Update app.py

app.py

@@ -4,7 +4,7 @@ import os
 import torch
 from fastapi import FastAPI
 from langchain.llms import VLLM
-from gptcache import cache
+from cachetools import TTLCache
 from sklearn.feature_extraction.text import TfidfVectorizer
 from sklearn.metrics.pairwise import cosine_similarity
 import nltk
@@ -27,10 +27,11 @@ model_2 = None
 model_3 = None
 model_4 = None
 
-cache_1 = cache.SimpleCache()
-cache_2 = cache.SimpleCache()
-cache_3 = cache.SimpleCache()
-cache_4 = cache.SimpleCache()
+# Using TTLCache from cachetools
+cache_1 = TTLCache(maxsize=100, ttl=600)  # maxsize=100 and ttl=600 (10 minutes)
+cache_2 = TTLCache(maxsize=100, ttl=600)
+cache_3 = TTLCache(maxsize=100, ttl=600)
+cache_4 = TTLCache(maxsize=100, ttl=600)
 
 previous_responses_1 = []
 previous_responses_2 = []
@@ -149,7 +150,7 @@ def create_langchain_model(model_name: str, device: torch.device, cache, previou
             prev_output = output_text.split()[-50:]
         output_chunks = split_output(output_text, MAX_TOKENS)
         best_response = get_best_response(output_chunks[0], previous_responses)
-        cache.put(input_text, best_response)
+        cache[input_text] = best_response
         previous_responses.append(best_response)
         return best_response
     return generate_for_model