Mirrors/ArchiveBox
2
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox synced 2024-11-10 14:44:18 +00:00
Code Issues Projects Releases Packages Wiki Activity

fix archive urls not pointing to .html wget versions!!

Browse source
This commit is contained in:
Nick Sweeting 2017-07-04 04:03:09 -05:00
parent e0f702bccd
commit d59bdbc86a
2 changed files with 21 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

21
archive.py
View file

@ -326,6 +326,22 @@ def valid_links(links):
"""remove chrome://, about:// or other schemed links that cant be archived"""
return (link for link in links if link['url'].startswith('http') or link['url'].startswith('ftp'))
def calculate_archive_url(link):
"""calculate the path to the wgetted html file, since wget may adjust some paths
to be different than the base_url path
See docs on wget --adjust-extension."""
split_url = link['url'].split('#', 1)
if re.search(".+\\.[Hh][Tt][Mm][Ll]?$", split_url[0], re.I | re.M):
# already ends in .html
return link['base_url']
else:
# .html needs to be appended
url = split_url[0] if not split_url[0].endswith('/') else split_url[0][:-1]
without_scheme = url.split('://', 1)[-1]
return '#'.join([without_scheme + '.html', *split_url[1:]])
def dump_index(links, service):
"""create index.html file for a given list of links and service"""
@ -337,7 +353,7 @@ def dump_index(links, service):
link_html = """\
<tr>
<td>{time}</td>
<td><a href="archive/{timestamp}/{base_url}" style="font-size:1.4em;text-decoration:none;color:black;" title="{title}">
<td><a href="archive/{timestamp}/{archive_url}" style="font-size:1.4em;text-decoration:none;color:black;" title="{title}">
<img src="archive/{timestamp}/favicon.ico">
{title} <small style="background-color: #eee;border-radius:4px; float:right">{tags}</small>
</td>
@ -352,6 +368,9 @@ def dump_index(links, service):
# since we dont screenshot or PDF links that are images or PDFs, change those links to point to the wget'ed file
link_info = {**link}
# append .html to archive links that dont have it, since wget appends .html to everything
link_info['archive_url'] = calculate_archive_url(link)
# add link type to title
if link['type']:
link_info.update({'title': '{title} ({type})'.format(**link)})

1
examples/firefox_export.html
View file

@ -21,6 +21,7 @@
<DT><A HREF="https://duckduckgo.com/?q=firefox+export+bookmarks&t=ffhp&ia=web" ADD_DATE="1497562974" LAST_MODIFIED="1497562974" ICON_URI="https://duckduckgo.com/favicon.ico" ICON="data:image/png;base64,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">firefox export bookmarks at DuckDuckGo</A>
<DT><A HREF="https://duckduckgo.com/?q=archive+firefox+bookmarks&t=ffab&ia=web" ADD_DATE="1497562974" LAST_MODIFIED="1497562974" ICON_URI="https://duckduckgo.com/favicon.ico" ICON="data:image/png;base64,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">archive firefox bookmarks at DuckDuckGo</A>
<DT><A HREF="https://github.com/nodiscc" ADD_DATE="1497562974" LAST_MODIFIED="1497562974" ICON_URI="https://assets-cdn.github.com/favicon.ico" ICON="data:image/png;base64,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">nodiscc (nodiscc) · GitHub</A>
<DT><A HREF="https://github.com/pirate/bookmark-archiver#troubleshooting" ADD_DATE="1497562975" LAST_MODIFIED="1497562975" ICON_URI="https://assets-cdn.github.com/favicon.ico" ICON="data:image/png;base64,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">pirate/bookmark-archiver · Github</A>
<DT><A HREF="https://www.ghacks.net/2009/07/23/firefox-bookmarks-archiver/" ADD_DATE="1497562974" LAST_MODIFIED="1497562974" ICON_URI="https://www.ghacks.net/wp-content/uploads/2005/10/favicon.ico" ICON="data:image/png;base64,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">Firefox Bookmarks Archiver - gHacks Tech News</A>
</DL><p>
<DT><H3 ADD_DATE="1409779227" LAST_MODIFIED="1470506008" PERSONAL_TOOLBAR_FOLDER="true">Bookmarks Toolbar</H3>