Update app.py

app.py

@@ -1,5 +1,5 @@
 import PyPDF2
-import subprocess
+import re
 import streamlit as st
 #import fitz
 from transformers import pipeline
@@ -21,29 +21,15 @@ st.title("PDF2SLIDE")
 uploaded_file = st.file_uploader("Selectionnez un PDF", type=["pdf"])
 
 if uploaded_file is not None:
-#	pdf_reader = PyPDF2.PdfReader(uploaded_file)
-
-#	paragraphs = []
-
-#	for page_number in range(len(pdf_reader.pages)):
-#		page = pdf_reader.pages[page_number]
-#		page_text = page.extract_text()
-#		#print(page_text)
-#		paragraphs.extend(page_text.split('\n\n'))
-#		for paragraph in paragraphs:
-#			print(paragraph)
-#			print(424242)
-
-
-    pdf_text = None
-    print(f"Chemin du fichier PDF : {uploaded_file.name}")
-    try:
-        pdf_text = subprocess.check_output(["pdftotext", uploaded_file.name, "-"]).decode("utf-8")
-    except Exception as e:
-        st.error(f"Erreur lors de la conversion du PDF en texte : {str(e)}")
-
-    if pdf_text:
-        paragraphs = pdf_text.split('\n\n')
+pdf_reader = PyPDF2.PdfFileReader(uploaded_file)
+    
+    paragraphs = []
+
+    for page_number in range(pdf_reader.numPages):
+        page = pdf_reader.getPage(page_number)
+        page_text = page.extractText()
+        paragraph_list = re.split(r'\n{2,}', page_text)
+        paragraphs.extend(paragraph_list)
         for paragraph in paragraphs:
             print(paragraph)
             print(424242)