Update app.py

app.py

@@ -415,28 +415,47 @@ async def read_root(response: Response):
 </body>
 </html>'''
 
-from openai import OpenAI
 
-client = OpenAI(base_url="http://localhost:8080/v1", api_key="no-key-required")
+from google import genai
+from google.genai import types
+import requests
+
+
+Api_key = os.getenv('API_KEY')
+System_instruction = '''**System Prompt for a Programmer-Oriented Coding Assistant:**\n\n> You are a highly focused, fast, and expert-level coding assistant built for professional programmers.\n> Your primary role is **to assist with code writing, debugging, refactoring, optimization, and architecture**.\n> Avoid unnecessary explanations unless asked. Do not teach—**support the user like a senior pair programmer** who assumes context and skill. Prioritize clean, correct, and efficient code.\n\n> Always:\n> * Get straight to the point.\n> * Suggest the most practical and scalable solution.\n> * Respond with complete code blocks when needed.\n> * Use strong defaults and modern conventions.\n> * Assume the user knows what they're doing.\n> * Think ahead: anticipate potential pitfalls or better approaches.\n> * Give fast, minimal answers when asked for quick help.\n\n> Only elaborate if specifically requested (e.g., “explain,” “why,” “teach,” “verbose”)'''
+
+
+client = genai.Client(api_key=Api_key)
 
 class ChatRequest(BaseModel):
     """Request model for the chat endpoint."""
     prompt: str
 
+def gen(prompt):
+    response = client.models.generate_content(
+        model="gemma-3-4b-it",
+        contents= prompt
+    )
+    return response.text
+
 @app.post("/response")
 async def handle_chat(chat_request: ChatRequest, token: str = Cookie(None)):
+    a= time.time()
     if token in Tokens:
         i = Tokens.index(token)
         History[i].append({"role": "user", "content": chat_request.prompt})
 
-        stream = client.chat.completions.create(
-                model="",
-                messages=History[i],
-            )
-        History[i].append({"role": "assistant", "content": stream.choices[0].message.content})
+        text = '<start_of_turn>system\n'+System_instruction+'<end_of_turn>\n<start_of_turn>user\n'
+        for j in History[i]:
+            if j['role']== 'user':
+                text = text + j['content'] + '<end_of_turn>\n<start_of_turn>model\n'
+            else : text = text + j['content'] + '<end_of_turn>\n<start_of_turn>user\n'
+        stream = gen(prompt)
+        History[i].append({"role": "assistant", "content": stream})
+        b = time.time()
         return {"text": stream.choices[0].message.content,
-               "time": int((stream.timings["prompt_ms"] + stream.timings["predicted_ms"])/1000 - 3),
-               "t_per_sec": round(stream.timings["predicted_per_second"] + 0.2, 2)}
+               "time": (b-a)/1000,
+               "t_per_sec": 0}
     else: return 'Please stop. Just refresh the page.'
 
 @app.post("/history")