Update app.py

app.py

@@ -1,61 +1,58 @@
 from fastapi import FastAPI, Request, HTTPException, Depends
 from fastapi.middleware.cors import CORSMiddleware
-from fastapi.security import HTTPBearer, HTTPAuthorizationCredentials
-from fastapi.responses import JSONResponse
 from llama_cpp import Llama
 import uvicorn
 
 app = FastAPI()
 
-# Allow all origins (for frontend access)
+# --- CORS SETTINGS (Crucial for GitHub Pages) ---
 app.add_middleware(
     CORSMiddleware,
     allow_origins=["*"],
+    allow_credentials=True,
     allow_methods=["*"],
     allow_headers=["*"],
 )
 
-# Simple API key auth
-security = HTTPBearer()
-MY_API_KEY = "my-secret-key-456"
+# --- CONFIGURATION ---
+MY_API_KEY = "my-secret-key-456" 
+# Optimized for HF Free Tier CPU
+llm = Llama(model_path="./model.gguf", n_ctx=2048, n_threads=4, n_gpu_layers=0)
 
-# Load GGUF model (CPU only, small threads for Spaces)
-llm = Llama(
-    model_path="./mpt-7b-chat.gguf",  # Make sure this is a tokenizer-included GGUF
-    n_ctx=2048,
-    n_threads=2,       # Reduce for free tier
-    n_gpu_layers=0     # Force CPU
-)
-
-def verify_token(credentials: HTTPAuthorizationCredentials = Depends(security)):
-    if credentials.credentials != MY_API_KEY:
+def verify_key(request: Request):
+    auth = request.headers.get("Authorization")
+    if auth != f"Bearer {MY_API_KEY}":
         raise HTTPException(status_code=403, detail="Unauthorized")
-    return credentials.credentials
-
-@app.post("/v1/chat")
-async def chat(request: Request, _ = Depends(verify_token)):
-    try:
-        data = await request.json()
-        user_prompt = data.get("prompt", "").strip()
-        if not user_prompt:
-            return JSONResponse(status_code=400, content={"error": "No prompt provided"})
 
-        # MPT chat format
-        prompt = f"<|im_start|>user\n{user_prompt}<|im_end|>\n<|im_start|>assistant\n"
+@app.get("/")
+def home():
+    return {"status": "Online", "branding": "ChatMPT Team"}
 
-        output = llm(
-            prompt,
-            max_tokens=512,
-            temperature=0.7,
-            stop=["<|im_end|>", "<|im_start|>"],
-            echo=False
-        )
-
-        reply = output["choices"][0]["text"].strip()
-        return JSONResponse(content={"reply": reply})
-
-    except Exception as e:
-        return JSONResponse(status_code=500, content={"error": str(e)})
+@app.post("/v1/chat")
+async def chat(request: Request, _ = Depends(verify_key)):
+    body = await request.json()
+    user_input = body.get("prompt", "")
+    
+    # Instruction-based prompt to fix the identity
+    prompt = (
+        "Assistant is a polite AI named ChatMPT, created by the ChatMPT Team.\n"
+        f"User: {user_input}\n"
+        "Assistant:"
+    )
+    
+    response = llm(
+        prompt, 
+        max_tokens=500, 
+        stop=["User:", "\n", "Assistant:"], 
+        temperature=0.7 
+    )
+    
+    raw_reply = response["choices"][0]["text"].strip()
+    
+    # Anti-Typo Safety Filter
+    final_reply = raw_reply.replace("ChatPBT", "ChatMPT").replace("ChatPP", "ChatMPT")
+    
+    return {"reply": final_reply}
 
 if __name__ == "__main__":
     uvicorn.run(app, host="0.0.0.0", port=7860)