Oraki

Bringing wikipedia to the hearing world

Summary

Oraki is an implementation of a mobile talking wikipedia. We download wikipedia and then turn the text into speech.

Parts

Backend

The oraki backend consists of the wikipeda parser. We use beautiful soup to take the raw knowledge of wikipedia and convert it into manageable chunks.

Endpoints

• /search
  • GET
    • request
      • query - your search term (e.g. "Lebron")
      • limit - (optional; default 10) number of results to return
    • response
      • list of titles (e.g. ["Lebron James", "Lebron (Author)", "etc."])
• /article_info
  • GET
    • request
      • title - the title of the article requested
    • response
      • sections
        • list of sections where each section constains
          • name - the name of the section
          • paragrapahs - list of paragraphs
    • errors
      • 404 -> title not found

Client

iOS application which downloads the text for the backend and converts it into wav files using text to speech. These wav files are later played back.

Components

1. SearchViewController - controls article searching
2. ArticleViewController - controls presenting and playing articles
3. FliteManager - manages Flite's text to speech computations

Resources

Flite CMU for text to speech Beautiful soup for html parsing