When I use the following code
from PyPDF2 import PdfFileMerger
merge = PdfFileMerger()
for newFile in nlst:
merge.append(newFile)
merge.write("newFile.pdf")
Something happened as following:
raise utils.PdfReadError("EOF marker not found")
PyPDF2.utils.PdfReadError: EOF marker not found
Anybody could tell me what happened?
After encountering this problem using camelot and PyPDF2, I did some digging and have solved the problem.
The end of file marker '%%EOF' is meant to be the very last line, but some PDF files put a huge chunk of javascript after this line, and the reader cannot find the EOF.
Illustration of what the EOF plus javascript looks like if you open it:
b'>>\r\n',
b'startxref\r\n',
b'275824\r\n',
b'%%EOF\r\n',
b'\n',
b'\n',
b'<script type="text/javascript">\n',
b'\twindow.parent.focus();\n',
b'</script><!DOCTYPE html>\n',
b'\n',
b'\n',
b'\n',
So you just need to truncate the file before the javascript begins.
Solution:
def reset_eof_of_pdf_return_stream(pdf_stream_in:list):
# find the line position of the EOF
for i, x in enumerate(txt[::-1]):
if b'%%EOF' in x:
actual_line = len(pdf_stream_in)-i
print(f'EOF found at line position {-i} = actual {actual_line}, with value {x}')
break
# return the list up to that point
return pdf_stream_in[:actual_line]
# opens the file for reading
with open('data/XXX.pdf', 'rb') as p:
txt = (p.readlines())
# get the new list terminating correctly
txtx = reset_eof_of_pdf_return_stream(txt)
# write to new pdf
with open('data/XXX_fixed.pdf', 'wb') as f:
f.writelines(txtx)
fixed_pdf = PyPDF2.PdfFileReader('data/XXX_fixed.pdf')
PDF is a file format, where a pdf parser normally starts reading the file by reading some global information located at the end of the file. At the very end of the document there needs to be a line with the content of
%%EOF
This is a marker, where the pdf parser knows, that the PDF document ends here and the global information it needs, should be before this (a startxref section).
I guess, that the error message you see, means, that one of the input documents was truncated and is missing this %%EOF-marker.
One simple solution for this problem (EOF marker not found). Open your .pdf file in other application (I used Libre office draw in Ubuntu 18.04). Then export the file as .pdf. Using this exported .pdf file the problem will not persist.
PyPDF2 cannot find the EOF marker in a PDF that is encrypted.
I came across the same error while I was working through the (excellent) Automate The Boring Stuff. Chapter 15, 2nd edition, page 355, project Combining Select Pages from Many PDFs.
I chose to combine all the PDFs I had made during this chapter into one document and one of them was an encrypted PDF and the project failed when it got to the end of the encrypted document with the error message:
PyPDF2.utils.PdfReadError: EOF marker not found
I moved the encrypted file to a different folder (so it would not be merged with the other pdfs) and the project worked fine.
So, it seems PyPDF2 cannot find the EOF marker in a PDF that is encrypted.
I've also got that problem and got a solution.
First, python reads PDF as 'rb' or 'wb' as a binary read and write format.
END OF FILE
Occurs when that there was an open parenthesis somewhere on a line, but not a matching closing parenthesis. Python reached the end of the file while looking for the closing parenthesis.
Here is the 1 solution:
Close that file that you've opened earlier using this command
newfile.close()
Check whether that pdf is opened using other variable and again close it
Same_file_with_another_variable.close()
Now open it only once and use it , you are good to go.
I wanted to add my hacky solution to this issue.
I had the same error with python requests (application/pdf). In my case the provider (a shipping labeling service) did give a 200 and a b'string which represents the PDF, but in some random cases it missed the EOF marker.
Because it was random, I came up with the following solution:
for obj in label_objects:
get_label = api.get_label(label_id=obj.label_id)
while not 'EOF' in str(get_label.content):
get_label = api.get_label(label_id=obj.label_id)
At a few tries it gives the b'string with EOF and we're good to proceed.
please use this code:
response = requests.get("your link")
pdf_io_bytes = io.BytesIO(response.content)
text_list = []
pdf = pypdf.PdfReader(pdf_io_bytes)
num_pages = len(pdf.pages)
for page in range(num_pages):
page_text = pdf.pages[page].extract_text()
text_list.append(page_text)
text = "\n".join(text_list)
Adding one more solution that worked around this issue for me - if the input files you are seeking to merge into a PDF are images, you could use a PIL function to merge as PDF via the append parameter while saving:
from PIL import Image
images = [
# the pdf page order will follow this list
Image.open( pth ).convert('RGBA')
for pth in jpg_pths
]
pth_pdf_ou = "my_merged.pdf"
images[0].save(
pth_pdf_ou,
"PDF",
resolution=100.0,
save_all=True,
append_images=images[1:]
)
i had the same problem. For me the solution was to close the previously opened file before working with it again.
© 2022 - 2024 — McMap. All rights reserved.