[Python] 纯文本查看 复制代码 import os
from pathlib import Path
import docx
from PyPDF2 import PdfFileReader
results = []
for root, dirs, files in os.walk('path/to/folder'):
for file in files:
file_path = os.path.join(root, file)
if file.endswith('.docx'):
doc = docx.Document(file_path)
page_count = len(doc.paragraphs)
paper_size = doc.sections[0].page_width, doc.sections[0].page_height
elif file.endswith('.pdf'):
pdf = PdfFileReader(open(file_path,'rb'))
page_count = pdf.getNumPages()
paper_size = pdf.getPage(0).mediaBox[2:], pdf.getPage(0).mediaBox[3:]
else:
continue
result = f"{file_path}|{page_count}|{paper_size[0]}x{paper_size[1]}"
results.append(result)
with open('paths.txt', 'w') as f:
f.write('\n'.join(results))
print('Results exported to paths.txt')
AI 生成的代码,未测试过,仅供参考 |