GH-5182 improve performance of SHACL sh:pattern

eclipse-rdf4j · Nov 7, 2024 · 32d167b · 32d167b
1 parent 3e4f94f
commit 32d167b
Show file tree

Hide file tree

Showing 2 changed files with 57 additions and 48 deletions.
diff --git a/...ava/org/eclipse/rdf4j/sail/shacl/ast/constraintcomponents/PatternConstraintComponent.java b/...ava/org/eclipse/rdf4j/sail/shacl/ast/constraintcomponents/PatternConstraintComponent.java
@@ -17,6 +17,7 @@
 import java.util.Objects;
 import java.util.Set;
 import java.util.function.Function;
+import java.util.regex.Pattern;
 
 import org.eclipse.rdf4j.model.IRI;
 import org.eclipse.rdf4j.model.Literal;
@@ -30,11 +31,16 @@
 import org.eclipse.rdf4j.sail.shacl.ast.planNodes.PatternFilter;
 import org.eclipse.rdf4j.sail.shacl.ast.planNodes.PlanNode;
 import org.eclipse.rdf4j.sail.shacl.wrapper.data.ConnectionsGroup;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
 
 public class PatternConstraintComponent extends AbstractSimpleConstraintComponent {
 
+	private static final Logger logger = LoggerFactory.getLogger(PatternConstraintComponent.class);
+
 	String pattern;
 	String flags;
+	private final Pattern compiledPattern;
 
 	public PatternConstraintComponent(String pattern, String flags) {
 		super();
@@ -44,6 +50,52 @@ public PatternConstraintComponent(String pattern, String flags) {
 		if (flags == null) {
 			this.flags = "";
 		}
+
+		if (flags != null && !flags.isEmpty()) {
+			int flag = 0b0;
+
+			if (flags.contains("i")) {
+				flag = flag | Pattern.CASE_INSENSITIVE;
+				logger.trace("PatternFilter constructed with case insensitive flag");
+			}
+
+			if (flags.contains("d")) {
+				flag = flag | Pattern.UNIX_LINES;
+				logger.trace("PatternFilter constructed with UNIX lines flag");
+			}
+
+			if (flags.contains("m")) {
+				flag = flag | Pattern.MULTILINE;
+				logger.trace("PatternFilter constructed with multiline flag");
+			}
+
+			if (flags.contains("s")) {
+				flag = flag | Pattern.DOTALL;
+				logger.trace("PatternFilter constructed with dotall flag");
+			}
+
+			if (flags.contains("u")) {
+				flag = flag | Pattern.UNICODE_CASE;
+				logger.trace("PatternFilter constructed with unicode case flag");
+			}
+
+			if (flags.contains("x")) {
+				flag = flag | Pattern.COMMENTS;
+				logger.trace("PatternFilter constructed with comments flag");
+			}
+
+			if (flags.contains("U")) {
+				flag = flag | Pattern.UNICODE_CHARACTER_CLASS;
+				logger.trace("PatternFilter constructed with unicode character class flag");
+			}
+
+			this.compiledPattern = Pattern.compile(pattern, flag);
+			logger.trace("PatternFilter constructed with pattern: {} and flags: {}", pattern, flags);
+
+		} else {
+			this.compiledPattern = Pattern.compile(pattern, 0b0);
+			logger.trace("PatternFilter constructed with pattern: {} and no flags", pattern);
+		}
 	}
 
 	@Override
@@ -87,7 +139,7 @@ public ConstraintComponent deepClone() {
 
 	@Override
 	Function<PlanNode, FilterPlanNode> getFilterAttacher(ConnectionsGroup connectionsGroup) {
-		return (parent) -> new PatternFilter(parent, pattern, flags, connectionsGroup);
+		return (parent) -> new PatternFilter(parent, compiledPattern, connectionsGroup);
 	}
 
 	@Override

diff --git a/core/sail/shacl/src/main/java/org/eclipse/rdf4j/sail/shacl/ast/planNodes/PatternFilter.java b/core/sail/shacl/src/main/java/org/eclipse/rdf4j/sail/shacl/ast/planNodes/PatternFilter.java
@@ -34,53 +34,9 @@ public class PatternFilter extends FilterPlanNode {
 
 	private final Pattern pattern;
 
-	public PatternFilter(PlanNode parent, String pattern, String flags, ConnectionsGroup connectionsGroup) {
+	public PatternFilter(PlanNode parent, Pattern pattern, ConnectionsGroup connectionsGroup) {
 		super(parent, connectionsGroup);
-		if (flags != null && !flags.isEmpty()) {
-			int flag = 0b0;
-
-			if (flags.contains("i")) {
-				flag = flag | Pattern.CASE_INSENSITIVE;
-				logger.trace("PatternFilter constructed with case insensitive flag");
-			}
-
-			if (flags.contains("d")) {
-				flag = flag | Pattern.UNIX_LINES;
-				logger.trace("PatternFilter constructed with UNIX lines flag");
-			}
-
-			if (flags.contains("m")) {
-				flag = flag | Pattern.MULTILINE;
-				logger.trace("PatternFilter constructed with multiline flag");
-			}
-
-			if (flags.contains("s")) {
-				flag = flag | Pattern.DOTALL;
-				logger.trace("PatternFilter constructed with dotall flag");
-			}
-
-			if (flags.contains("u")) {
-				flag = flag | Pattern.UNICODE_CASE;
-				logger.trace("PatternFilter constructed with unicode case flag");
-			}
-
-			if (flags.contains("x")) {
-				flag = flag | Pattern.COMMENTS;
-				logger.trace("PatternFilter constructed with comments flag");
-			}
-
-			if (flags.contains("U")) {
-				flag = flag | Pattern.UNICODE_CHARACTER_CLASS;
-				logger.trace("PatternFilter constructed with unicode character class flag");
-			}
-
-			this.pattern = Pattern.compile(pattern, flag);
-			logger.trace("PatternFilter constructed with pattern: {} and flags: {}", pattern, flags);
-
-		} else {
-			this.pattern = Pattern.compile(pattern, 0b0);
-			logger.trace("PatternFilter constructed with pattern: {} and no flags", pattern);
-		}
+		this.pattern = pattern;
 	}
 
 	private static Literal str(Value argValue, ValueFactory valueFactory) {
@@ -104,8 +60,9 @@ boolean checkTuple(Reference t) {
 		Value literal = t.get().getValue();
 		literal = str(literal, SimpleValueFactory.getInstance());
 
-		if (literal == null)
+		if (literal == null) {
 			return false;
+		}
 
 		if (QueryEvaluationUtility.isStringLiteral(literal)) {
 			boolean result = pattern.matcher(((Literal) literal).getLabel()).find();