Constituency Tree and Dependency Tree

[Ramesh Kumar]

Hi everyone, 

      Is there any way to get a constituency tree and dependency tree using NLTK only; I am not taking about using Stanford Parser in NLTK. But I want to know, whether NLTK has its own Parser that includes constituency and dependency tree?

Thanks 

[Alex Rudnick]

Hey Ramesh,

I don't think it does, at least not to the extent that you would hope.
Parsing is a pretty complex topic, and the grammars that you need to
build a robust parser are both quite large and sensitive to domain
differences.

If you're not familiar with the background for why this would be true,
the NLTK book chapter on the topic is quite good:
http://www.nltk.org/book/ch08.html

Your best bet, I think, is to use some external parser like Stanford
Parser or MaltParser -- and both of those have nice robust models for
several different languages -- and hook them up to NLTK.

Best of luck! Let us know how it goes!

--
-- alexr