Full text reading experience

[chuanqisun]

Inspired by #37 and #6, it would be great to have a full text reading experience.

Some considerations:

• The messed up history of formats
  • Atom feeds have dedicated summary field for synopsis. and content field for full text. Many sites still use content for synopsis.
  • RSS feeds have a single description field. Most sites use it for synopsis, few use it for full text. There is a content:encoded field, but its use is unpredictable.
• The default template uses dark theme and accordion. The design aims to provide an index that helps you scan and click through to the actual website. Reading the full text would require a different template
  • Probably a master/detail two column layout. For narrow viewport, a modal might become necessary.
  • Probably a light theme by default, optimized for reading speed.
• Downloading full text can cause performance issue. We need to cache articles individually, and fetch them from the client side only when user wants to read the full text.
• For articles that don't provide full text in the xml feed, or provide some crazy formatting with encoded HTML, we need to either crawl the actual web site or run some formatting/sanitization logic. mercury parser is a good tool for this.