org.corpus_tools.salt.util.internal

Class Diff

java.lang.Object

org.corpus_tools.salt.util.internal.Diff

public class Diff
extends Object

This class compares two SDocumentGraph objects for isomorphie. It uses characteristics of document graphs to use a more specific algorithm for checking whether a graph is isomorphic. At first tokens are compared. Both graphs are compared starting with offset "0". Two tokens are the same, when their textual offset and the overlapped text is the same. As with all following SNodes, SAnnotations and SFeatures are checked. The next step is to look for SNodes that are the source of a incoming relations of a SToken. Of these only those are picked, that have SNodes on their SOutgoingrelations, that already have been checked (at this point: only STokens). These SNodes are then compared with each other. Whenever a relation is used in the way described above, the SRelation is checked for SFeatures and SRelations. To adapt the isomorphie check and the computation of differences you can pass an option map via Diff(SDocumentGraph, SDocumentGraph, Map). Possible options are:

• #OPTION_CHECK_ANNOTATION - When true, even all annotations of nodes and relations are used for isomorphie check and difference computation.
• #OPTION_CHECK_ANNOTATION_DIFF - ???
• #OPTION_CHECK_ID - When true identifiers needs to be the same in isomorphie check and difference computation.

Author:

florian, André Röhrig

Field Summary

Fields

Modifier and Type	Field and Description
static String	OPTION_IGNORE_ANNOTATIONS When true, differences in SAnnotation have no influence on isomorphie and will not produce any differences.
static String	OPTION_IGNORE_FEATURES When true, differences in SFeature have no influence on isomorphie and will not produce any differences.
static String	OPTION_IGNORE_ID When true, differences in id have no influence on isomorphie and will not produce any differences.
static String	OPTION_IGNORE_LAYER When true, layers will not be checked when computing isomorphie and differences.
static String	OPTION_IGNORE_META_ANNOTATIONS When true, differences in SMetaAnnotation have no influence on isomorphie and will not produce any differences.
static String	OPTION_IGNORE_NAME When true, differences in name have no influence on isomorphie and will not produce any differences.
static String	OPTION_IGNORE_PROCESSING_ANNOTATIONS When true, differences in SProcessingAnnotation have no influence on isomorphie and will not produce any differences.

Constructor Summary

Constructors

Constructor and Description
Diff(SDocumentGraph template, SDocumentGraph other) Initializes Diff object with the two graphs template and other.
Diff(SDocumentGraph template, SDocumentGraph other, Map<String,Boolean> optionMap)

Method Summary

Methods

Modifier and Type	Method and Description
boolean	compareAnnotationContainers(SAnnotationContainer template, SAnnotationContainer other, Set<Difference> subDiffs) Checks whether two SAnnotationContainer objects are isomorph and returns true if this is the case, false otherwise.
boolean	compareIdentifiableElements(IdentifiableElement template, IdentifiableElement other, Set<Difference> subDiffs) Checks whether two IdentifiableElement objects are isomorph.
boolean	compareLayers(SDocumentGraph template, SDocumentGraph other, Boolean diff) Compares all layers in template graph with all layers in other graph and searches for isomorphic partners.
boolean	compareRelations(SDocumentGraph template, List<? extends SRelation> templateRels, SDocumentGraph other, List<? extends SRelation> otherRels, Boolean diff) Compares all SPointingRelation of other and template graph.
Set<Difference>	findDiffs() Compares the set graphs and returns if they are isomorphic or not.
com.google.common.collect.BiMap<SNode,SNode>	getIsoNodes() Return BiMap with isomorphic SNodes key= template, value= other
boolean	isIsomorph() Compares the set graphs and returns if they are isomorphic or not.

Methods inherited from class java.lang.Object

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

Field Detail

OPTION_IGNORE_FEATURES

public static final String OPTION_IGNORE_FEATURES

When true, differences in SFeature have no influence on isomorphie and will not produce any differences.

See Also:

Constant Field Values

OPTION_IGNORE_ANNOTATIONS

public static final String OPTION_IGNORE_ANNOTATIONS

When true, differences in SAnnotation have no influence on isomorphie and will not produce any differences.

See Also:

Constant Field Values

OPTION_IGNORE_META_ANNOTATIONS

public static final String OPTION_IGNORE_META_ANNOTATIONS

When true, differences in SMetaAnnotation have no influence on isomorphie and will not produce any differences.

See Also:

Constant Field Values

OPTION_IGNORE_PROCESSING_ANNOTATIONS

public static final String OPTION_IGNORE_PROCESSING_ANNOTATIONS

When true, differences in SProcessingAnnotation have no influence on isomorphie and will not produce any differences.

See Also:

Constant Field Values

OPTION_IGNORE_ID

public static final String OPTION_IGNORE_ID

When true, differences in id have no influence on isomorphie and will not produce any differences.

See Also:

Constant Field Values

OPTION_IGNORE_NAME

public static final String OPTION_IGNORE_NAME

When true, differences in name have no influence on isomorphie and will not produce any differences.

See Also:

Constant Field Values

OPTION_IGNORE_LAYER

public static final String OPTION_IGNORE_LAYER

When true, layers will not be checked when computing isomorphie and differences.

See Also:

Constant Field Values

Constructor Detail

Diff

public Diff(SDocumentGraph template,
    SDocumentGraph other)

Initializes Diff object with the two graphs template and other.

Parameters:

template -

other -

Diff

public Diff(SDocumentGraph template,
    SDocumentGraph other,
    Map<String,Boolean> optionMap)

Method Detail

isIsomorph

public boolean isIsomorph()

Compares the set graphs and returns if they are isomorphic or not. In case of the graphs are not isomorphic, this method does not record all differences. TO speed up the computation. it just detects the first difference and returns false. To get a full list of all differences, use method findDiffs().

Returns:

true, if graphs are isomorphic, false otherwise.

findDiffs

public Set<Difference> findDiffs()

Compares the set graphs and returns if they are isomorphic or not. If graphs are not isomporphic, this method finds and records all differences and lists them. This means the entire graphs have to be compared and could slow down the computation. If you are only interested in the result and not in the particular differences, use method isIsomorph().

Returns:

true, if graphs are isomorphic, false otherwise.

getIsoNodes

public com.google.common.collect.BiMap<SNode,SNode> getIsoNodes()

Return BiMap with isomorphic SNodes key= template, value= other

Returns:

BiMap with isomorphic SNodes

compareIdentifiableElements

public boolean compareIdentifiableElements(IdentifiableElement template,
                                  IdentifiableElement other,
                                  Set<Difference> subDiffs)

Checks whether two IdentifiableElement objects are isomorph. If option OPTION_IGNORE_ID is false, this method returns true, if the id values of both are equal, false otherwise.If subdiffs is not null, the differences will be added to that list, if that list is empty the method returns false when the first difference was found.

Parameters:

templateNode -

otherNode -

subDiffs -

Returns:

true if elements have the same id, false otherwise

compareAnnotationContainers

public boolean compareAnnotationContainers(SAnnotationContainer template,
                                  SAnnotationContainer other,
                                  Set<Difference> subDiffs)

Checks whether two SAnnotationContainer objects are isomorph and returns true if this is the case, false otherwise. If subdiffs is not null, the differences will be added to that list, if that list is empty the method returns false when the first difference was found.
The behavior of this method depends on options: OPTION_IGNORE_META_ANNOTATIONS, OPTION_IGNORE_PROCESSING_ANNOTATIONS, OPTION_IGNORE_ANNOTATIONS and OPTION_IGNORE_FEATURES

Parameters:

templateNode -

otherNode -

subDiffs - if differences shall be computed, this list must not have to be empty

Returns:

compareRelations

public boolean compareRelations(SDocumentGraph template,
                       List<? extends SRelation> templateRels,
                       SDocumentGraph other,
                       List<? extends SRelation> otherRels,
                       Boolean diff)

Compares all SPointingRelation of other and template graph.
When diff is true, the return value is not necessarily correct, since in that case the differences in getDifferences() is computed. To check the isomorphie there must be no differences.

Parameters:

template -

other -

diff -

Returns:

compareLayers

public boolean compareLayers(SDocumentGraph template,
                    SDocumentGraph other,
                    Boolean diff)

Compares all layers in template graph with all layers in other graph and searches for isomorphic partners. To layers are isomorph, when have the same name and contain the same number of nodes and relations and have the same labels.
When diff is true, the return value is not necessarily correct, since in that case the differences in getDifferences() is computed. To check the isomorphie there must be no differences.

Parameters:

template -

other -

diff -

Returns: