CODEWITHSUNDEEP
javaparsingstanford-nlpantlr4
skrtbhtngr
skrtbhtngr

Reputation: 2251

Using POS-tagged string with ANTLR

I have a string which I am tagging using Stanford CoreNLP POS Tagger. Now, I want the words in the string to act like values (terminals) and the corresponding tag of each word to act as the token (non-terminal).

For example if the tagged string is:

Three/CD, critics/NNS, review/VBP, a/DT, book/NN, ./.

Then, I want these tags (CD, NNS, VBP, etc.) to be passed to ANTLR 4 parser as token stream with the words in the sentence (Three, critics, review, etc.) being values of these tokens which can be inferred later.

I want to ask if there is an efficient way I can parse the sentence using ANTLR but use my own methods to supply the tokens?

PS: I am using ANTLR 4 IDE plugin for Eclipse. No tag for it yet!

Upvotes: 0

Views: 98

Answers (1)

GRosenberg
GRosenberg

Reputation: 6001

Extend CommonToken to create your own custom token with methods and fields appropriate for the tagging. Extend CommonTokenFactory to supply your tokens to the lexer when called; use Lexer.setTokenFactory() to make the factory available.

Upvotes: 0

Related Questions