New text-parser updates

November 10, 2009 1 min read
  • Modified the container-selection algorithm to increase accuracy, especially to reduce the chances of stripping legitimate content. A bit more extraneous content may now be permitted on certain sites. Still optimizing.
  • Fixed minor bugs.
  • Added multipage detection to more popular sites.
  • Fixed character-encoding issue that would show up as missing apostrophes or dashes on pages such as Chicago Sun-Times and New York Times.
  • Fixed issues with Reuters, Bloomberg, Blogger, BoingBoing, LA Times, New York Times, Times Online UK, Touch Arcade, and Huffington Post.

This is still very much in beta, but it’s getting there. Thanks, everyone.

Related Posts