Update app.py

app.py

@@ -1,15 +1,14 @@
 import os
 import logging
+import time
 from dotenv import load_dotenv
 import streamlit as st
 from PyPDF2 import PdfReader
 from langchain.text_splitter import CharacterTextSplitter
-# from langchain.embeddings import HuggingFaceInstructEmbeddings
 from langchain_cohere import CohereEmbeddings
 from langchain.vectorstores import FAISS
 from langchain.memory import ConversationBufferMemory
 from langchain.chains import ConversationalRetrievalChain
-# from langchain.llms import Ollama
 from langchain_groq import ChatGroq
 
 # Load environment variables
@@ -18,7 +17,7 @@ load_dotenv()
 # Set up logging
 logging.basicConfig(
     level=logging.INFO,
-    format='%(asctime)s - %(levelname)s - %(message)s'
+    format="%(asctime)s - %(levelname)s - %(message)s"
 )
 
 # Function to extract text from PDF files
@@ -41,24 +40,38 @@ def get_text_chunks(text):
     chunks = text_splitter.split_text(text)
     return chunks
 
-# Function to create a FAISS vectorstore
-# def get_vectorstore(text_chunks):
-#     embeddings = HuggingFaceInstructEmbeddings(model_name="hkunlp/instructor-xl")
-#     vectorstore = FAISS.from_texts(texts=text_chunks, embedding=embeddings)
-#     return vectorstore
-
+# Function to create a FAISS vectorstore with rate-limiting and retry logic
 def get_vectorstore(text_chunks):
     cohere_api_key = os.getenv("COHERE_API_KEY")
     embeddings = CohereEmbeddings(model="embed-english-v3.0", cohere_api_key=cohere_api_key)
-    vectorstore = FAISS.from_texts(texts=text_chunks, embedding=embeddings)
+    
+    vectorstore = None
+    batch_size = 10  # Process chunks in batches of 10
+    for i in range(0, len(text_chunks), batch_size):
+        batch = text_chunks[i:i+batch_size]
+        retry_count = 0
+        
+        while retry_count < 5:  # Retry up to 5 times
+            try:
+                if vectorstore is None:
+                    vectorstore = FAISS.from_texts(texts=batch, embedding=embeddings)
+                else:
+                    vectorstore.add_texts(batch, embedding=embeddings)
+                break  # Exit retry loop if successful
+            except Exception as e:
+                if "rate limit" in str(e).lower():
+                    logging.warning(f"Rate limit exceeded. Retrying batch {i//batch_size + 1} in {2 ** retry_count} seconds...")
+                    time.sleep(2 ** retry_count)  # Exponential backoff
+                    retry_count += 1
+                else:
+                    raise e  # Raise other errors
     return vectorstore
 
 # Function to set up the conversational retrieval chain
 def get_conversation_chain(vectorstore):
     try:
-        # llm = Ollama(model="llama3.2:1b")
         llm = ChatGroq(model="llama-3.1-70b-versatile", temperature=0.5)
-        memory = ConversationBufferMemory(memory_key='chat_history', return_messages=True)
+        memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
         
         conversation_chain = ConversationalRetrievalChain.from_llm(
             llm=llm,
@@ -75,8 +88,8 @@ def get_conversation_chain(vectorstore):
 # Handle user input
 def handle_userinput(user_question):
     if st.session_state.conversation is not None:
-        response = st.session_state.conversation({'question': user_question})
-        st.session_state.chat_history = response['chat_history']
+        response = st.session_state.conversation({"question": user_question})
+        st.session_state.chat_history = response["chat_history"]
 
         for i, message in enumerate(st.session_state.chat_history):
             if i % 2 == 0:
@@ -113,21 +126,5 @@ def main():
                 vectorstore = get_vectorstore(text_chunks)
                 st.session_state.conversation = get_conversation_chain(vectorstore)
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     main()
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-