uk.ac.man.documentparser.input
Class OTMI

java.lang.Object
  uk.ac.man.documentparser.input.OTMI

All Implemented Interfaces:

java.lang.Iterable<Document>, java.util.Iterator<Document>, DocumentIterator

public class OTMI
extends java.lang.Object
implements DocumentIterator

Field Summary

private org.w3c.dom.Node

root

Constructor Summary

OTMI(java.io.File file)

Method Summary

private java.util.ArrayList<java.lang.String>	getSnippets(org.w3c.dom.NodeList sections)
private java.util.HashMap<java.lang.String,java.lang.Integer>	getWordCounts(org.w3c.dom.NodeList sections)
boolean	hasNext()
java.util.Iterator<Document>	iterator()
Document	next()
void	remove()
void	skip()

Methods inherited from class java.lang.Object

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

Field Detail

root

private org.w3c.dom.Node root

Constructor Detail

OTMI

public OTMI(java.io.File file)

Method Detail

skip

public void skip()

Specified by:

skip in interface DocumentIterator

hasNext

public boolean hasNext()

Specified by:

hasNext in interface java.util.Iterator<Document>

iterator

public java.util.Iterator<Document> iterator()

Specified by:

iterator in interface java.lang.Iterable<Document>

getSnippets

private java.util.ArrayList<java.lang.String> getSnippets(org.w3c.dom.NodeList sections)

getWordCounts

private java.util.HashMap<java.lang.String,java.lang.Integer> getWordCounts(org.w3c.dom.NodeList sections)

next

public Document next()

Specified by:

next in interface java.util.Iterator<Document>

remove

public void remove()

Specified by:

remove in interface java.util.Iterator<Document>