Ideas for Google Summer of Code/UD Annotatrix

From Apertium
Jump to navigation Jump to search

UD Annotatrix is a tool for annotating Universal Dependencies. The objective of this project is to extend it in useful ways:


  • General: Relates to the whole codebase
  • Interface: Relates to the front end graph drawing and buttons/tabs etc. (HTML, CSS, JS)
  • Backend: The JS that does format parsing and validation (JS)
  • Server: The code that runs the server (Python)
Task General Interface Backend Server
Comprehensive testing framework x
Interface for disambiguation (in CG mode) x
Integrate d3's graphing library and TreeX x
Support for ambiguous sentences x x
Full keyboard support x
Support for enhanced dependencies x x
Integration with GitHub accounts x
Collaborative editting support x x x
Search functionality x x x
Full logging support x
Automatic punctuation support x
Active learning x x
Automatic validation x x x(?)
Morphology filling from API x x(?)
Decent support for labels (like GitHub) x x
  1. Comprehensive testing framework
    • Find out how the code works, and add tests for functionalities (interface and backend).
  2. Integrate d3's graphing library and TreeX
    • Better graph/edge layouts
  3. Full keyboard support
  4. Collaborative editting support

Coding challenge

  • Download UD Annotatrix
  • Fix one issue in the server version
  • Fix one issue in the client
  • Annotate 20 trees
  • Proof of concept for graphing with d3 (hard-coded values)