com.sharethis.textrank

Class TextRank

java.lang.Object

com.sharethis.textrank.TextRank

All Implemented Interfaces:

Callable<Collection<MetricVector>>

public class TextRank
extends Object
implements Callable<Collection<MetricVector>>

Java implementation of the TextRank algorithm by Rada Mihalcea, et al. http://lit.csci.unt.edu/index.php/Graph-based_NLP

Author:

paco@sharethis.com

Field Summary

Fields

Modifier and Type	Field and Description
protected long	elapsed_time
protected Graph	graph
protected LanguageModel	lang Protected members.
static int	MAX_NGRAM_LENGTH
static long	MAX_WORDNET_GRAPH
static long	MAX_WORDNET_TEXT
protected Map<NGram,MetricVector>	metric_space
static double	MIN_NORMALIZED_RANK
protected Graph	ngram_subgraph
static String	NLP_RESOURCES Public definitions.
protected long	start_time
protected String	text
protected boolean	use_wordnet

Constructor Summary

Constructors

Constructor and Description
TextRank(String res_path, String lang_code) Constructor.

Method Summary

Modifier and Type	Method and Description
Collection<MetricVector>	call() Run the TextRank algorithm on the given semi-structured text (e.g., results of parsed HTML from crawled web content) to build a graph of weighted key phrases.
Graph	getGraph() Accessor for the graph.
void	prepCall(String text, boolean use_wordnet) Prepare to call algorithm with a new text to analyze.
String	toString() Serialize resulting graph to a string.

Methods inherited from class java.lang.Object

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

Field Detail

NLP_RESOURCES

public static final String NLP_RESOURCES

Public definitions.

See Also:

Constant Field Values

MIN_NORMALIZED_RANK

public static final double MIN_NORMALIZED_RANK

See Also:

Constant Field Values

MAX_NGRAM_LENGTH

public static final int MAX_NGRAM_LENGTH

See Also:

Constant Field Values

MAX_WORDNET_TEXT

public static final long MAX_WORDNET_TEXT

See Also:

Constant Field Values

MAX_WORDNET_GRAPH

public static final long MAX_WORDNET_GRAPH

See Also:

Constant Field Values

lang

protected final LanguageModel lang

Protected members.

text

protected String text

use_wordnet

protected boolean use_wordnet

graph

protected Graph graph

ngram_subgraph

protected Graph ngram_subgraph

metric_space

protected Map<NGram,MetricVector> metric_space

start_time

protected long start_time

elapsed_time

protected long elapsed_time

Constructor Detail

TextRank

public TextRank(String res_path,
                String lang_code)

Constructor.

Method Detail

prepCall

public void prepCall(String text,
                     boolean use_wordnet)

Prepare to call algorithm with a new text to analyze.

call

public Collection<MetricVector> call()

Run the TextRank algorithm on the given semi-structured text (e.g., results of parsed HTML from crawled web content) to build a graph of weighted key phrases.

Specified by:

call in interface Callable<Collection<MetricVector>>

getGraph

public Graph getGraph()

Accessor for the graph.

toString

public String toString()

Serialize resulting graph to a string.

Overrides:

toString in class Object