server.py

import os
import tempfile
import streamlit as st
import ocrmypdf

import logging

logging.basicConfig(
    level=logging.INFO,
    format="%(asctime)s [%(levelname)s] %(message)s",
    handlers=[
        logging.FileHandler("debug.log"),
        logging.StreamHandler()
    ]
)

STATE_FIXED_PDF_PATH="fixed_pdf_path"
STATE_UPLOADED_FILE="uploaded_file"
STATE_UPLOADED_FILE_NAME="uploaded_file_name"

ocrmypdf_ocr=False
ocrmypdf_remove_background=False
ocrmypdf_deskew=False
ocrmypdf_optimize=1

st.set_page_config(page_title='OCRmyPDFonWEB')

def uploader_callback():
    if st.session_state[STATE_UPLOADED_FILE] is not None:
        fd, path = tempfile.mkstemp()

        with st.spinner('For large files, the duration of the process can be very long. Therefore, it is recommended to be patient and wait as long as no error message is displayed.'):
            try:
                with os.fdopen(fd, 'wb') as tmp:
                    uploaded_file=st.session_state[STATE_UPLOADED_FILE]

                    tmp.write(uploaded_file.getvalue())
                    tmp.close()

                    fixed_path=path + '_fixed.pdf'

                    ocrmypdf.ocr(path, fixed_path, 
                        remove_background=ocrmypdf_remove_background, 
                        optimize=ocrmypdf_optimize,
                        deskew=ocrmypdf_deskew,
                        tesseract_timeout=400 if ocrmypdf_ocr else 0, 
                        skip_text=True, 
                        max_image_mpixels=901167396
                    )

                    st.session_state[STATE_FIXED_PDF_PATH] = fixed_path
                    st.session_state[STATE_UPLOADED_FILE_NAME] = uploaded_file.name + '_ompow.pdf'                    
            finally:
                os.remove(path)

def download_callback():
    for key in st.session_state.keys():
        del st.session_state[key]

c1 = st.container()
c1.title("OCRmyPDFonWEB")

if STATE_FIXED_PDF_PATH not in st.session_state:
    c1.write("Select options and upload PDF. Once this process is successfully completed, the edited PDF file will be available for download.")
    ocrmypdf_ocr = c1.checkbox('Optical character recognition (OCR)')
    ocrmypdf_remove_background = c1.checkbox('Remove background')
    ocrmypdf_deskew = c1.checkbox('Deskew')
    ocrmypdf_optimize = c1.slider(
        label = "Optimize file size (0 off, 1 without quality loss, 3 smallest but maybe with slight quality loss)",
        min_value = 0,
        max_value = 3,
        value = 1
    )
    c1.file_uploader(label="Upload PDF", on_change=uploader_callback, key=STATE_UPLOADED_FILE)
else:
    c1.write("Editing of the PDF file has been successfully completed. Now the optimized version is ready for download.")
    try:
        with open(st.session_state[STATE_FIXED_PDF_PATH], 'rb') as f:
            st.download_button(label = 'Download PDF', data = f, file_name=st.session_state[STATE_UPLOADED_FILE_NAME], on_click = download_callback)
    finally:
        os.remove(st.session_state[STATE_FIXED_PDF_PATH])