Difference between revisions of "Task ideas for Google Code-in/Language detection in simple-html and apertium-apy"

From Apertium
Jump to navigation Jump to search
 
(2 intermediate revisions by the same user not shown)
Line 4: Line 4:
 
==Objective==
 
==Objective==
   
The language detection in the [[simple HTML]] interface currently uses a 2.9M javascript file. The objective of this set of tasks is to get it to use [[apertium-apy]] for the language detection.
+
The language detection in the [[simple-html]] interface currently uses a 2.9M javascript file. The objective of this set of tasks is to get it to use [[apertium-apy]] for the language detection.
   
 
==Tasks==
 
==Tasks==
Line 18: Line 18:
 
===Use external language detection in simple-html===
 
===Use external language detection in simple-html===
   
  +
Make the [[simple-html]] interface not use 2.9MB javascript module for language detection/identification. Instead it should query [[apertium-apy]] with text to get a list of languages with probabilities.
 
   
 
===Interface behaviour for language guessing===
 
===Interface behaviour for language guessing===
   
  +
Based on results of language detection function, make [[simple-html]] highlight in the menu the ''n'' (e.g., 3) most probable languages, and select the most probable.
   
 
[[Category:Tasks for Google Code-in|Language detection in apertium-apy]]
 
[[Category:Tasks for Google Code-in|Language detection in apertium-apy]]

Latest revision as of 20:49, 13 November 2013


Objective[edit]

The language detection in the simple-html interface currently uses a 2.9M javascript file. The objective of this set of tasks is to get it to use apertium-apy for the language detection.

Tasks[edit]

Implement language detection in apertium-apy[edit]

Currently apertium-apy does not do language detection. Make a new function that allows the language of some input text to be identified. This function should return a dict of languages and probabilities. For this task you will also need to train models for the language identifier.

Localised 'available languages' in apertium-apy[edit]

Make a new function for apertium-apy, is takes as input a language code, and as output gives the list of available pairs, and their translations in the language specified by the language code. You will probably need to know JavaScript and Python.

Use external language detection in simple-html[edit]

Make the simple-html interface not use 2.9MB javascript module for language detection/identification. Instead it should query apertium-apy with text to get a list of languages with probabilities.

Interface behaviour for language guessing[edit]

Based on results of language detection function, make simple-html highlight in the menu the n (e.g., 3) most probable languages, and select the most probable.