Difference between revisions of "User:Jonpan4/gsoc2020proposal"

From Apertium
Jump to navigation Jump to search
 
(17 intermediate revisions by the same user not shown)
Line 18: Line 18:


== Reasons why Google and Apertium should sponsor it and how and who it will benefit in society ==
== Reasons why Google and Apertium should sponsor it and how and who it will benefit in society ==
Universal Dependencies (UD) is a framework for consistent annotation of grammar across many languages through a system of dependency trees. UD Annotatrix helps provide a tool for creating and editing these dependency trees based on the guidelines set by UD. My project will allow for easier visualisations of graphs as well as implement new features to make editing more streamlined. There have been difficult style issues that needed addressing, and by switching to a d3JS-based editor, these style changes can be more easily implemented.
Universal Dependencies (UD) is a framework for consistent annotation of grammar across many languages through a system of dependency trees. UD Annotatrix helps provide a tool for creating and editing these dependency trees based on the guidelines set by UD. One part of my project will allow for easier visualisations of graphs. There have been difficult style issues that needed addressing, and by switching to a d3JS-based editor, these style changes can be more easily implemented. I will also implement/put finishing touches on new features in order to make editing more streamlined.


== Skills and Qualifications ==
== Skills and Qualifications ==
Line 36: Line 36:
!|Goals
!|Goals
|-
|-
! Week 1 (June 1 - June 7)
| Week 1 (June 1 - June 7)
|
|
* Building new graph editor
* Building new graph editor (part 1)
* Adding the necessary visual features like POS labels and token numbers.
* Adding zooming/panning.
* Add ability to split and merge tokens.
|-
|-
! Week 2 (June 8 - June 14)
|Week 2 (June 8 - June 14)
|
|
* Building new graph editor
* Building new graph editor (part 2)
* Fixing styling issues such as staggering the vertical alignment of deprels or shifting tokens to account for deprel labels
* Merging it with the current UD Annotatrix codebase
|-
|-
! Week 3 (June 15 - June 21)
|Week 3 (June 15 - June 21)
|
|
* Fixing vertical alignment view (it is currently very messed up in the current version of UD Annotatrix).
* Better key support
* Better RTL support (automatic detection of rtl text and correct formatting for embedded rtl text).
⚫
* Undo/redo history fixes
|-
|-
! Week 4 (June 22 - June 28)
|Week 4 (June 22 - June 28)
|
|
* Increase support for enhanced dependencies
* Increase support for enhanced dependencies
|-
|-
! Week 5 (June 29 - July 5)
|Week 5 (June 29 - July 5)
|
|
* Active learning for labels
* Active learning for labels
* '''First evaluation''' (June 29 - July 3)
* '''First evaluation''' (June 29 - July 3)
** Deliverable: Fully functional graph editor that fixes all of the current styling issues.
|-
|-
! Week 6 (July 6 - July 12)
|Week 6 (July 6 - July 12)
|
|
* Finishing up Github integration
* Finishing up Github integration
|-
|-
! Week 7 (July 13 - July 19)
|Week 7 (July 13 - July 19)
|
|
* Better support for editing via textbox
* Better support for editing via textbox
|-
|-
! Week 8 (July 20 - July 26)
|Week 8 (July 20 - July 26)
|
|
* Better key support (either globalizing all current key bindings or creating a way to tell the current scope of keybinds)
*
⚫
* Undo/redo history fixes
|-
|-
! Week 9 (July 27 - August 2)
|Week 9 (July 27 - August 2)
|
|
* Morphology filling from API (using APy for tokenisation and analysis of sentences)
*
* Interface for disambiguation
* '''Second evaluation''' (July 27 - July 31)
* '''Second evaluation''' (July 27 - July 31)
** Deliverable: An improved streamline editing process for UD Annotatrix.
|-
|-
! Week 10 (August 3 - August 9)
|Week 10 (August 3 - August 9)
|
|
* Creating an alternate view similar to treex using d3JS.
*
|-
|-
! Week 11 (August 10 - August 16)
|Week 11 (August 10 - August 16)
|
|
* Tooltips to indicate errors
*
* Minor fixes:
** [https://github.com/jonorthwash/ud-annotatrix/issues/347 #347]
** [https://github.com/jonorthwash/ud-annotatrix/issues/335 #335]
* Documentation
|-
|-
! Week 12 (August 17 - August 24)
|Week 12 (August 17 - August 24)
|
|
* Flex time and documentation
*
* '''Final evaluation''' (August 24 - August 31)
** Deliverable: Fixes to some old issues and documentation of the implemented changes in UD Annotatrix.
|}
|}



== Other Commitments ==
== Other Commitments ==
I am probably going to take an online literature class during the summer to fulfill some college req (it will take up about 5 hours a week), but I will definitely have at least 30 free hours a week.
I am probably going to take an online literature class during the summer to fulfill a college requirement (it will take up about 5 hours a week), but I will definitely have at least 30 free hours a week.

Latest revision as of 03:00, 1 April 2020

Contact Information[edit]

Name: Jonathan Pan

Location: California

E-mail: jonpan4@gmail.com

IRC: jjjppp

GitHub: JPJPJPOPOP

Why is it that you are interested in Apertium?[edit]

I participated in GCI and helped mentor with Apertium, and I have worked a little on UD Annotatrix before and wanted to do a more in-depth project with it.

Which of the published tasks are you interested in? What do you plan to do?[edit]

Improvements to UD Annotatrix. The work plan containing the tasks that I plan to do is below.

Reasons why Google and Apertium should sponsor it and how and who it will benefit in society[edit]

Universal Dependencies (UD) is a framework for consistent annotation of grammar across many languages through a system of dependency trees. UD Annotatrix helps provide a tool for creating and editing these dependency trees based on the guidelines set by UD. One part of my project will allow for easier visualisations of graphs. There have been difficult style issues that needed addressing, and by switching to a d3JS-based editor, these style changes can be more easily implemented. I will also implement/put finishing touches on new features in order to make editing more streamlined.

Skills and Qualifications[edit]

I am a freshman majoring in computer science at UC Berkeley. Coding experience includes:

  • HTML
  • CSS
  • Javascript, d3JS, ReactJS
  • Python

Coding challenge for creating a graphing editor using d3JS: https://github.com/JPJPJPOPOP/d3-graph

Workplan[edit]

Week Goals
Week 1 (June 1 - June 7)
  • Building new graph editor (part 1)
  • Adding the necessary visual features like POS labels and token numbers.
  • Adding zooming/panning.
  • Add ability to split and merge tokens.
Week 2 (June 8 - June 14)
  • Building new graph editor (part 2)
  • Fixing styling issues such as staggering the vertical alignment of deprels or shifting tokens to account for deprel labels
  • Merging it with the current UD Annotatrix codebase
Week 3 (June 15 - June 21)
  • Fixing vertical alignment view (it is currently very messed up in the current version of UD Annotatrix).
  • Better RTL support (automatic detection of rtl text and correct formatting for embedded rtl text).
Week 4 (June 22 - June 28)
  • Increase support for enhanced dependencies
Week 5 (June 29 - July 5)
  • Active learning for labels
  • First evaluation (June 29 - July 3)
    • Deliverable: Fully functional graph editor that fixes all of the current styling issues.
Week 6 (July 6 - July 12)
  • Finishing up Github integration
Week 7 (July 13 - July 19)
  • Better support for editing via textbox
Week 8 (July 20 - July 26)
  • Better key support (either globalizing all current key bindings or creating a way to tell the current scope of keybinds)
  • Undo/redo history fixes
Week 9 (July 27 - August 2)
  • Morphology filling from API (using APy for tokenisation and analysis of sentences)
  • Interface for disambiguation
  • Second evaluation (July 27 - July 31)
    • Deliverable: An improved streamline editing process for UD Annotatrix.
Week 10 (August 3 - August 9)
  • Creating an alternate view similar to treex using d3JS.
Week 11 (August 10 - August 16)
  • Tooltips to indicate errors
  • Minor fixes:
  • Documentation
Week 12 (August 17 - August 24)
  • Flex time and documentation
  • Final evaluation (August 24 - August 31)
    • Deliverable: Fixes to some old issues and documentation of the implemented changes in UD Annotatrix.

Other Commitments[edit]

I am probably going to take an online literature class during the summer to fulfill a college requirement (it will take up about 5 hours a week), but I will definitely have at least 30 free hours a week.

Retrieved from "https://wiki.apertium.org/w/index.php?title=User:Jonpan4/gsoc2020proposal&oldid=71972"