outlyer / daily_digest Goto Github PK

Sites which use non-standard image URLs cause problems. i.e.

<img src="//images.eurogamer.net/2015/articles//a/1/8/6/4/2/2/1/20161020085927_1.jpg.jpg/EG11/resize/600x-1/quality/80/format/jpg" width="600" height="337.5" alt="1" data-uri="2015/articles//a/1/8/6/4/2/2/1/20161020085927_1.jpg.jpg" data-original-width="1920" data-original-height="1080" onclick=""/>

The regular expression simply trims off the .jpg.jpg.*$ and the image ends up broken. Seems like the link works without the /EG11/resize/600x-1/quality/80/format/jpg suffix, but the regular expression trims everything after the first .jpg.

Readability goes away in December

https://mercury.postlight.com/web-parser/

This is remarkably difficult in the wiki and normal editor for no good reason

For Safari Reading List

require "plist"
begin
  bookmarks_xml = `plutil -convert xml1 -o - ~/Library/Safari/Bookmarks.plist`
  urls = Plist.parse_xml(bookmarks_xml)["Children"].select{|e|e["Title"]=="com.apple.ReadingList"}[0]["Children"].map{|e|e["URLString"]}
  puts urls
rescue NoMethodError => e
  puts "Reading List is empty."
end

Research further:
https://developers.google.com/gmail/api/

• Remove dependency on wget from article_renderer.rb
• Remove dependency on convert from article_renderer.rb
• Deal with encoding issues on sites like arstechnica.com
• Deal with image issues on sites using multiple DPI images (i.e. have a 2x in the image tag)

https://github.com/minimagick/minimagick

Try to keep everything self-contained via bundler.

Use one of these:

• https://github.com/huyha85/opengraph_parser
• https://github.com/intridea/opengraph

Add a link to the end of stories with "Archive"

See:
https://getpocket.com/developer/docs/v3/modify#action_archive

Potentially use this:
https://github.com/turadg/pocket-ruby