mirror/archivebox
1
0
Fork 0
mirror of https://github.com/pirate/ArchiveBox.git synced 2025-08-27 00:05:27 +02:00
Code Issues Releases Wiki Activity

fix fetch page title default

Browse Source
This commit is contained in:
Nick Sweeting
2019-01-11 05:19:22 -05:00
parent 67d103a293
commit d35c6cf8b5
1 changed files with 7 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
archivebox/util.py
View File

@@ -212,15 +212,19 @@ def download_url(url):
return source_path return source_path
def fetch_page_title(url, default=None): def fetch_page_title(url, default=True):
"""Attempt to guess a page's title by downloading the html""" """Attempt to guess a page's title by downloading the html"""
if default is True:
default = url
try: try:
html_content = urllib.request.urlopen(url).read().decode('utf-8') html_content = urllib.request.urlopen(url).read().decode('utf-8')
match = re.search('<title>(.*?)</title>', html_content) match = re.search('<title>(.*?)</title>', html_content)
return match.group(1) if match else default return match.group(1) if match else default or None
except Exception: except Exception:
if default is False:
raise
return default return default