org.apache.lucene.mark
Class ContextMarker

Object
  ContextMarker

public class ContextMarker
extends Object

Workhorse class that handles marking hits, context surrounding hits, and search terms.

Created: Dec 26, 2004

Author:

Martin Haye

Field Summary

private MarkCollector	collector Client instance which receives the resulting marks
private String	field Field name (for debugging)
private WordIter	iter0 Iterator used for locating the start of the hit/context
private WordIter	iter1 Iterator used for locating the end of the hit/context
static int	MARK_ALL_TERMS See MARK_NO_TERMS
static int	MARK_CONTEXT_TERMS See MARK_NO_TERMS
static int	MARK_NO_TERMS The following modes can be used for term marking: MARK_NO_TERMS: Terms are not marked MARK_SPAN_TERMS: Search terms are marked only within span hits.
static int	MARK_SPAN_TERMS See MARK_NO_TERMS
private int	maxContext Target size (in chars) of the context surrounding each hit
private int	prevEndWord End of the previous context
private Set	stopSet Set of stop-words to avoid marking outside of hits
private int	termMode Whether to mark terms inside/outside hits, context, etc.
private Set	terms Set of search terms to mark
private int	termsMarkedPos Word position up to which we've marked all terms
private MarkPos	tmpPos Used to temporary position storage

Constructor Summary

ContextMarker(int maxContext, int termMode, Set terms, Set stopSet, WordIter wordIter, MarkCollector collector, String field)
Construct a new marker

Method Summary

(package private) void	emitMarks(Span posSpan, MarkPos contextStart, MarkPos contextEnd) Emit all the marks for the given hit.
(package private) void	findContext(Span posSpan, Span nextSpan, MarkPos contextStart, MarkPos contextEnd) Locate the start and end of context for the given hit.
void	mark(Span[] posOrderSpans, int maxContext) Mark a series of spans.
static void	markField(FieldSpans fieldSpans, String field, WordIter iter, int maxContext, int termMode, Set stopSet, MarkCollector collector) Mark context, spans, and terms a field of data.
void	markField(String field, FieldSpans fieldSpans, MarkCollector collector) Mark context, spans, and terms within the given field of this document.
void	markField(String field, FieldSpans fieldSpans, WordIter iter, int maxContext, int termMode, Set stopSet, MarkCollector collector) Mark context, spans, and terms within the given field of this document.
private void	markTerms(WordIter iter, int fromPos, int toPos, boolean markStopWords) Mark terms up to (but not including) 'wordPos'

Methods inherited from class Object

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

Field Detail

MARK_NO_TERMS

public static final int MARK_NO_TERMS

The following modes can be used for term marking:

MARK_NO_TERMS: Terms are not marked

MARK_SPAN_TERMS: Search terms are marked only within span hits.

MARK_CONTEXT_TERMS: Search terms are marked within span hits and, if found, within the context surrounding those hits.

MARK_ALL_TERMS: Search terms are marked wherever they are found.

See Also:

Constant Field Values

MARK_SPAN_TERMS

public static final int MARK_SPAN_TERMS

See MARK_NO_TERMS

See Also:

Constant Field Values

MARK_CONTEXT_TERMS

public static final int MARK_CONTEXT_TERMS

See MARK_NO_TERMS

See Also:

Constant Field Values

MARK_ALL_TERMS

public static final int MARK_ALL_TERMS

See MARK_NO_TERMS

See Also:

Constant Field Values

maxContext

private int maxContext

Target size (in chars) of the context surrounding each hit

iter0

private WordIter iter0

Iterator used for locating the start of the hit/context

iter1

private WordIter iter1

Iterator used for locating the end of the hit/context

collector

private MarkCollector collector

Client instance which receives the resulting marks

terms

private Set terms

Set of search terms to mark

stopSet

private Set stopSet

Set of stop-words to avoid marking outside of hits

termMode

private int termMode

Whether to mark terms inside/outside hits, context, etc. See MARK_SPAN_TERMS, etc.

termsMarkedPos

private int termsMarkedPos

Word position up to which we've marked all terms

tmpPos

private MarkPos tmpPos

Used to temporary position storage

prevEndWord

private int prevEndWord

End of the previous context

field

private String field

Field name (for debugging)

Constructor Detail

ContextMarker

public ContextMarker(int maxContext,
                     int termMode,
                     Set terms,
                     Set stopSet,
                     WordIter wordIter,
                     MarkCollector collector,
                     String field)

Construct a new marker

Method Detail

markField

public void markField(String field,
                      FieldSpans fieldSpans,
                      MarkCollector collector)

Mark context, spans, and terms within the given field of this document. Context around each hit will be up to 80 characters (including the text of the hit itself). Search terms will only be marked within hits. If you would like to override these defaults, use one of the other variations of this method.

Parameters:

field - field name to mark

fieldSpans - spans to mark with

collector - collector to receive the marks

markField

public void markField(String field,
                      FieldSpans fieldSpans,
                      WordIter iter,
                      int maxContext,
                      int termMode,
                      Set stopSet,
                      MarkCollector collector)

Mark context, spans, and terms within the given field of this document.

Parameters:

field - field name to mark

iter - iterator over the words in the field

maxContext - target number of characters for context around each hit (including the text of the hit itself.) 80 is often a good choice. Specify zero to turn off context marking.

termMode - what areas to mark hits - see MARK_NO_TERMS.

stopSet - set of stop words to avoid marking outside hits

collector - collector to receive the marks

markField

public static void markField(FieldSpans fieldSpans,
                             String field,
                             WordIter iter,
                             int maxContext,
                             int termMode,
                             Set stopSet,
                             MarkCollector collector)

Mark context, spans, and terms a field of data.

Parameters:

field - field name to mark

iter - iterator over the words in the field

maxContext - target number of characters for context around each hit (including the text of the hit itself.) 80 is often a good choice. Specify zero to turn off context marking.

termMode - what areas to mark hits - see MARK_NO_TERMS.

stopSet - set of stop words to avoid marking outside hits

collector - collector to receive the marks

mark

public void mark(Span[] posOrderSpans,
                 int maxContext)

Mark a series of spans.

Parameters:

posOrderSpans - Spans to mark, in ascending position order.

maxContext - Target # of chars for context around hits (0 for none)

findContext

void findContext(Span posSpan,
                 Span nextSpan,
                 MarkPos contextStart,
                 MarkPos contextEnd)

Locate the start and end of context for the given hit.

Parameters:

posSpan - hit for which to find context

nextSpan - following hit (or null if none)

contextStart - OUT: start of context

contextEnd - OUT: end of context

emitMarks

void emitMarks(Span posSpan,
               MarkPos contextStart,
               MarkPos contextEnd)

Emit all the marks for the given hit.

Parameters:

posSpan - hit for which to emit marks

contextStart - start of context (or null if context disabled)

contextEnd - end of context (or null if context disabled)

markTerms

private void markTerms(WordIter iter,
                       int fromPos,
                       int toPos,
                       boolean markStopWords)

Mark terms up to (but not including) 'wordPos'