temp_code/python/pdf/pdf_comp/pdf_comp.py

# -*- coding:utf-8 -*-
# author: peng
# file: mypdf.py
# time: 2021/9/8 17:47
# desc:压缩PDF，对纯图片的PDF效果效果较好，有文字内容的可能会比较模糊，推荐高质量的压缩

import fitz
from PIL import Image
import os
from shutil import copyfile, rmtree
from math import ceil

from time import strftime, localtime, time
import easygui as g

from functools import wraps
# 时间计数装饰器，func如果有return值，必须返回才能有值
def runtime(func):
    @wraps(func)
    def wrapper(*args, **kwargs):
        print(strftime("%Y-%m-%d %H:%M:%S", localtime()))
        start = time()
        func_return = func(*args, **kwargs)
        end = time()
        print(func.__name__, args[-1], args[-2], " spend time ", end - start, " sec")
        return func_return

    return wrapper


class Luban(object):

    def __init__(self, quality, ignoreBy=102400):
        self.ignoreBy = ignoreBy
        self.quality = quality

    def setPath(self, path):
        self.path = path

    def setTargetDir(self, foldername="target"):
        self.dir, self.filename = os.path.split(self.path)
        self.targetDir = os.path.join(self.dir, foldername)

        if not os.path.exists(self.targetDir):
            os.makedirs(self.targetDir)

        self.targetPath = os.path.join(self.targetDir, "c_" + self.filename)

    def load(self):
        self.img = Image.open(self.path)

        if self.img.mode == "RGB":
            self.type = "JPEG"
        elif self.img.mode == "RGBA":
            self.type = "PNG"
        else:  # 其他的图片就转成JPEG
            self.img = self.img.convert("RGB")
            self.type = "JPEG"

    def computeScale(self):
        # 计算缩小的倍数

        srcWidth, srcHeight = self.img.size

        srcWidth = srcWidth + 1 if srcWidth % 2 == 1 else srcWidth
        srcHeight = srcHeight + 1 if srcHeight % 2 == 1 else srcHeight

        longSide = max(srcWidth, srcHeight)
        shortSide = min(srcWidth, srcHeight)

        scale = shortSide / longSide
        if (scale <= 1 and scale > 0.5625):
            if (longSide < 1664):
                return 1
            elif (longSide < 4990):
                return 2
            elif (longSide > 4990 and longSide < 10240):
                return 4
            else:
                return max(1, longSide // 1280)

        elif (scale <= 0.5625 and scale > 0.5):
            return max(1, longSide // 1280)

        else:
            return ceil(longSide / (1280.0 / scale))

    def compress(self):
        self.setTargetDir()
        # 先调整大小，再调整品质
        if os.path.getsize(self.path) <= self.ignoreBy:
            copyfile(self.path, self.targetPath)
        else:
            self.load()

            scale = self.computeScale()
            srcWidth, srcHeight = self.img.size
            cache = self.img.resize((srcWidth // scale, srcHeight // scale),
                                    Image.ANTIALIAS)

            cache.save(self.targetPath, self.type, quality=self.quality)


# 提取成图片
def covert2pic(doc, totaling, zooms=None):
    '''
    :param totaling: pdf的页数
    :param zooms: 值越大，分辨率越高，文件越清晰，列表内两个浮点数，每个尺寸的缩放系数，默认为分辨率的2倍
    :return:
    '''
    if zooms is None:
        zooms = [2.0, 2.0]
    if os.path.exists('.pdf'):  # 临时文件，需为空
        rmtree('.pdf')
    os.mkdir('.pdf')
    print(f"pdf页数为 {totaling} \n创建临时文件夹.....")
    for pg in range(totaling):
        page = doc[pg]
        print(f"\r{page}", end="")
        trans = fitz.Matrix(*zooms).preRotate(0)  # 0为旋转角度
        pm = page.getPixmap(matrix=trans, alpha=False)
        lurl = '.pdf/%s.jpg' % str(pg + 1)
        pm.writePNG(lurl) #保存
    doc.close()

# 图片合成pdf
def pic2pdf(obj, ratio, totaling):
    doc = fitz.open()
    compressor = Luban(quality=ratio)
    for pg in range(totaling):
        path = '.pdf/%s.jpg' % str(pg + 1)
        compressor.setPath(path)
        compressor.compress()
        print(f"\r 插入图片 {pg + 1}/{totaling} 中......", end="")
        img = '.pdf/target/c_%s.jpg' % str(pg + 1)
        imgdoc = fitz.open(img)  # 打开图片
        pdfbytes = imgdoc.convertToPDF()  # 使用图片创建单页的 PDF
        os.remove(img)
        imgpdf = fitz.open("pdf", pdfbytes)
        doc.insertPDF(imgpdf)  # 将当前页插入文档
    if os.path.exists(obj):  # 若pdf文件存在先删除
        os.remove(obj)
    doc.save(obj)  # 保存pdf文件
    doc.close()


@runtime
def pdfz(doc, obj, ratio, totaling):
    covert2pic(doc, totaling)
    pic2pdf(obj, ratio, totaling)


def pic_quality():
    print("输入压缩等级1~3：")
    comp_level = input("压缩等级（1=高画质50%，2=中画质70%，3=低画质80%）：（输入数字并按回车键）")
    # 用字典模拟Switch分支，注意输入的值是str类型
    ratio = {'1': 40, '2': 20, '3': 10}
    # 字典中没有则默认 低画质压缩
    return ratio.get(comp_level, 10)


if __name__ == "__main__":
    print("请选择需要压缩的PDF文件")
    while True:
        '''打开选择文件夹对话框'''
        filepath = g.fileopenbox(title=u"选择PDF", filetypes=['*.pdf'])
        if filepath == None:
            input("还未选择文件，输入任意键继续.......")
            continue
        else:
            filedir, filename = os.path.split(filepath)
            print(u'已选中文件【%s】' % (filename))
            if filename.endswith(".pdf") == False:
                input("选择的文件类型不对，输入任意键继续.......")
                continue
            ratio = pic_quality()
            obj = "new_" + filename
            doc = fitz.open(filepath)
            totaling = doc.pageCount

            pdfz(doc, obj, ratio, totaling)
            rmtree('.pdf')
            oldsize = os.stat(filepath).st_size
    
            newsize = os.stat(obj).st_size
            print('压缩结果 %.2f M >>>> %.2f M'%(oldsize/(1024 * 1024),newsize/(1024 * 1024)))
            input(f"压缩已完成，文件保存在改程序目录下{filedir}，如需继续压缩请按任意键")