opennlp.tools.tokenize
Class TokenizerME

java.lang.Object
  opennlp.tools.tokenize.TokenizerME

All Implemented Interfaces:

Tokenizer

public class TokenizerME
extends Object

A Tokenizer for converting raw text into separated tokens. It uses Maximum Entropy to make its decisions. The features are loosely based off of Jeff Reynar's UPenn thesis "Topic Segmentation: Algorithms and Applications.", which is available from his homepage: .

This tokenizer needs a statistical model to tokenize a text which reproduces the tokenization observed in the training data used to create the model. The TokenizerModel class encapsulates the model and provides methods to create it from the binary representation.

A tokenizer instance is not thread safe. For each thread one tokenizer must be instantiated which can share one TokenizerModel instance to safe memory.

To train a new model {#train(String, Iterator, boolean) method can be used.

Sample usage:

InputStream modelIn;

...

TokenizerModel model = TokenizerModel(modelIn);

Tokenizer tokenizer = new TokenizerME(model);

String tokens[] = tokenizer.tokenize("A sentence to be tokenized.");

See Also:

Tokenizer, TokenizerModel, TokenSample

Field Summary

static Pattern	alphaNumeric Alpha-Numeric Pattern
static String	NO_SPLIT Constant indicates no token split.
static String	SPLIT Constant indicates a token split.

Constructor Summary

TokenizerME(TokenizerModel model)

Method Summary

double[]	getTokenProbabilities() Returns the probabilities associated with the most recent calls to AbstractTokenizer.tokenize(String) or tokenizePos(String).
String[]	tokenize(String s) Splits a string into its atomic parts
Span[]	tokenizePos(String d) Tokenizes the string.
static TokenizerModel	train(String languageCode, ObjectStream<TokenSample> samples, boolean useAlphaNumericOptimization) Trains a model for the TokenizerME with a default cutoff of 5 and 100 iterations.
static TokenizerModel	train(String languageCode, ObjectStream<TokenSample> samples, boolean useAlphaNumericOptimization, int cutoff, int iterations) Trains a model for the TokenizerME.
boolean	useAlphaNumericOptimization() Returns the value of the alpha-numeric optimization flag.

Methods inherited from class java.lang.Object

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

Field Detail

SPLIT

public static final String SPLIT

Constant indicates a token split.

See Also:

Constant Field Values

NO_SPLIT

public static final String NO_SPLIT

Constant indicates no token split.

See Also:

Constant Field Values

alphaNumeric

public static final Pattern alphaNumeric

Alpha-Numeric Pattern

Constructor Detail

TokenizerME

public TokenizerME(TokenizerModel model)

Method Detail

getTokenProbabilities

public double[] getTokenProbabilities()

Returns the probabilities associated with the most recent calls to AbstractTokenizer.tokenize(String) or tokenizePos(String).

Returns:

probability for each token returned for the most recent call to tokenize. If not applicable an empty array is returned.

tokenizePos

public Span[] tokenizePos(String d)

Tokenizes the string.

Parameters:

d - The string to be tokenized.

Returns:

A span array containing individual tokens as elements.

train

public static TokenizerModel train(String languageCode,
                                   ObjectStream<TokenSample> samples,
                                   boolean useAlphaNumericOptimization,
                                   int cutoff,
                                   int iterations)
                            throws IOException

Trains a model for the TokenizerME.

Parameters:

languageCode - the language of the natural text

samples - the samples used for the training.

useAlphaNumericOptimization - - if true alpha numerics are skipped

cutoff - number of times a feature must be seen to be considered

iterations - number of iterations to train the maxent model

Returns:

the trained TokenizerModel

Throws:

IOException - it throws an IOException if an IOException is thrown during IO operations on a temp file which is created during training. Or if reading from the ObjectStream fails.

train

public static TokenizerModel train(String languageCode,
                                   ObjectStream<TokenSample> samples,
                                   boolean useAlphaNumericOptimization)
                            throws IOException,
                                   ObjectStreamException

Trains a model for the TokenizerME with a default cutoff of 5 and 100 iterations.

Parameters:

languageCode - the language of the natural text

samples - the samples used for the training.

useAlphaNumericOptimization - - if true alpha numerics are skipped

Returns:

the trained TokenizerModel

Throws:

IOException - it throws an IOException if an IOException is thrown during IO operations on a temp file which is

ObjectStreamException - if reading from the ObjectStream fails created during training.

useAlphaNumericOptimization

public boolean useAlphaNumericOptimization()

Returns the value of the alpha-numeric optimization flag.

Returns:

true if the tokenizer should use alpha-numeric optimization, false otherwise.

tokenize

public String[] tokenize(String s)

Description copied from interface: Tokenizer

Splits a string into its atomic parts

Specified by:

tokenize in interface Tokenizer

Parameters:

s - The string to be tokenized.

Returns:

The String[] with the individual tokens as the array elements.