Application that converts a sentence from English to German using the semantics of the sentence. This is unlike other statistical translators which blindly perform a word to word replacement with probability statistics models.
This approach is different from existing syntactic translators which use probability to guess the translations and ordering of words. We translate using the meaning of the sentence to German. Check this Vauquois triangle for levels of translation
- [optional] Copy .classpath_template and rename the copy to .classpath
- Import this repo as an existing project in eclipse
- Install Python 2.7 if not already installed
- Install Jython from here using the installer and NOT the standalore JAR. Remember the directory where it is installing. We need it later. Pick Standard Installation during the installation process
- Install
nktl
python package as mentioned here. Tip:sudo easy_install pip
to installpip
command on a MAC. - By default
nltk
will be installed inLib/
folder of your python installation directory, usually/Users/<username>/Python2.7
. Copynltk/
andnltk-***egg.info/
folders from that directory toLib/
folder of Jython's installation directory - Update
Jython*.jar
's path in your project's build path in eclipse
-
Download the external libraries from here
-
Update
kparser.jar
and other jar files invendor/
directory paths in your build path in eclipse -
Update the value
<param name="dictionary_path" value="__SOME_PATH__/vendor/resources/WordNet-3.0/dict"/>
inwn_file_properties.xml
file insidewsd_resources/
folder you downloaded from above link -
Run TestKParser.java file in the src/ folder
Note: If you use WordNet Stemmer in nltk, you have to download WordNet corpus using nltk.download()
in python terminal (REPL) and install to default location pointed by the downloader. Other stemmers work just fine.