so2jia / acl2013-charparsing

ACL2013-CharParsing

This is an example script to run the code of my ACL 2013 paper Chinese Parsing Exploiting Characters

Please follow the below steps:
(1) Download the code of ZPar.
(2) Modify the "setting" file, finding the following line
     #chinese constituency parser: acl13(default), jcad
   ensuring that CHINESE_CONSTITUENCY_PARSER jcad
(3) run "python configure.py chinese conparser"
(4) run command cmake . generate Makefile
(5) compile using make
(6) copy the bin files in the bin folder into folder ZPar-RunExample
(7) run "./ctb60-exp.sh" in ZPar-RunExample

================ There is another folder named TreeGeneration, telling one user that how to generate training corpus and character-level phrase structures.

• Training corpus generation
     python ../tools/binarize_noroot.py ../tools/rule.txt test.cs test.zpar
   python ../tools/append_wordstructure.py -d ../tools/word.inner.zpar -i test.zpar -o test.chars.zpar -r remain.words
• Character-level phrase structures
     python ../tools/unbinarizeclt.py -i test.chars.zpar -o test.chars.cs

================ Any problems, please concat [email protected].