Difference between revisions of "Tatar and Russian"

From Apertium
Jump to navigation Jump to search
Line 79: Line 79:
 
cd ..
 
cd ..
 
cd nursery/apertium-tat-rus
 
cd nursery/apertium-tat-rus
./qa.sh
+
./qa
 
</pre>
 
</pre>
   
<code>./qa.sh</code> runs the whole regression test suite. It requires <code>apertium-eval-translator.pl</code> script from trunk and assumes that your directory structure follows that of the apertium repository:
+
<code>./qa</code> runs the whole regression test suite. It requires <code>apertium-eval-translator.pl</code> script from trunk and assumes that your directory structure follows that of the apertium repository:
 
<pre>
 
<pre>
 
..
 
..

Revision as of 21:59, 11 March 2015

This is a language pair translating from Tatar to Russian. The pair is currently located in nursery.

Current state

Last updated Testvoc (clean or not) Corpus testvoc
(no *, no */@, no */@/#)
Stems in bidix WER, PER on dev. corpus Average WER, PER on unseen texts
08/08/2014 No
  • news(86.3, 86.3, 82.3)
  • wp(83.0, 83.0, 80.1)
  • Quran(85.3, 85.3, 82.1)
  • NT(83.0, 83.0, 80.1)
  • aytmatov(90.0, 90.0 88.4)
5973 74.92%, 58.06% --
  • Testvoc = apertium-tat-rus/testvoc/standard/testvoc.sh
  • Corpus testvoc = apertium-tat-rus/testvoc/corpus/trimmed-coverage.sh. Corpora can be found in the turkiccorpora repository.
    • news = tat.news.2005-2011_300K-sentences.txt.bz2, wp = tat.wikipedia.2013-02-25.txt.bz2, NT = tat.NT.txt.bz2. Others are unambiguous.
  • Number of stems taken from the header "aperium-dixtools format-1line 10 55 apertium-tat-rus.tat-rus.dix" produces.
  • Development corpus = apertium-tat-rus/corpus (-test = tat-rus-nova.txt, -ref = tat-rus-posted.txt). WER / PER results are given when unknown-word marks (stars) are not removed.

Installation

You will need HFST, lttoolbox, vislcg3, apertium and apertium-lex-tools installed on your computer to be able to compile and use apertium-tat-rus.

If you are using a Debian-based distro, the easiest way to get those dependencies is to install them with apt-get from User:Tino Didriksen's repository.

If you are using the Apertium VirtualBox image, all those tools will be already installed.

Apertium-tat-rus fetches morphological transducers and CG-disambiguators of Tatar and Russian from the apertium-tat and apertium-rus directories in the /languages module. So you have to check out and compile these two monolingual packages first:

mkdir languages
cd languages
svn co http://svn.code.sf.net/p/apertium/svn/languages/apertium-tat/
cd apertium-tat
./autogen.sh
make
cd ..

svn co http://svn.code.sf.net/p/apertium/svn/languages/apertium-rus/
cd apertium-rus
./autogen.sh
make
cd ../..

After you're done with that, you have to check out and compile apertium-tat-rus itself, specifying where monolingual packages you've just compiled are located:

mkdir nursery
cd nursery
svn co http://svn.code.sf.net/p/apertium/svn/nursery/apertium-tat-rus/
cd apertium-tat-rus
./autogen.sh --with-lang1=../../languages/apertium-tat/ --with-lang2=../../languages/apertium-rus/
make

You can test the translator now:

echo "Мин китап укыйм." | apertium -d . tat-rus
Я читаю книгу.

cd ../..
mkdir trunk
cd trunk
svn co http://svn.code.sf.net/p/apertium/svn/trunk/apertium-eval-translator/
cd ..
cd nursery/apertium-tat-rus
./qa

./qa runs the whole regression test suite. It requires apertium-eval-translator.pl script from trunk and assumes that your directory structure follows that of the apertium repository:

                        ..
         /                   |                 \
    languages             nursery             trunk
    /        \                \                  |
apertium-tat  apertium-rus  apertium-tat-rus  apertium-eval-translator

Workplan (GSoC 2014)

This is a workplan for development efforts for the Tatar to Russian translator in Google Summer of Code 2014.

Major goals

  • Clean testvoc
  • 10000 top stems in bidix and at least 80% trimmed coverage
  • Constraint grammar of Tatar containing at least 1000 rules, which makes 90-95% of all words unambiguous, with at least 95% retaining the correct analysis.
  • Average WER on unseen texts below 50

Overview

Weeks 1-6 Weeks 7-12 Saturdays
get categor(y/ies) testvoc clean
with one word ->
<- add more stems to categor(y/ies)
while preserving testvoc clean
disambiguation lexical selection adding stems & cleaning testvoc
transfer rules for pending wiki tests (focus on phrases and clauses, not single words)

Weekly schedule

Week Dates Target Achieved Evaluation Notes
Testvoc (category, type) Stems Testvoc clean? Stems
1 19/05—25/05 Nouns, lite -- 236 --
2 26/05—01/06 Adjectives, lite All nouns
from tat.lexc
2141 -- Not all nouns were added. Around 500 are pending. They require adding stems to rus.dix as well.
3 02/06—08/06 Verbs, lite 4106 95.27% 3258 -- Adjectives, verbs added. Some more pending in apertium-rus/dev/to_[add/check].txt
4 09/06—15/06 Adverbs, full 6071 5331 -- Adverbs clean when I comment out everything except adverbs in Root lexicon. 96% if I don't. Probably something else gets translated with adverb(s).
5 16/06—22/06 Numerals, full 8036 5488 --
6 23/06—29/06 Pronouns, full 10000 1. 500 words x 2
2. Try out assimilation evaluation toolkit if it's usable by that time.
Midterm evaluation
Results when unknown word-marks (stars) are not removed
tat-rus/texts/text1.* (full coverage):
WER 66.73%, PER 56.48%
tat-rus/texts/text2.* (not fully covered):
WER 78.42%, PER 63.58%
7 30/06—06/07 Manually disambiguate a Tatar corpus (in a way so that it will be usable in the cg3ide later) -- See apertium-tat/texts/corpus.ana.txt
8 07/07—13/07 --
9 14/07—20/07 --
10 21/07—27/07 --
11 28/07—03/07 Corpus testvoc clean on all of the available corpora -- Difference between error-free coverage of the analyser and error-free coverage of the full translator is between 1.6% and 4% (see stats above).
12 04/08—10/08 Write Constraint Grammar for Tatar --
13 11/08—18/08 All categories, full 10000 Gisting Final evaluation
  • Testvoc-lite (apertium-tat-rus/testvoc/lite$ ./testvoc.sh) for a category means taking one word per each sub-category and making the full paradigm of the word pass through the translator without debug symbols. E.g. "testvoc-lite for nouns clean" means that if we leave only one word in each of the N1, N-COMPOUND-PX etc. lexicons, they will pass through translator without [@#] errors.
  • Till midterm, bilingual dictionary and, if necessary, Russian transducer are expanded with translations of stems taken from apertium-tat.tat.lexc.
  • Evaluation (except for gisting evaluation) is taking words and performing an evaluation for post-edition word error rate (WER). The output for those words should be clean.

See also