dinhquangson commited on
Commit
cae8ab9
1 Parent(s): cdde907

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +5 -3
app.py CHANGED
@@ -37,9 +37,11 @@ def get_pdf_pages(pdf_docs):
37
  """
38
  pages = []
39
  for pdf in pdf_docs:
40
- data=pdf.read()
41
- print(str(len(data)))
42
- pdf_loader = UnstructuredPDFLoader(data)
 
 
43
  pdf_pages = pdf_loader.load_and_split()
44
  pages=paegs+pdf_pages
45
  return pages
 
37
  """
38
  pages = []
39
  for pdf in pdf_docs:
40
+ pdf_path=os.path.join("tempDir",pdf.name)
41
+ with open(pdf_path, "wb") as f:
42
+ f.write(pdf.getbuffer())
43
+
44
+ pdf_loader = UnstructuredPDFLoader(pdf_path)
45
  pdf_pages = pdf_loader.load_and_split()
46
  pages=paegs+pdf_pages
47
  return pages