def compare_diff(qiniuhtml, coshtml):
urls1 = set()
id_pattern = re.compile(r' data-clipboard-text="(.*?)"', re.MULTILINE)
with open(qiniuhtml, encoding='utf-8') as f:
content = f.read()
for m in id_pattern.finditer(content):
url = m.group(1)
if (url.endswith('.zip') or url.endswith('/')):
continue
urls1.add(os.path.split(url)[1])
urls2 = set()
id_pattern2 = re.compile(r' filename="(.*?)"', re.MULTILINE)
with open(coshtml, encoding='utf-8') as f:
content = f.read()
for m in id_pattern2.finditer(content):
url = m.group(1)
if (url.endswith('.zip') or url.endswith('/')):
continue
urls2.add(url)
print(urls1.difference(urls2))
print(urls2.difference(urls1))
print(urls1.symmetric_difference(urls2))
评论列表
文章目录