Update app.py

app.py

@@ -1,35 +1,31 @@
 from fastapi import FastAPI, HTTPException, Request
-from fastapi.responses import StreamingResponse
+from fastapi.responses import StreamingResponse, FileResponse
 from fastapi.middleware.cors import CORSMiddleware
 import aiohttp
 import json
-import time
-import random
 import ast
-import urllib.parse
-from apscheduler.schedulers.background import BackgroundScheduler
 import os
 from pydantic import BaseModel
+from datetime import datetime, timedelta
+from aiocache import cached, SimpleMemoryCache
+from aiocache.serializers import JsonSerializer
 
 SAMBA_NOVA_API_KEY = os.environ.get("SAMBA_NOVA_API_KEY", None)
 
 app = FastAPI()
 
-# Time-Limited Infinite Cache
-cache = {}
-CACHE_DURATION = 120
+# Constants for Caching
+CACHE_TIME_SECONDS = 24 * 3600  # 24 hours in seconds
 
-# Function to clean up expired cache entries
-def cleanup_cache():
-    current_time = time.time()
-    for key, (value, timestamp) in list(cache.items()):
-        if current_time - timestamp > CACHE_DURATION:
-            del cache[key]
-
-# Initialize and start the scheduler
-scheduler = BackgroundScheduler()
-scheduler.add_job(cleanup_cache, 'interval', seconds=60)  # Run cleanup every 60 seconds
-scheduler.start()
+# Cache Middleware to add caching headers
+@app.middleware("http")
+async def add_cache_headers(request: Request, call_next):
+    response = await call_next(request)
+    if response.status_code == 200 and request.url.path != "/" and request.url.path != "/script1.js" and request.url.path != "/script2.js" and request.url.path != "/styles.css":
+        expires = datetime.utcnow() + timedelta(seconds=CACHE_TIME_SECONDS)
+        response.headers["Cache-Control"] = f"public, max-age={CACHE_TIME_SECONDS}"
+        response.headers["Expires"] = expires.strftime("%a, %d %b %Y %H:%M:%S GMT")
+    return response
 
 class StreamTextRequest(BaseModel):
     query: str
@@ -38,43 +34,31 @@ class StreamTextRequest(BaseModel):
     api_key: str = None
 
 @app.post("/stream_text")
+@cached(ttl=600, cache=SimpleMemoryCache, serializer=JsonSerializer())
 async def stream_text(request: StreamTextRequest):
-    current_time = time.time()
-    cache_key = (request.query, request.history, request.model)
-
-    # Check if the request is in the cache and not expired
-    if cache_key in cache:
-        cached_response, timestamp = cache[cache_key]
-        return StreamingResponse(iter([f"{cached_response}"]), media_type='text/event-stream')
-
     # Model selection logic
     if "405" in request.model:
         fmodel = "Meta-Llama-3.1-405B-Instruct"
-    if "70" in request.model:
-        fmodel = "Meta-Llama-3.1-70B-Instruct"
+    elif "70" in request.model:
+        fmodel = "Meta-Llama-3.3-70B-Instruct"
     else:
-        fmodel = "Meta-Llama-3.1-8B-Instruct"
+        fmodel = "Meta-Llama-3.1-1B-Instruct"
 
     system_message = """You are Voicee, a friendly and intelligent voice assistant created by KingNish. Your primary goal is to provide accurate, concise, and engaging responses while maintaining a positive and upbeat tone. Always aim to provide clear and relevant information that directly addresses the user's query, but feel free to sprinkle in a dash of humor—after all, laughter is the best app! Keep your responses brief and to the point, avoiding unnecessary details or tangents, unless they’re hilariously relevant. Use a friendly and approachable tone to create a pleasant interaction, and don’t shy away from a cheeky pun or two! Tailor your responses based on the user's input and previous interactions, ensuring a personalized experience that feels like chatting with a witty friend. Invite users to ask follow-up questions or clarify their needs, fostering a conversational flow that’s as smooth as butter on a hot pancake. Aim to put a smile on the user's face with light-hearted and fun responses, and be proactive in offering additional help or suggestions related to the user's query. Remember, your goal is to be the go-to assistant for users, making their experience enjoyable and informative—like a delightful dessert after a hearty meal!"""
 
     messages = [{'role': 'system', 'content': system_message}]
-
     messages.extend(ast.literal_eval(request.history))
-
     messages.append({'role': 'user', 'content': request.query})
     
     data = {'messages': messages, 'stream': True, 'model': fmodel}
-
     api_key = request.api_key if request.api_key != 'none' else SAMBA_NOVA_API_KEY
 
-
     async def stream_response():
         async with aiohttp.ClientSession() as session:
-            async with session.post('https://api.sambanova.ai/v1/chat/completions', headers = { 'Authorization': f'Bearer {api_key}', 'Content-Type': 'application/json' }, json=data) as response:
+            async with session.post('https://api.sambanova.ai/v1/chat/completions', headers={'Authorization': f'Bearer {api_key}', 'Content-Type': 'application/json'}, json=data) as response:
                 if response.status != 200:
                     raise HTTPException(status_code=response.status, detail="Error fetching AI response")
 
-                response_content = ""
                 async for line in response.content:
                     line = line.decode('utf-8').strip()
                     if line.startswith('data: {'):
@@ -84,22 +68,13 @@ async def stream_text(request: StreamTextRequest):
                             content = parsed_data.get("choices", [{}])[0].get("delta", {}).get("content", '')
                             if content:
                                 content = content.replace("\n", " ")
-                                response_content += f"data: {content}\n\n"
                                 yield f"data: {content}\n\n"
                         except json.JSONDecodeError as e:
                             print(f"Error decoding JSON: {e}")
                             yield f"data: Error decoding JSON\n\n"
 
-                # Cache the full response
-                cache[cache_key] = (response_content, current_time)
-
     return StreamingResponse(stream_response(), media_type='text/event-stream')
 
-
-
-# Serve index.html from the same directory as your main.py file 
-from starlette.responses import FileResponse 
-
 @app.get("/script1.js")
 async def script1_js():
     return FileResponse("script1.js")