Replace PyPDF2 with pypdfium2

app.py

@@ -1,6 +1,6 @@
 import streamlit as st
 from dotenv import load_dotenv
-from PyPDF2 import PdfReader
+import pypdfium2 as pdfium
 from langchain.text_splitter import CharacterTextSplitter
 from langchain.embeddings import OpenAIEmbeddings, HuggingFaceInstructEmbeddings
 from langchain.vectorstores import FAISS
@@ -10,12 +10,15 @@
 from htmlTemplates import css, bot_template, user_template
 from langchain.llms import HuggingFaceHub
 
+
 def get_pdf_text(pdf_docs):
     text = ""
     for pdf in pdf_docs:
-        pdf_reader = PdfReader(pdf)
-        for page in pdf_reader.pages:
-            text += page.extract_text()
+        pdf_reader = pdfium.PdfDocument(pdf)
+        for i in range(len(pdf_reader)):
+            page = pdf_reader.get_page(i)
+            textpage = page.get_textpage()
+            text += textpage.get_text_range() + "\n"
     return text
 
 
@@ -88,7 +91,6 @@ def main():
             with st.spinner("Processing"):
                 # get pdf text
                 raw_text = get_pdf_text(pdf_docs)
-
                 # get the text chunks
                 text_chunks = get_text_chunks(raw_text)
 

requirements.txt

@@ -1,11 +1,11 @@
 langchain==0.0.184
-PyPDF2==3.0.1
 python-dotenv==1.0.0
 streamlit==1.18.1
 openai==0.27.6
 faiss-cpu==1.7.4
 altair==4
 tiktoken==0.4.0
+pypdfium2==4.18.0
 # uncomment to use huggingface llms
 # huggingface-hub==0.14.1