com.carrotsearch.labs.langid.ILangIdClassifier Maven / Gradle / Ivy
package com.carrotsearch.labs.langid;
import java.nio.ByteBuffer;
import java.util.List;
/**
* Language detection (classifier) public interface.
*
*
* Use case scenarios are: streaming ({@link #reset()},
* {@link #append(CharSequence)}, {@link #classify(boolean)}) and single-call
* classification ({@link #classify(CharSequence, boolean)}).
*
*
* Pay attention to contracts on individual methods (the objects they return are
* mutable and their contents will change on subsequent calls).
*
*
* Thread safety: in general instances of this interface will
* not be thread safe unless marked otherwise. See particular
* implementation for details.
*
* @see LangIdV3
*/
public interface ILangIdClassifier {
/**
* Classify the language of an input character sequence. Whether all of the
* sequence or just subsambles will be used for classification is up to the
* implementation (at the moment the whole input is consumed so you may want
* to trim it if it exceeds 2k characters which is typically enough to extract
* a high-quality language profile).
*
*
* This method is an all-in-one call to {@link #reset()},
* {@link #append(CharSequence)} and {@link #classify(boolean)}.
*
* @param str
* The input character sequence to identify.
* @param normalizeConfidence
* Normalize prediction confidence to 0-1 range.
* @return Returns the most likely language in which str is
* written. May return null if no data (or not enough
* data) is available.
*/
public abstract DetectedLanguage classify(CharSequence str,
boolean normalizeConfidence);
/**
* Reset internal buffers and state to start classifying a new example.
*/
public abstract void reset();
/**
* Update internal buffers and feature vectors with more text.
*/
public abstract void append(CharSequence str);
/**
* Update internal buffers and feature vectors with more UTF8-encoded text.
* Care should be given to proper text segmentation into unicode points
* (otherwise broken features may be identified).
*/
public abstract void append(ByteBuffer buffer);
/**
* Update internal buffers and feature vectors with more UTF8-encoded text.
* Care should be given to proper text segmentation into unicode points
* (otherwise broken features may be identified).
*/
public abstract void append(byte[] array, int start, int length);
/**
* Apply classification to the current buffer state. This may be called while
* appending (to abort early if the desired confidence has been reached).
*/
public abstract DetectedLanguage classify(boolean normalizeConfidence);
/**
* Return a list of ranked languages for the current buffer. The list is not
* sorted, cannot be manipulated and will be reused on any subsequent calls to
* this object, including {@link DetectedLanguage} objects inside. If the
* result is to be stored somewhere, it needs to be deeply cloned.
*/
public abstract List rank(boolean normalizeConfidence);
}