gov.llnl.ontology.text.parse
Class MaltParser

java.lang.Object
  gov.llnl.ontology.text.parse.MaltParser

All Implemented Interfaces:

Parser

Direct Known Subclasses:

MaltLinearParser, MaltSvmParser

public class MaltParser
extends Object
implements Parser

A Parser wrapper around the Malt Parser.

Author:

Keith Stevens

Constructor Summary

MaltParser(String maltParserModelPath, opennlp.tools.tokenize.Tokenizer tokenizer, opennlp.tools.postag.POSTagger tagger)
Creates a new MaltParser using the provided model paths.

Method Summary

String	parseText(String header, String document) Returns a dependency parse tree formatted in the default CoNLL format.
String	parseText(String header, StringPair[] sentence) Returns a dependency parse tree formatted in the default CoNLL format.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

MaltParser

public MaltParser(String maltParserModelPath,
                  opennlp.tools.tokenize.Tokenizer tokenizer,
                  opennlp.tools.postag.POSTagger tagger)

Creates a new MaltParser using the provided model paths. Note that this Parser cannot be readily used within a map reduce job.

Method Detail

parseText

public String parseText(String header,
                        String document)

Returns a dependency parse tree formatted in the default CoNLL format. The returned string may represent multiple parse tress, espeically if the given document contains multiple sentences. Each parse tree will be padded with new lines. If header is not null, then it will be first line of each parse tree, otherwise only padded parse trees will be returned.

Specified by:

parseText in interface Parser

Parameters:

header - The header text that should label each parsed tree

document - The raw text to be parsed

parseText

public String parseText(String header,
                        StringPair[] sentence)

Returns a dependency parse tree formatted in the default CoNLL format. The returned string will contain only one parse tree If header is not null, then it will be first line of the parse tree, otherwise only the parse tree will be returned.

Specified by:

parseText in interface Parser

Parameters:

header - The header text that should label each parsed tree

sentence - The part of speech tagged words in a sentence.