Cured bit rot in the query containment module

SmartDataAnalytics · Feb 12, 2019 · 20a0487 · 20a0487
1 parent 2e45edf
commit 20a0487
Show file tree

Hide file tree

Showing 3 changed files with 60 additions and 25 deletions.
diff --git a/...ontainment/src/main/java/org/aksw/jena_sparql_api/query_containment/index/NodeMapper.java b/...ontainment/src/main/java/org/aksw/jena_sparql_api/query_containment/index/NodeMapper.java
@@ -5,7 +5,7 @@
 import com.codepoetics.protonpack.functions.TriFunction;
 
 public interface NodeMapper<A, B, M, C, V>
-	extends TriFunction<A, B, TreeMapping<A, B, M, V>, Entry<C, V>>
+//	extends TriFunction<A, B, TreeMapping<A, B, M, V>, Entry<C, V>>
 {
-
+	Entry<C, V> apply(A a, B b, TreeMapping<A, B, M, V> treeMapping);
 }
diff --git a/...ht/MainLsqQueryContainmentEvaluation.java → ...na/MainLsqQueryContainmentEvaluation.java b/...ht/MainLsqQueryContainmentEvaluation.java → ...na/MainLsqQueryContainmentEvaluation.java
@@ -1,4 +1,4 @@
-package org.aksw.jena_sparql_api.jgrapht;
+package org.aksw.commons.graph.index.jena;
 
 import java.io.File;
 import java.io.FileInputStream;
@@ -18,25 +18,24 @@
 import java.util.Objects;
 import java.util.Set;
 import java.util.concurrent.TimeUnit;
-import java.util.function.BiFunction;
 import java.util.stream.Collectors;
 
 import javax.persistence.criteria.Root;
 
 import org.aksw.commons.collections.tagmap.TagMapSetTrie;
 import org.aksw.commons.collections.tagmap.ValidationUtils;
 import org.aksw.commons.graph.index.core.SubgraphIsomorphismIndex;
-import org.aksw.commons.graph.index.jena.SubgraphIsomorphismIndexJena;
 import org.aksw.jena_sparql_api.core.SparqlService;
 import org.aksw.jena_sparql_api.mapper.jpa.core.RdfEntityManager;
 import org.aksw.jena_sparql_api.mapper.jpa.core.SparqlEntityManagerFactory;
 import org.aksw.jena_sparql_api.mapper.util.JpaUtils;
 import org.aksw.jena_sparql_api.query_containment.index.NodeMapperOp;
-import org.aksw.jena_sparql_api.query_containment.index.NodeMapperOpEquality;
+import org.aksw.jena_sparql_api.query_containment.index.NodeMapperOpContainment;
 import org.aksw.jena_sparql_api.query_containment.index.OpContext;
-import org.aksw.jena_sparql_api.query_containment.index.QueryContainmentIndex;
-import org.aksw.jena_sparql_api.query_containment.index.QueryContainmentIndexImpl;
 import org.aksw.jena_sparql_api.query_containment.index.ResidualMatching;
+import org.aksw.jena_sparql_api.query_containment.index.SparqlQueryContainmentIndex;
+import org.aksw.jena_sparql_api.query_containment.index.SparqlQueryContainmentIndexImpl;
+import org.aksw.jena_sparql_api.query_containment.index.SparqlTreeMapping;
 import org.aksw.jena_sparql_api.query_containment.index.TreeMapping;
 import org.aksw.jena_sparql_api.stmt.SparqlQueryParserImpl;
 import org.aksw.jena_sparql_api.update.FluentSparqlService;
@@ -49,10 +48,11 @@
 import org.apache.jena.sparql.algebra.Algebra;
 import org.apache.jena.sparql.algebra.Op;
 import org.apache.jena.sparql.util.NodeUtils;
-import org.jgrapht.DirectedGraph;
+import org.jgrapht.Graph;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
+import com.codepoetics.protonpack.functions.TriFunction;
 import com.google.common.base.Stopwatch;
 import com.google.common.collect.ArrayListMultimap;
 import com.google.common.collect.BiMap;
@@ -74,7 +74,7 @@ public static void main(String[] args) throws Exception {
         SparqlService ss = FluentSparqlService.http("http://lsq.aksw.org/sparql").create();
 
         emf.setSparqlService(ss);
-        emf.addScanPackageName(MainSparqlQueryToGraph.class.getPackage().getName());
+        emf.addScanPackageName(SubGraphIsomorphismIndexTests.class.getPackage().getName());
         RdfEntityManager em = emf.getObject();
 
         List<LsqQuery> lsqQueries;
@@ -457,19 +457,22 @@ public static void doRun(List<LsqQuery> lsqQueries) throws Exception {
         }
 
 
-		BiFunction<OpContext, OpContext, NodeMapperOp> nodeMapperFactory = (a, b) -> new NodeMapperOpEquality();
-
+		//BiFunction<OpContext, OpContext, NodeMapperOp> nodeMapperFactory = (a, b) -> new NodeMapperOpEquality();
+
+    	TriFunction<OpContext, OpContext, Table<Op, Op, BiMap<Node, Node>>, NodeMapperOp> nodeMapperFactory = NodeMapperOpContainment::new; //(aContext, bContext) -> new NodeMapperOpContainment(aContext, bContext);
+
+        //TriFunction<OpContext, OpContext, Table<Op, Op, BiMap<Node, Node>>, NodeMapperOp> nodeMapperFactory = NodeMapperOpEquality::new;
         //QueryContainmentIndex<Node, DirectedGraph<Node, Triple>, Node, Op, Op> indexA = QueryContainmentIndexImpl.create(nodeMapper);
         //QueryContainmentIndex<Node, DirectedGraph<Node, Triple>, Node, Op, Op> indexB = QueryContainmentIndexImpl.createFlat(nodeMapper);
 
-        SubgraphIsomorphismIndex<Entry<Node, Long>, DirectedGraph<Node, Triple>, Node> siiTreeTags = SubgraphIsomorphismIndexJena.create();
-        SubgraphIsomorphismIndex<Entry<Node, Long>, DirectedGraph<Node, Triple>, Node> siiFlat = SubgraphIsomorphismIndexJena.createFlat();
-        SubgraphIsomorphismIndex<Entry<Node, Long>, DirectedGraph<Node, Triple>, Node> siiTagBased = SubgraphIsomorphismIndexJena.createTagBased(new TagMapSetTrie<>(NodeUtils::compareRDFTerms));
+        SubgraphIsomorphismIndex<Entry<Node, Long>, Graph<Node, Triple>, Node> siiTreeTags = SubgraphIsomorphismIndexJena.create();
+        SubgraphIsomorphismIndex<Entry<Node, Long>, Graph<Node, Triple>, Node> siiFlat = SubgraphIsomorphismIndexJena.createFlat();
+        SubgraphIsomorphismIndex<Entry<Node, Long>, Graph<Node, Triple>, Node> siiTagBased = SubgraphIsomorphismIndexJena.createTagBased(new TagMapSetTrie<>(NodeUtils::compareRDFTerms));
 
-        SubgraphIsomorphismIndex<Entry<Node, Long>, DirectedGraph<Node, Triple>, Node> siiValidating = ValidationUtils.createValidatingProxy(SubgraphIsomorphismIndex.class, siiTreeTags, siiTagBased);
-        SubgraphIsomorphismIndex<Entry<Node, Long>, DirectedGraph<Node, Triple>, Node> sii = siiTreeTags; //siiValidating;
+        SubgraphIsomorphismIndex<Entry<Node, Long>, Graph<Node, Triple>, Node> siiValidating = ValidationUtils.createValidatingProxy(SubgraphIsomorphismIndex.class, siiTreeTags, siiTagBased);
+        SubgraphIsomorphismIndex<Entry<Node, Long>, Graph<Node, Triple>, Node> sii = siiTreeTags; //siiValidating;
 
-        QueryContainmentIndex<Node, DirectedGraph<Node, Triple>, Node, Op, ResidualMatching> index = QueryContainmentIndexImpl.create(sii, nodeMapperFactory);
+        SparqlQueryContainmentIndex<Node, ResidualMatching>index = SparqlQueryContainmentIndexImpl.create(sii, nodeMapperFactory);
 
         //QueryContainmentIndex<Node, DirectedGraph<Node, Triple>, Node, Op, Op> index = ValidationUtils.createValidatingProxy(QueryContainmentIndex.class, indexA, indexB);
 
@@ -653,7 +656,7 @@ public static void doRun(List<LsqQuery> lsqQueries) throws Exception {
 //                    siiA.printTree();
 
                     //Thread.sleep(5000);
-	        		Multimap<Node, TreeMapping<Op, Op, BiMap<Node, Node>, ResidualMatching>> matches = ArrayListMultimap.create();
+	        		Multimap<Node, SparqlTreeMapping<ResidualMatching>> matches = ArrayListMultimap.create();
 
                     System.out.println("Querying view candidates of: " + e.getKey());                    
                     Op op = e.getValue();
@@ -681,7 +684,8 @@ public static void doRun(List<LsqQuery> lsqQueries) throws Exception {
 
 	               	 // Find all isomorphic ones:
 	               	Set<Node> isomorphicKeys = matches.entries().stream().filter(jj -> {
-	               		TreeMapping<Op, Op, BiMap<Node, Node>, ResidualMatching> treeMapping = jj.getValue();
+	               		SparqlTreeMapping<ResidualMatching> treeMapping = jj.getValue();
+	               		//TreeMapping<Op, Op, BiMap<Node, Node>, ResidualMatching> treeMapping = jj.getValue();
 	               		Op rootA = treeMapping.getaTree().getRoot();
 	               		Op rootB = treeMapping.getbTree().getRoot();
 

diff --git a/...inment/src/test/java/org/aksw/commons/graph/index/jena/SubGraphIsomorphismIndexTests.java b/...inment/src/test/java/org/aksw/commons/graph/index/jena/SubGraphIsomorphismIndexTests.java
@@ -11,25 +11,46 @@
 import org.aksw.commons.jena.jgrapht.PseudoGraphJenaGraph;
 import org.aksw.jena_sparql_api.core.SparqlService;
 import org.aksw.jena_sparql_api.mapper.jpa.core.SparqlEntityManagerFactory;
+import org.aksw.jena_sparql_api.query_containment.core.SparqlQueryContainmentUtils;
 import org.aksw.jena_sparql_api.update.FluentSparqlService;
 import org.apache.jena.graph.Graph;
 import org.apache.jena.graph.Node;
+import org.apache.jena.query.Query;
 import org.apache.jena.rdf.model.Model;
 import org.apache.jena.riot.Lang;
 import org.apache.jena.riot.RDFDataMgr;
+import org.apache.jena.sparql.algebra.Algebra;
+import org.apache.jena.sparql.algebra.Op;
+import org.apache.jena.sparql.algebra.OpAsQuery;
+import org.jgrapht.graph.DefaultGraphType;
 import org.junit.BeforeClass;
-import org.junit.Ignore;
 import org.junit.Test;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
 
 import com.google.common.collect.BiMap;
 import com.google.common.collect.Multimap;
 
 
-@Ignore
+//@Ignore
 public class SubGraphIsomorphismIndexTests {
+
+	private static final Logger logger = LoggerFactory.getLogger(SubGraphIsomorphismIndexTests.class);
 
+
     protected static EntityManager em;
 
+    public static String clearProjection(String rawQueryStr) {
+        Query rawQuery = SparqlQueryContainmentUtils.queryParser.apply(rawQueryStr);
+        Op rawOp = Algebra.compile(rawQuery);
+        Op op = QueryToGraph.normalizeOpReplaceConstants(rawOp);
+        Query query = OpAsQuery.asQuery(op);
+        query.setQueryResultStar(true);
+
+        String result = "" + query;
+        return result;
+    }
+
     @BeforeClass
     public static void setup() throws Exception {
         Model model = RDFDataMgr.loadModel("lsq-sparqlqc-synthetic-simple.ttl", Lang.TURTLE);
@@ -43,10 +64,12 @@ public static void setup() throws Exception {
 
     public SubgraphIsomorphismIndex<String, String, Node> buildIndex(Collection<String> queryIds) {
 
+    	logger.warn("Claiming a simple directed graph - although it may be a pseudo graph");
+
         SubgraphIsomorphismIndex<String, Graph, Node> base =
                 SubgraphIsomorphismIndexWrapper.wrap(
                         SubgraphIsomorphismIndexJena.create(),
-                        PseudoGraphJenaGraph::new);
+                        jenaGraph -> new PseudoGraphJenaGraph(jenaGraph, DefaultGraphType.directedSimple()));
 
         SubgraphIsomorphismIndex<String, String, Node> result =
                 SubgraphIsomorphismIndexWrapper.wrap(base, QueryToGraph::queryToGraph);
@@ -55,7 +78,12 @@ public SubgraphIsomorphismIndex<String, String, Node> buildIndex(Collection<Stri
         for(String queryId : queryIds) {
             LsqQuery lsqQuery = em.find(LsqQuery.class, queryId);
 
-            String queryStr = lsqQuery.getText();
+            // Transform the query so that no constants remain in basic graph patterns
+            // Example: ?s a ?o -> ?s ?p ?o . FILTER(?p = rdf:type)
+
+
+            String rawQueryStr = lsqQuery.getText();
+            String queryStr = clearProjection(rawQueryStr);
             result.put(queryId, queryStr);
         }
 
@@ -70,8 +98,11 @@ public void lookup(SubgraphIsomorphismIndex<String, String, Node> index, Collect
         for(String queryId : queryIds) {
             LsqQuery lsqQuery = em.find(LsqQuery.class, queryId);
 
-            String queryStr = lsqQuery.getText();
+            String rawQueryStr = lsqQuery.getText();
+            String queryStr = clearProjection(rawQueryStr);
             System.out.println("Lookup result for " + queryId + ": " + queryStr);
+
+
             Multimap<String, BiMap<Node, Node>> r = index.lookup(queryStr, false);
 
             r.asMap().forEach((k, isos) -> {