切分pdf并提取内容

pdf里面包含多篇报告，报告以文章编号结尾，部分存在文章连接 。目的提取报告的标题，版号，版面，作者，文章编号，原文连接，以及文章主体部分大概文字

代码如下：
blocks：

import fitz
from PyPDF2 import PdfFileReader, PdfFileWriter
import pandas as pd
doc = fitz.Document(r'C:\Users\Crown\Desktop\WS6.0下载文本样式PDF 15Nov2022-clipbook.pdf')#作为提取文本的对象
pdf = PdfFileReader(r'C:\Users\Crown\Desktop\WS6.0下载文本样式PDF 15Nov2022-clipbook.pdf')#作为切分pdf的对象

title = []#存放pdf的多篇文档的标题
pages = []#存放pdf文档的起始页
list_data = []#存放数据的列表
for list in doc.get_toc():#遍历pdf文档的目录
    pages.append(list[2])#存储文档起始页，便于分割
    title.append(list[1])#存储标题
for pg in range(len(pages)):#切分文档
    pdf_writer = PdfFileWriter()#建立对象
    if pg == len(pages)-1:#最后一部分文档应该是最后文档的起始页到pdf的页数结尾
        for j in range(pages[pg]-1,len(pdf.pages)):
            pdf_writer.addPage(pdf.getPage(j))
        with open('../测试/temp.pdf', 'wb') as page_data:
            pdf_writer.write(page_data)
    else:#两篇文档的起始页就是一篇完整的文档
        for j in range(pages[pg]-1,pages[pg+1]-1):
            pdf_writer.addPage(pdf.getPage(j))
        with open('../测试/temp.pdf', 'wb') as page_data:#存储
            pdf_writer.write(page_data)
    temp = fitz.Document('../测试/temp.pdf')#读取完整文档
    sum = 0#统计个数
    data = {}#存储数据的字典
    banhao = ''#版号
    meiti = ''#媒体
    riqi = ''#日期
    banmian = ''#版面
    biaoti = ''#标题
    zuozhi = ''#作者
    Docld = ''#文章编号
    lianjie = ''#原文连接
    for page in range(0,len(temp)):#读取文档的每一页
        blocks = temp[page].get_text('blocks')#以模块方式读取文本信息
        if page == 0 and pg == 0:#pdf的起始页有一行写明文章总数，与其他页面不同，单独处理
            for i in range(0,len(blocks)-1):#对文本模块进行遍历，最后一个模块是注释故不需要
                if '文章总数' in blocks[i][4]:
                    continue
                if '|' in blocks[i][4]:#获取标题，日期，媒体，作者
                    meiti = blocks[i][4].split('|')[0].strip()
                    if len(blocks[i][4].split('|')) > 4:#有些有版号有些没有
                        banhao = blocks[i][4].split('|')[2].strip()
                        banmian = blocks[i][4].split('|')[3].strip()
                        if 'By' in blocks[i][4].split('|')[-1]:
                            zuozhi = blocks[i][4].split('By')[1].strip()
                    else:
                        banmian = blocks[i][4].split('|')[2].strip()
                    riqi = blocks[i][4].split('|')[1].split('\n')[0].strip()
                    riqi = riqi.split('-')[1] + '-' + riqi.split('-')[2] + '-' + riqi.split('-')[0]#改变日期格式
                if i == 2:
                    biaoti = blocks[i][4].replace('\n','').strip()
                if '原文' in blocks[i][4] and '/' in blocks[i][4]:
                    lianjie = blocks[i][4].split(':')[1].replace('\n', '')
                if '文章编号' in blocks[i][4]:
                    Docld = blocks[i][4].split(':')[1].replace('\n', '')
                sum += len(blocks[i][4]) - 1
        elif page == 0:#针对第一页
            for i in range(0,len(blocks)-1):
                #print(blocks[i])
                if '文章总数' in blocks[i][4]:
                    continue
                if '|' in blocks[i][4]:
                    meiti = blocks[i][4].split('|')[0].strip()

                    if len(blocks[i][4].split('|')) > 4:
                        banhao = blocks[i][4].split('|')[2].strip()
                        banmian = blocks[i][4].split('|')[3].strip()
                        if 'By' in blocks[i][4].split('|')[-1]:
                            zuozhi = blocks[i][4].split('By')[1].strip()
                    else:
                        banhao = ''
                        banmian = blocks[i][4].split('|')[2].strip()

                    riqi = blocks[i][4].split('|')[1].split('\n')[0].strip()
                    riqi = riqi.split('-')[1] + '-' + riqi.split('-')[2] + '-' + riqi.split('-')[0]
                if i == 1:
                    biaoti = blocks[i][4].replace('\n','').strip()
                sum += len(blocks[i][4]) - 1
                if '原文' in blocks[i][4] and '/' in blocks[i][4]:
                    lianjie = blocks[i][4].split(':')[1].replace('\n', '')
                if '文章编号' in blocks[i][4]:
                    Docld = blocks[i][4].split(':')[1].replace('\n', '')
        elif page < len(temp):#针对后续的页面
            for i in range(0, len(blocks)-1):
                #print(blocks[i])
                if '|' in blocks[i][4]:
                    continue
                if '文章编号' in blocks[i][4]:
                    Docld = blocks[i][4].split(':')[1].replace('\n', '')
                    continue
                elif '原文' in blocks[i][4] and '/' in blocks[i][4]:
                    lianjie = blocks[i][4].split(':')[1].replace('\n', '')
                    continue
                sum += len(blocks[i][4]) - 1
    data['标题'] = title[pg]
    data['日期'] = riqi
    data['媒体类型'] = '报刊'
    data['媒体'] = meiti
    data['版号'] = banhao
    data['版面'] = banmian
    data['作者'] = zuozhi
    data['字数'] = sum - 20
    data['Docld'] = Docld
    data['原文链接'] = lianjie
    print(data)
    list_data.append(data)
datas = pd.DataFrame(list_data)
datas.to_excel('1.xlsx')

data：