text generation via Pollinations
         prompt = "\n".join([f"{m['role']}: {m['content']}" for m in messages])
-        import aiohttp
         async with aiohttp.ClientSession() as session:
             async with session.post(POLLINATIONS_URL, json={"prompt": prompt, "model": "openai"}, params={"stream": "true"}) as resp:
                 async for line in resp.content:
@@ -50,13 +49,11 @@ async def route_chat(model: str, messages: List[Dict]) -> AsyncGenerator[str, No
         raise ValueError(f"Unknown model: {model}")
 
 async def route_autocomplete(model: str, prefix: str, suffix: str, max_tokens: int) -> str:
-    # For autocomplete, use Groq with FIM (fill-in-middle) if supported, else just completion.
     key = random.choice(groq_keys)
     client = Groq(api_key=key)
-    # Format for code completion
     prompt = f"<fim_prefix>{prefix}<fim_suffix>{suffix}<fim_middle>"
     response = client.completions.create(
-        model="codellama-70b",  # or "llama3-8b"
+        model="codellama-70b",
         prompt=prompt,
         max_tokens=max_tokens,
         temperature=0.1,