[Pdf Erorder] #python

pdf_eroder.py

import os
import shutil
import sys
import traceback
import uuid
import time
import cv2
import numpy as np
from fpdf import FPDF
from pdf2image import convert_from_path, pdfinfo_from_path
from tqdm import tqdm

try:
    pdf_path:str=sys.argv[1].strip()
    print(f"Prcessing :{pdf_path}")
    # print(sys.argv)
except:
    print("Imvalid argument found in python")
    sys.exit()
if not pdf_path.endswith(".pdf"):
    print("The given file is not pdf")
    sys.exit()


if not os.path.isfile(pdf_path):
    print(f"No such path \"{pdf_path}\"")
    sys.exit()

WORKING = os.path.split(pdf_path)[0]


#############################
def show_list_progress(
        the_list,
        interactive: bool = True,
        unit: str = "file",
        description: str = None,
):
    if not interactive:
        return the_list
    else:
        return tqdm(
            the_list,
            dynamic_ncols=True,
            unit=unit,
            desc=description,
            bar_format="{l_bar}{bar}|[{elapsed}<{remaining}, {rate_fmt}]",
        )


def get_file_in_subfolder(path):
    onlyfiles = [os.path.join(path, f) for f in os.listdir(path) if os.path.isfile(os.path.join(path, f))]
    return onlyfiles


#############################
tmp_folder_name = "tmp_" + str(uuid.uuid4())
TMP_FOLDER = os.path.join(WORKING, tmp_folder_name)

######################
if os.path.isfile(TMP_FOLDER) or os.path.isdir(TMP_FOLDER):
    print(f'"{TMP_FOLDER}" is already exist!!!')
    sys.exit()
os.makedirs(TMP_FOLDER)
######################

info = pdfinfo_from_path(pdf_path, userpw=None, poppler_path=None)

maxPages = info["Pages"]
PAGE_CHUNK = 2
for page in show_list_progress(range(1, maxPages + 1, PAGE_CHUNK), description="Spliting pages to images", unit="page"):
    images_chunk = convert_from_path(pdf_path, dpi=600, first_page=page, last_page=min(page + PAGE_CHUNK - 1, maxPages))
    count = page - 1
    for image in images_chunk:
        count += 1
        image.save(os.path.join(TMP_FOLDER, f'{count}.jpg'), 'JPEG')
#############################
# lets modified the images_chunk
eroded_folder = os.path.join(TMP_FOLDER, "eroded")
os.makedirs(eroded_folder)
iamge_list = get_file_in_subfolder(TMP_FOLDER)
for img in show_list_progress(iamge_list, unit="image", description="eroding emages..."):
    out = os.path.join(eroded_folder, os.path.split(img)[1])
    kernel = np.ones((2, 2), np.uint8)
    #kernel = np.ones((4, 4), np.uint8)
    img_grey = cv2.imread(img, cv2.IMREAD_GRAYSCALE)
    img_erosion = cv2.erode(img_grey, kernel, iterations=4)

    # /////////////greay scalling
    thresh = 128
    img_binary = cv2.threshold(img_erosion, thresh, 255, cv2.THRESH_BINARY)[1]
    #save image
    cv2.imwrite(out, img_binary)


    # cv2.imwrite(out, img_erosion)


def make_pdf_file(filename_list, parent_image_folder, original_pdf: str):
    suffix = "_eroded_pdf_file"
    pdf_file = os.path.join(WORKING, os.path.split(original_pdf)[1] + f"{suffix}.pdf")
    odd_file = os.path.join(WORKING, os.path.split(original_pdf)[1] + f"{suffix}_odd_actual_even_reversed_second.pdf")
    even_file = os.path.join(WORKING, os.path.split(original_pdf)[1] + f"{suffix}_even_actiual_odd_first.pdf")
    if os.path.isfile(pdf_file):
        print("Already printed")
        sys.exit()

    pdf = FPDF('P', 'mm', 'A4')
    odd_file_paths = []
    even_file_paths = []

    for num in range(1, len(filename_list) + 1):
        image_name = os.path.join(parent_image_folder, str(num)) + ".jpg"
        if num % 2 == 0:
            even_file_paths.append(image_name)
        else:
            odd_file_paths.append(image_name)
        pdf.add_page()
        pdf.image(image_name, 0, 0, 210, 297)
    print(f"write images: {pdf_file}")
    pdf.output(pdf_file, "F")
    #########################
    ###
    # check page count
    flag_page_mitch_match_odd = False
    if len(odd_file_paths) != len(even_file_paths):
        if len(odd_file_paths) > len(even_file_paths):
            flag_page_mitch_match_odd = True
            even_file_paths.append(None)
            #raise Exception("Alert.. page miscount")
        if len(odd_file_paths) < len(even_file_paths):
            flag_page_mitch_match_odd = True
            odd_file_paths.append(None)
            raise Exception("Alert.. page miscount")
    ###
    if len(odd_file_paths) != len(even_file_paths):
        raise Exception("Alert.. page miscount[Final]")

    odd_pdf = FPDF('P', 'mm', 'A4')
    odd_file_paths.reverse()
    for filename in odd_file_paths:
        odd_pdf.add_page()
        if filename is None:
            continue
        odd_pdf.image(filename, 0, 0, 210, 297)
    print(f"write images: {odd_file}")
    odd_pdf.output(odd_file, "F")
    #####################
    even_pdf = FPDF('P', 'mm', 'A4')
    for filename in even_file_paths:
        even_pdf.add_page()
        if filename is None:
            continue
        even_pdf.image(filename, 0, 0, 210, 297)

    print(f"write images: {even_file}")
    even_pdf.output(even_file, "F")


try:
    make_pdf_file(get_file_in_subfolder(eroded_folder), eroded_folder, pdf_path)
except Exception as e:
    traceback.print_exc()
    pass
time.sleep(2)
shutil.rmtree(TMP_FOLDER)

pdf_eroder.sh

#!/bin/bash

if test "$#" -ne 1; then
    echo "Illegal number of parameters"
    exit
fi

re='^[0-9]+$'
if  [[ $1 =~ $re ]] ; then
   echo "Please pass a device namet!!!"
   exit
fi

FILE=$PWD/$1

if test -f "$FILE"; then
    real_pdf=$FILE
elif test -f "$1"; then
    real_pdf=$1
fi

if [[ -z "${real_pdf}" ]]; then
  echo $1" doesnot exist"
else
  echo $real_pdf
  python3 ./pdf_eroder.py "$real_pdf"
fi