org.sindice.siren.analysis.attributes

Interface NodeNumericTermAttribute

All Superinterfaces:

org.apache.lucene.util.Attribute

All Known Implementing Classes:

NodeNumericTermAttributeImpl

public interface NodeNumericTermAttribute
extends org.apache.lucene.util.Attribute

Expert: This class provides an Attribute for the NumericTokenizer for indexing numeric values that can be used by NodeNumericRangeQuery.

This attribute provides a stream of tokens which iterates over the different precisions of a given numeric value.

The string representation of each precision is prefixed by:

• the numeric type of the value;
• the precision step;

This prefix is in fact encoding the numeric type and precision step inside the dictionary. This prefix is necessary for two reasons:

• it avoids overlapping value of different numeric type, and therefore avoid getting false-positive;
• enables better clustering of the values of a particular numeric type in the dictionary.

Method Summary

Methods

Modifier and Type	Method and Description
org.apache.lucene.document.FieldType.NumericType	getNumericType() Return the numeric type of the value
int	getPrecisionStep() Returns the precision step
int	getShift() Returns the current shift value
int	getValueSize() Returns the value size in bits (32 for float, int; 64 for double, long)
boolean	incrementShift(org.apache.lucene.analysis.tokenattributes.CharTermAttribute termAtt) Increment the shift and generate the next token.
void	init(org.apache.lucene.document.FieldType.NumericType numericType, long value, int valSize) Initialise this attribute
void	resetShift() Reset the current shift value to 0
void	setPrecisionStep(int precisionStep) Set the precision step

Method Detail

getNumericType

org.apache.lucene.document.FieldType.NumericType getNumericType()

Return the numeric type of the value

getShift

int getShift()

Returns the current shift value

Undefined before first call to #incrementShift(CharTermAttribute, NumericType)

getValueSize

int getValueSize()

Returns the value size in bits (32 for float, int; 64 for double, long)

setPrecisionStep

void setPrecisionStep(int precisionStep)

Set the precision step

getPrecisionStep

int getPrecisionStep()

Returns the precision step

init

void init(org.apache.lucene.document.FieldType.NumericType numericType,
        long value,
        int valSize)

Initialise this attribute

resetShift

void resetShift()

Reset the current shift value to 0

incrementShift

boolean incrementShift(org.apache.lucene.analysis.tokenattributes.CharTermAttribute termAtt)

Increment the shift and generate the next token.

The original Lucene's NumericTokenStream.NumericTermAttribute implements TermToBytesRefAttribute. There is a conflict problem with the CharTermAttribute used in higher-level SIREn's analyzers, which also implements TermToBytesRefAttribute. The problem is that the AttributeSource is not able to choose between the two when requested an attribute implementing TermToBytesRefAttribute, e.g., in TermsHashPerField.

The current solution is to fill the BytesRef attribute of the CharTermAttribute with the encoded numeric value.

Returns:

True if there are still tokens, false if we reach the end of the stream.