去除 扫描全能王 pdf 的广告二维码

remove_ad_for_sao_miao_quan_neng_wang.py

import sys
import img2pdf
# =========== 总流程 ============
# pdf --> jpg (抽取内部原有的jpg)
# 过滤掉不要的图片
# jpg --> pdf
# ===============================

def extract_all_jpg_from_pdf(pdf_file):
    res = []
    with open(pdf_file, "rb") as file:
        pdf = file.read()

    img_counter = 0
    pointer = 0
    while True:
        pointer = pdf.find(b"stream", pointer)
        if pointer < 0:
            break

        x = pdf.find(b"\xff\xd8", pointer)
        if x < 0:
            pointer = pointer + 1
            continue
        else:
            extension = "jpg"

        limit = pdf.find(b"endstream", pointer)
        if limit < 0:
            break

        y = pdf.find(b"\xff\xd9", pointer, limit) + 2

        pointer = limit + 9
        if y < 2:
            continue        
        
        img = pdf[x:y]
        res.append(img)

        # img_counter = img_counter + 1
        # save_file = f"{img_counter:3d}.{extension}"
        # with open(save_file, "wb") as jpgfile:
        #     jpgfile.write(img)
    return res


if __name__ == '__main__':
    input_file = sys.argv[1]
    output_file = input_file+'-noAD.pdf'

    # 抽取原来的图片文件
    images = extract_all_jpg_from_pdf(input_file)

    # 扫描全能王生成的pdf，有一半是二维码。此处去掉
    images = [images[i] for i in range(1, len(images), 2)]

    # 重新保存为pdf
    a4_input = (img2pdf.mm_to_pt(210),img2pdf.mm_to_pt(297))
    layout_fun = img2pdf.get_layout_fun(a4_input, auto_orient=True)
    with open(output_file, "wb") as f:
        f.write(img2pdf.convert(images, layout_fun=layout_fun))