feat: make the extraction of a code matching optional

Author: AlexanderSchultheiss

docker-resources/custom.properties

@@ -12,6 +12,8 @@ extraction.print-enabled=false
 # Should changes to only the presence condition of source code be ignored? If set to true, VEVOS will only consider
 # the old presence condition for a line of source code that has otherwise not been changed
 extraction.ignore-pc-changes=true
+# Should the extraction also extract a matching of the code before and after a commit's changes?
+extraction.extract-code-matching=false
 # Path in which diff detective keeps its intermediate results. No need to change this.
 diff-detective.output-dir=ground-truth/dd
 # Path in which diff detective keeps the repos. No need to change this.

docker-resources/verification.properties

@@ -8,6 +8,8 @@ extraction.print-enabled=false
 # Should changes to only the presence condition of source code be ignored? If set to true, VEVOS will only consider
 # the old presence condition for a line of source code that has otherwise not been changed
 extraction.ignore-pc-changes=true
+# Should the extraction also extract a matching of the code before and after a commit's changes?
+extraction.extract-code-matching=false
 # Path in which diff detective keeps its intermediate results. No need to change this.
 diff-detective.output-dir=ground-truth/dd
 # Path in which diff detective keeps the repos. No need to change this.

docker-resources/without_linux.properties

@@ -8,6 +8,8 @@ extraction.print-enabled=false
 # Should changes to only the presence condition of source code be ignored? If set to true, VEVOS will only consider
 # the old presence condition for a line of source code that has otherwise not been changed
 extraction.ignore-pc-changes=true
+# Should the extraction also extract a matching of the code before and after a commit's changes?
+extraction.extract-code-matching=false
 # Path in which diff detective keeps its intermediate results. No need to change this.
 diff-detective.output-dir=ground-truth/dd
 # Path in which diff detective keeps the repos. No need to change this.

src/main/java/org/variantsync/vevos/extraction/FastExtraction.java

@@ -34,6 +34,8 @@ public class FastExtraction {
             = "diff-detective.num-threads";
     public static final String BATCH_SIZE
             = "diff-detective.batch-size";
+    private static final String EXTRACT_CODE_MATCHING
+            = "extraction.extract-code-matching";
     private final Properties properties;
 
     public FastExtraction(Properties properties) {
@@ -158,7 +160,9 @@ private BiConsumer<Repository, Path> buildRunner() {
             Path resultsRoot = extractionDir.resolve(repo.getRepositoryName());
             boolean printEnabled = Boolean.parseBoolean(this.properties.getProperty(PRINT_ENABLED));
 
-            FastPCAnalysis analysis = new FastPCAnalysis(printEnabled, resultsRoot, Boolean.parseBoolean(properties.getProperty(IGNORE_PC_CHANGES)));
+            FastPCAnalysis analysis = new FastPCAnalysis(printEnabled, resultsRoot,
+                    Boolean.parseBoolean(properties.getProperty(IGNORE_PC_CHANGES)),
+                    Boolean.parseBoolean(properties.getProperty(EXTRACT_CODE_MATCHING)));
             final BiFunction<Repository, Path, Analysis> AnalysisFactory = (r, out) -> new Analysis(
                     "PCAnalysis",
                     List.of(
@@ -169,14 +173,14 @@ private BiConsumer<Repository, Path> buildRunner() {
             );
             final int availableProcessors;
             String numThreads = this.properties.getProperty(NUM_THREADS);
-            if (numThreads == null || numThreads.trim().equals("") || numThreads.trim().equals("0")) {
+            if (numThreads == null || numThreads.trim().isEmpty() || numThreads.trim().equals("0")) {
                 availableProcessors = Runtime.getRuntime().availableProcessors();
             } else {
                 availableProcessors = Integer.parseInt(numThreads);
             }
             final int batchSize;
             String configuredSize = this.properties.getProperty(BATCH_SIZE);
-            if (configuredSize == null || configuredSize.trim().equals("") || configuredSize.trim().equals("0")) {
+            if (configuredSize == null || configuredSize.trim().isEmpty() || configuredSize.trim().equals("0")) {
                 batchSize = 256;
             } else {
                 batchSize = Integer.parseInt(configuredSize);

src/main/java/org/variantsync/vevos/extraction/FastPCAnalysis.java

@@ -37,13 +37,15 @@ public class FastPCAnalysis implements Analysis.Hooks, PCAnalysis {
 
     private final boolean ignorePCChanges;
     private final Path resultsRoot;
+    private final boolean extractCodeMatching;
 
-    public FastPCAnalysis(boolean printEnabled, Path resultsRoot, boolean ignorePCChanges) {
+    public FastPCAnalysis(boolean printEnabled, Path resultsRoot, boolean ignorePCChanges, boolean extractCodeMatching) {
         this.printEnabled = printEnabled;
         this.resultsRoot = resultsRoot;
         this.threadBatches = new ConcurrentHashMap<>();
         this.failedCommits = ConcurrentHashMap.newKeySet();
         this.ignorePCChanges = ignorePCChanges;
+        this.extractCodeMatching = extractCodeMatching;
         try {
             Files.createDirectories(resultsRoot);
         } catch (IOException e) {
@@ -143,18 +145,21 @@ public void endCommit(Analysis analysis) {
 
         String pcAsCSVBefore = groundTruthBefore.asPcCsvString();
         String pcAsCSVAfter = groundTruthAfter.asPcCsvString();
-        String matchingAsCSVBefore = groundTruthBefore.asMatchingCsvString();
-        String matchingAsCSVAfter = groundTruthAfter.asMatchingCsvString();
 
         Serde.writeToFile(commitSaveDir.resolve(CODE_VARIABILITY_CSV_BEFORE),
                 pcAsCSVBefore);
         Serde.writeToFile(commitSaveDir.resolve(CODE_VARIABILITY_CSV_AFTER),
                 pcAsCSVAfter);
 
-        Serde.writeToFile(commitSaveDir.resolve(CODE_MATCHING_CSV_BEFORE),
-                matchingAsCSVBefore);
-        Serde.writeToFile(commitSaveDir.resolve(CODE_MATCHING_CSV_AFTER),
-                matchingAsCSVAfter);
+        if (extractCodeMatching) {
+            String matchingAsCSVBefore = groundTruthBefore.asMatchingCsvString();
+            String matchingAsCSVAfter = groundTruthAfter.asMatchingCsvString();
+
+            Serde.writeToFile(commitSaveDir.resolve(CODE_MATCHING_CSV_BEFORE),
+                    matchingAsCSVBefore);
+            Serde.writeToFile(commitSaveDir.resolve(CODE_MATCHING_CSV_AFTER),
+                    matchingAsCSVAfter);
+        }
 
         Serde.writeToFile(commitSaveDir.resolve(COMMIT_MESSAGE_FILE), commit.getFullMessage());
 

src/main/java/org/variantsync/vevos/extraction/FullExtraction.java

@@ -23,11 +23,14 @@
 import java.util.function.BiConsumer;
 import java.util.function.BiFunction;
 
+
 public class FullExtraction {
     public static final String PRINT_ENABLED
             = "extraction.print-enabled";
     public static final String IGNORE_PC_CHANGES
             = "extraction.ignore-pc-changes";
+    public static final String EXTRACT_CODE_MATCHING
+            = "extraction.extract-code-matching";
     public static final String GT_SAVE_DIR
             = "extraction.gt-save-dir";
     public static final String DATASET_FILE
@@ -176,7 +179,7 @@ public static void quitOnError() {
 
     private BiConsumer<Repository, Path> buildRunner(String diffDetectiveCache) {
         return (repo, repoOutputDir) -> {
-            FullPCAnalysis analysis = new FullPCAnalysis(Path.of(diffDetectiveCache), Boolean.parseBoolean(properties.getProperty(IGNORE_PC_CHANGES)));
+            FullPCAnalysis analysis = new FullPCAnalysis(Path.of(diffDetectiveCache), Boolean.parseBoolean(properties.getProperty(IGNORE_PC_CHANGES)),  Boolean.parseBoolean(properties.getProperty(EXTRACT_CODE_MATCHING)));
             final BiFunction<Repository, Path, Analysis> AnalysisFactory = (r, out) -> new Analysis(
                     "PCAnalysis",
                     List.of(

src/main/java/org/variantsync/vevos/extraction/FullPCAnalysis.java

@@ -23,11 +23,14 @@ public class FullPCAnalysis implements Analysis.Hooks, PCAnalysis {
     private final Hashtable<String, GroundTruth> groundTruthMap;
     private final Path diffDetectiveCache;
     private final boolean ignorePCChanges;
+    // TODO: extract code matching during full analysis
+    private final boolean extractCodeMatching;
 
-    public FullPCAnalysis(Path diffDetectiveCache, boolean ignorePCChanges) {
+    public FullPCAnalysis(Path diffDetectiveCache, boolean ignorePCChanges, boolean extractCodeMatching) {
         this.groundTruthMap = new Hashtable<>();
         this.diffDetectiveCache = diffDetectiveCache;
         this.ignorePCChanges = ignorePCChanges;
+        this.extractCodeMatching = extractCodeMatching;
     }
 
     @Override

src/main/java/org/variantsync/vevos/extraction/PCAnalysis.java

@@ -17,7 +17,7 @@ public interface PCAnalysis {
      * @param fileGT          The ground truth that is modified by analyzing the node
      * @param node            The node that is to be analyzed
      * @param time            Whether we should handle the node as before or after the edit
-     * @param ignorePCChanges
+     * @param ignorePCChanges Whether changes to only the presence condition should be ignored
      */
     static void analyzeNode(FileGT.Mutable fileGT, DiffNode<DiffLinesLabel> node, Time time, boolean ignorePCChanges) throws MatchingException {
         if (time == Time.BEFORE && node.diffType == DiffType.ADD) {
@@ -65,9 +65,8 @@ static void analyzeNode(FileGT.Mutable fileGT, DiffNode<DiffLinesLabel> node, Ti
         // Grow the root mapping
         fileGT.growIfRequired(toLine);
 
-        // Insert the matchings
+        // Insert the matching for artifacts
         if (!node.isAnnotation()) {
-            // set the matching
             fileGT.setMatching(currentRange, counterpartRange);
         }
 

src/main/resources/extraction.properties

@@ -3,6 +3,8 @@
 #####################################
 extraction.print-enabled=false
 extraction.ignore-pc-changes=true
+# Should the extraction also extract a matching of the code before and after a commit's changes?
+extraction.extract-code-matching=false
 extraction.gt-save-dir=/home/alex/data/EXTRACTION/ground-truth
 #diff-detective.dataset-file = src/main/resources/debug.md
 diff-detective.dataset-file=src/main/resources/verification.md