GSoC/GCI Archive
Google Code-in 2012 Apertium

Fix 200 testvoc errors in Armenian nouns

completed by: conor-f

mentors: Francis Tyers

<spectie> the task is for each item in the list
<spectie> do the following:
<spectie> (1) look in the apertium-hye-eng.hye-eng.dix
<spectie> (1a) if the word is there, with the wrong animacy, change the animacy. (e.g. եվրաինտեգրում) should be "nn" not "aa"
<spectie> (1b) if the word is not there, look it up on wiktionary, and add the entries
* conor_f nods
<spectie> (1b) (ii) if the word on wiktionary is an alternative spelling, edit the apertium-hye.hye.dix
<spectie> (e.g. http://en.wiktionary.org/wiki/%D5%B0%D5%A5%D5%A9%D5%AB%D5%A9%D5%A5%D6%80%D5%A5%D5%B6)
<spectie>     <e lm="հեթիթերեն"><i>հեթիթերեն</i><par n="հայերեն__n_nn"/></e>
<spectie> and change the entry to:
<spectie>  <e lm="հեթիթերեն" r="LR"><p><l>հեթիթերեն</l><r>խեթերեն</r></p><par n="հայերեն__n_nn"/></e>
<spectie> you might be able to categorise the work to make it quicker
<spectie> by splitting the unknown word list into words that _are_ in the dictionary, and words that _aren't_
<spectie> every ~100 words, run the testvoc again to check that you're not breaking anything ;)