Package com.tonic.analysis.similarity

Class MethodSimilarityAnalyzer

java.lang.Object

com.tonic.analysis.similarity.MethodSimilarityAnalyzer

public class MethodSimilarityAnalyzer
extends Object

Pairwise method-similarity analyzer over a ClassPool, used to find duplicates and renamed copies.

Constructor Summary

Constructors

Constructor	Description
MethodSimilarityAnalyzer(ClassPool classPool)	Creates an analyzer over the given pool; call buildIndex before querying.

Method Summary

Modifier and Type	Method	Description
void	buildIndex()	Builds signatures for every method with code in the pool, replacing any previous index.
SimilarityResult	compare(MethodSignature sig1, MethodSignature sig2, SimilarityMetric primaryMetric)	Scores two signatures on every metric.
List<SimilarityResult>	findAllSimilar(SimilarityMetric metric, double minScore)	Compares every indexed pair and keeps those scoring at or above the threshold.
List<SimilarityResult>	findDuplicates()	Finds pairs scoring at least 0.95 on the combined metric.
List<SimilarityResult>	findRenamedCopies()	Finds pairs with near-identical opcode sequences but different method names.
List<List<MethodSignature>>	findSimilarityGroups(SimilarityMetric metric, double minScore)	Clusters similar methods into connected components over the pairwise similarity graph.
List<SimilarityResult>	findSimilarTo(MethodSignature target, SimilarityMetric metric, double minScore)	Finds indexed methods similar to a given signature.
List<SimilarityResult>	findSimilarTo(String className, String methodName, String descriptor, SimilarityMetric metric, double minScore)	Finds indexed methods similar to the referenced method.
int	getMethodCount()
MethodSignature	getSignature(String className, String methodName, String descriptor)	Looks up an indexed signature by method reference.
List<MethodSignature>	getSignatures()
void	setProgressCallback(Consumer<String> callback)	Sets a callback to receive progress messages during indexing and comparison.

Methods inherited from class java.lang.Object

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

Constructor Detail

MethodSimilarityAnalyzer

public MethodSimilarityAnalyzer(ClassPool classPool)

Creates an analyzer over the given pool; call buildIndex before querying.

Parameters:

classPool - the pool of classes to analyze

Method Detail

setProgressCallback

public void setProgressCallback(Consumer<String> callback)

Sets a callback to receive progress messages during indexing and comparison.

Parameters:

callback - the consumer of progress strings

buildIndex

public void buildIndex()

Builds signatures for every method with code in the pool, replacing any previous index.

findAllSimilar

public List<SimilarityResult> findAllSimilar(SimilarityMetric metric,
                                             double minScore)

Compares every indexed pair and keeps those scoring at or above the threshold.

Parameters:

metric - the metric to score with

minScore - the minimum overall score to keep

Returns:

matching pairs, best first

findSimilarTo

public List<SimilarityResult> findSimilarTo(String className,
                                            String methodName,
                                            String descriptor,
                                            SimilarityMetric metric,
                                            double minScore)

Finds indexed methods similar to the referenced method.

Parameters:

className - the owning class name

methodName - the method name

descriptor - the method descriptor

metric - the metric to score with

minScore - the minimum overall score to keep

Returns:

matching results, best first; empty if the method is not indexed

findSimilarTo

public List<SimilarityResult> findSimilarTo(MethodSignature target,
                                            SimilarityMetric metric,
                                            double minScore)

Finds indexed methods similar to a given signature.

Parameters:

target - the signature to match against

metric - the metric to score with

minScore - the minimum overall score to keep

Returns:

matching results, best first

compare

public SimilarityResult compare(MethodSignature sig1,
                                MethodSignature sig2,
                                SimilarityMetric primaryMetric)

Scores two signatures on every metric.

Parameters:

sig1 - the first signature

sig2 - the second signature

primaryMetric - the metric that determines the overall score

Returns:

the result holding all metric scores

findDuplicates

public List<SimilarityResult> findDuplicates()

Finds pairs scoring at least 0.95 on the combined metric.

Returns:

the potential duplicate pairs, best first

findRenamedCopies

public List<SimilarityResult> findRenamedCopies()

Finds pairs with near-identical opcode sequences but different method names.

Returns:

the candidate renamed-copy pairs

findSimilarityGroups

public List<List<MethodSignature>> findSimilarityGroups(SimilarityMetric metric,
                                                        double minScore)

Clusters similar methods into connected components over the pairwise similarity graph.

Parameters:

metric - the metric to score with

minScore - the minimum score for an edge

Returns:

groups of two or more methods, largest first

getSignatures

public List<MethodSignature> getSignatures()

Returns:

an unmodifiable view of the indexed signatures

getMethodCount

public int getMethodCount()

Returns:

the number of indexed methods

getSignature

public MethodSignature getSignature(String className,
                                    String methodName,
                                    String descriptor)

Looks up an indexed signature by method reference.

Parameters:

className - the owning class name

methodName - the method name

descriptor - the method descriptor

Returns:

the matching signature, or null if not indexed