Update app.py

app.py

@@ -1,11 +1,13 @@
+import os
+import httpx
 from fastapi import FastAPI, Request
 from fastapi.middleware.cors import CORSMiddleware
-import httpx
-import os
+from fastapi.responses import StreamingResponse
 
+# सिस्टम की नींव: FastAPI एप्लीकेशन
 app = FastAPI()
 
-# CORS इनेबल करना
+# सुरक्षा के लिए CORS का सेटअप
 app.add_middleware(
     CORSMiddleware,
     allow_origins=["*"],
@@ -13,44 +15,47 @@ app.add_middleware(
     allow_headers=["*"],
 )
 
-# कॉन्फ़िगरेशन - Secrets से लें
-NVIDIA_API_KEY = os.getenv("NVIDIA_API_KEY")
-MODEL_NAME = os.getenv("MODEL_NAME", "microsoft/phi-4-multimodal-instruct")
+# सीक्रेट्स से डेटा प्राप्त करना
+API_KEY = os.getenv("API_KEY")
+MODEL_NAME = os.getenv("MODEL_NAME", "meta/llama-4-maverick-17b-128e-instruct")
 API_URL = "https://integrate.api.nvidia.com/v1/chat/completions"
 
-# सिस्टम प्रॉम्ट जो हमेशा सर्वर पर रहेगा
-SYSTEM_PROMPT = "You are Aura Gen 2.0, a state-of-the-art AI assistant developed by Divy Bhai. You possess world-class expertise in software engineering, advanced algorithmic logic, and complex problem-solving. Always provide highly optimized, secure, and precise solutions in English by default. Use Markdown formatting extensively (code blocks, tables, lists) to structure your answers clearly and effectively."
+# सिस्टम निर्देश - आपके द्वारा बताए गए नाम के साथ
+SYSTEM_INSTRUCTION = "You are Aura Gen 2.0, a state-of-the-art AI assistant developed by Divy. You possess world-class expertise in software engineering, advanced algorithmic logic, and complex problem-solving. Always provide highly optimized, secure, and precise solutions in English by default. You are capable of analyzing both text and images provided by the user. Use Markdown formatting extensively."
 
-@app.post("/v1/chat/completions")
-async def chat_proxy(request: Request):
+@app.post("/chat")
+async def process_chat(request: Request):
     data = await request.json()
     user_messages = data.get("messages", [])
     
-    # हमेशा सिस्टम प्रॉम्ट को सबसे ऊपर जोड़ें
-    full_messages = [{"role": "system", "content": SYSTEM_PROMPT}] + user_messages
+    # संदेशों को व्यवस्थित करना
+    full_messages = [{"role": "system", "content": SYSTEM_INSTRUCTION}] + user_messages
     
-    async with httpx.AsyncClient() as client:
-        response = await client.post(
-            API_URL,
-            headers={
-                "Authorization": f"Bearer {NVIDIA_API_KEY}",
-                "Content-Type": "application/json"
-            },
-            json={
-                "model": MODEL_NAME,
-                "messages": full_messages,
-                "max_tokens": data.get("max_tokens", 4096),
-                "temperature": data.get("temperature", 0.6),
-                "top_p": 0.70,
-                "frequency_penalty": 0.00,
-                "presence_penalty": 0.00,
-                "stream": False
-            },
-            timeout=60.0
-        )
-    return response.json()
+    # पेलोड बनाना (इमेज और टेक्स्ट दोनों के लिए तैयार)
+    payload = {
+        "model": MODEL_NAME,
+        "messages": full_messages,
+        "max_tokens": 1024,
+        "temperature": 0.7,
+        "top_p": 0.9,
+        "stream": True
+    }
+
+    # स्ट्रीमिंग क्लाइंट बनाना
+    async def generate_stream():
+        async with httpx.AsyncClient(timeout=60.0) as client:
+            async with client.stream(
+                "POST", 
+                API_URL, 
+                headers={"Authorization": f"Bearer {API_KEY}", "Content-Type": "application/json"},
+                json=payload
+            ) as response:
+                async for line in response.aiter_lines():
+                    if line:
+                        yield line + "\n"
+
+    return StreamingResponse(generate_stream(), media_type="text/event-stream")
 
 if __name__ == "__main__":
     import uvicorn
-    # Hugging Face Spaces 7860 पोर्ट का उपयोग करता है
-    uvicorn.run(app, host="0.0.0.0", port=7860)
+    uvicorn.run(app, host="0.0.0.0", port=7860)