from sdk.utils.util_class import PathParser
from sdk.temp.temp_supports import IsSolution, DM
这些全是自定义的工具包，自己去其他文章下找

# !/usr/bin/python3
# -*- coding:utf-8 -*-
"""
@author: JHC000abc@gmail.com
@file: demo.py
@time: 2024/3/6 16:10 
@desc: docx/pdf 提取图片 转存成 markdown文档"""# python-docx
from docx import Document
# pymupdf
import fitz
import re
from sdk.utils.util_class import PathParser
from sdk.temp.temp_supports import IsSolution, DMclass Solution(IsSolution):"""Solution"""def __init__(self, **kwargs):"""初始化函数:param kwargs: 字典类型的参数字典，包含可选的关键字参数"""super(Solution, self).__init__()self.__dict__.update({k: v for k, v in [i for i in locals().values() if isinstance(i, dict)][0].items()})def exit_handler(self):"""程序退出自动执行:return:"""print("程序退出")def muti_thread_function(self, *args):"""处理数据函数:param args::return:"""img_out_file_lis,file_split = argswith open(self.folder.merge_path([self.save_path, f"{file_split.name}.{file_split.tail}.md"]), "w",encoding="utf-8") as f:for _file in img_out_file_lis:_file_split = PathParser(_file)img_folder = "/".join(self.folder.split_path(_file)[-2:])f.write(f"![{_file_split.name}](./{img_folder})\n")@DM.add_project()def parse_imgs_from_docx(self, file, file_split):"""从doc文档中解析，保存图片，并返回图片本地保存路径:param file::return:"""print("提交", file)img_out_file_lis = []doc = Document(file)dict_rel = doc.part.relsfor index, (r_id, rel) in enumerate(dict_rel.items()):if not (  # 如果文件不是在media或者embeddings中的，直接跳过str(rel.target_ref).startswith('media')or str(rel.target_ref).startswith('embeddings')):continue# 如果文件不是我们想要的后缀，也直接跳过_file_split = PathParser(str(rel.target_ref))if _file_split.tail.lower() not in ['png', "jpeg", 'jpg', 'bin']:continuefile_name = _file_split.name.lower()+"."+_file_split.tail_save_path = self.make_out_path(self.save_path,[f"{file_split.name}.{file_split.tail}"])img_save_file = self.folder.merge_path([_save_path,file_name])self.save_imgs(img_save_file,rel.target_part.blob)img_out_file_lis.append(img_save_file)self.muti_thread_function(img_out_file_lis, file_split)print("结束",file)@DM.add_project()def parse_imgs_from_pdf(self,file,file_split):""":param file::param file_split::return:"""img_out_file_lis = []checkXO = r"/Type(?= */XObject)"checkIM = r"/Subtype(?= */Image)"# 打开pdfdoc = fitz.open(file)# 图片计数imgcount = 0# 获取对象数量长度lenXREF = doc.xref_length()for i in range(1, lenXREF):# 定义对象字符串text = doc.xref_object(i)isXObject = re.search(checkXO, text)# 使用正则表达式查看是否是图片isImage = re.search(checkIM, text)# 如果不是对象也不是图片，则continueif not isXObject or not isImage:continueimgcount += 1# 根据索引生成图像pix = fitz.Pixmap(doc, i)# 根据pdf的路径生成图片的名称new_name = f"{file_split.name}_{i}.png"# 如果pix.n<5,可以直接存为PNG_save_path = self.make_out_path(self.save_path,[f"{file_split.name}.{file_split.tail}"])_save_file = self.folder.merge_path([_save_path, new_name])if pix.n < 5:pix.writePNG(_save_file)# 否则先转换CMYKelse:pix0 = fitz.Pixmap(fitz.csRGB, pix)pix0.writePNG(_save_file)pix0 = None# 释放资源pix = Noneimg_out_file_lis.append(_save_file)self.muti_thread_function(img_out_file_lis, file_split)def save_imgs(self,file,date):"""保存图片:param file::param date::return:"""with open(file, "wb") as f:f.write(date)def process(self, **kwargs):"""处理文件:param kwargs: 关键字参数:return: 无返回值"""self.in_path = kwargs["in_path"]self.save_path = kwargs["save_path"]self.folder.create_folder(self.save_path)self.name_set = set()for file, name in self.get_file(self.in_path, status=True):print(file, name)  # 打印文件名和名称file_split = PathParser(file)if file_split.tail == "docx":self.parse_imgs_from_docx(file, file_split)elif file_split.tail == "pdf":self.parse_imgs_from_pdf(file, file_split)DM.close_pool()if __name__ == '__main__':in_path = R"D:\Desktop\1"save_path = R"D:\Desktop\2"e = Solution()e.process(in_path=in_path, save_path=save_path)