Package org.apache.lucene.demo.knn

Class DemoEmbeddings

java.lang.Object

org.apache.lucene.demo.knn.DemoEmbeddings

public class DemoEmbeddings
extends Object

This class provides computeEmbedding(String) and computeEmbedding(Reader) for calculating "semantic" embedding vectors for textual input.

Constructor Summary

Constructors

Constructor	Description
DemoEmbeddings(KnnVectorDict vectorDict)	Sole constructor

Method Summary

Modifier and Type	Method	Description
float[]	computeEmbedding(Reader input)	Tokenize and lower-case the input, look up the tokens in the dictionary, and sum the token vectors.
float[]	computeEmbedding(String input)	Tokenize and lower-case the input, look up the tokens in the dictionary, and sum the token vectors.

Methods inherited from class java.lang.Object

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

Constructor Details

DemoEmbeddings

public DemoEmbeddings(KnnVectorDict vectorDict)

Sole constructor

Parameters:

vectorDict - a token to vector dictionary

Method Details

computeEmbedding

public float[] computeEmbedding(String input)
                         throws IOException

Tokenize and lower-case the input, look up the tokens in the dictionary, and sum the token vectors. Unrecognized tokens are ignored. The resulting vector is normalized to unit length.

Parameters:

input - the input to analyze

Returns:

the KnnVector for the input

Throws:

IOException

computeEmbedding

public float[] computeEmbedding(Reader input)
                         throws IOException

Tokenize and lower-case the input, look up the tokens in the dictionary, and sum the token vectors. Unrecognized tokens are ignored. The resulting vector is normalized to unit length.

Parameters:

input - the input to analyze

Returns:

the KnnVector for the input

Throws:

IOException