Menu
▾
▴
[Bigdata-commit] SF.net SVN: bigdata:[4530] branches/QUADS_QUERY_BRANCH
|
From: <tho...@us...> - 2011-05-20 15:05:07
|
Revision: 4530
http://bigdata.svn.sourceforge.net/bigdata/?rev=4530&view=rev
Author: thompsonbry
Date: 2011-05-20 15:04:59 +0000 (Fri, 20 May 2011)
Log Message:
-----------
Changed the default minRelevance value to 0.25 (from 0.0) and updated the unit tests and code to match.
Modified Paths:
--------------
branches/QUADS_QUERY_BRANCH/bigdata-rdf/src/java/com/bigdata/rdf/lexicon/ITextIndexer.java
branches/QUADS_QUERY_BRANCH/bigdata-rdf/src/java/com/bigdata/rdf/store/BD.java
branches/QUADS_QUERY_BRANCH/bigdata-rdf/src/test/com/bigdata/rdf/lexicon/TestFullTextIndex.java
branches/QUADS_QUERY_BRANCH/bigdata-sails/src/java/com/bigdata/rdf/sail/BigdataEvaluationStrategyImpl3.java
branches/QUADS_QUERY_BRANCH/bigdata-sails/src/java/com/bigdata/rdf/sail/FreeTextSearchExpander.java
branches/QUADS_QUERY_BRANCH/bigdata-sails/src/java/com/bigdata/rdf/sail/bench/NanoSparqlClient.java
branches/QUADS_QUERY_BRANCH/bigdata-sails/src/test/com/bigdata/rdf/sail/TestBigdataSailEvaluationStrategyImpl.java
branches/QUADS_QUERY_BRANCH/bigdata-sails/src/test/com/bigdata/rdf/sail/TestNamedGraphs.java
branches/QUADS_QUERY_BRANCH/bigdata-sails/src/test/com/bigdata/rdf/sail/TestSearchQuery.java
Modified: branches/QUADS_QUERY_BRANCH/bigdata-rdf/src/java/com/bigdata/rdf/lexicon/ITextIndexer.java
===================================================================
--- branches/QUADS_QUERY_BRANCH/bigdata-rdf/src/java/com/bigdata/rdf/lexicon/ITextIndexer.java 2011-05-20 00:12:12 UTC (rev 4529)
+++ branches/QUADS_QUERY_BRANCH/bigdata-rdf/src/java/com/bigdata/rdf/lexicon/ITextIndexer.java 2011-05-20 15:04:59 UTC (rev 4530)
@@ -79,40 +79,42 @@
*/
public boolean getIndexDatatypeLiterals();
- /**
- * Do free text search
- *
- * @param query
- * The query (it will be parsed into tokens).
- * @param languageCode
- * The language code that should be used when tokenizing the
- * query -or- <code>null</code> to use the default {@link Locale}
- * ).
- * @param prefixMatch
- * When <code>true</code>, the matches will be on tokens which
- * include the query tokens as a prefix. This includes exact
- * matches as a special case when the prefix is the entire token,
- * but it also allows longer matches. For example,
- * <code>free</code> will be an exact match on <code>free</code>
- * but a partial match on <code>freedom</code>. When
- * <code>false</code>, only exact matches will be made.
- * @param minCosine
- * The minimum cosine that will be returned.
- * @param maxCosine
- * The maximum cosine that will be returned. Useful for
- * evaluating in relevance ranges.
- * @param maxRank
- * The upper bound on the #of hits in the result set.
- * @param matchAllTerms
- * if true, return only hits that match all search terms
- * @param timeout
- * The timeout -or- ZERO (0) for NO timeout (this is equivalent
- * to using {@link Long#MAX_VALUE}).
- * @param unit
- * The unit in which the timeout is expressed.
- *
- * @return The result set.
- */
+ /**
+ * Do free text search
+ *
+ * @param query
+ * The query (it will be parsed into tokens).
+ * @param languageCode
+ * The language code that should be used when tokenizing the
+ * query -or- <code>null</code> to use the default {@link Locale}
+ * ).
+ * @param prefixMatch
+ * When <code>true</code>, the matches will be on tokens which
+ * include the query tokens as a prefix. This includes exact
+ * matches as a special case when the prefix is the entire token,
+ * but it also allows longer matches. For example,
+ * <code>free</code> will be an exact match on <code>free</code>
+ * but a partial match on <code>freedom</code>. When
+ * <code>false</code>, only exact matches will be made.
+ * @param minCosine
+ * The minimum cosine that will be returned (in [0:maxCosine]).
+ * If you specify a minimum cosine of ZERO (0.0) you can drag in
+ * a lot of basically useless search results.
+ * @param maxCosine
+ * The maximum cosine that will be returned (in [minCosine:1.0]).
+ * Useful for evaluating in relevance ranges.
+ * @param maxRank
+ * The upper bound on the #of hits in the result set.
+ * @param matchAllTerms
+ * if true, return only hits that match all search terms
+ * @param timeout
+ * The timeout -or- ZERO (0) for NO timeout (this is equivalent
+ * to using {@link Long#MAX_VALUE}).
+ * @param unit
+ * The unit in which the timeout is expressed.
+ *
+ * @return The result set.
+ */
public Hiterator<A> search(final String query, final String languageCode,
final boolean prefixMatch,
final double minCosine, final double maxCosine,
Modified: branches/QUADS_QUERY_BRANCH/bigdata-rdf/src/java/com/bigdata/rdf/store/BD.java
===================================================================
--- branches/QUADS_QUERY_BRANCH/bigdata-rdf/src/java/com/bigdata/rdf/store/BD.java 2011-05-20 00:12:12 UTC (rev 4529)
+++ branches/QUADS_QUERY_BRANCH/bigdata-rdf/src/java/com/bigdata/rdf/store/BD.java 2011-05-20 15:04:59 UTC (rev 4530)
@@ -165,40 +165,63 @@
* }
*
* </pre>
+ *
+ * The default is {@value #DEFAULT_MAX_HITS}.
*/
final URI MAX_HITS = new URIImpl(SEARCH_NAMESPACE+"maxHits");
-
+
/**
- * Magic predicate used to query for free text search metadata. Use
- * in conjunction with {@link #SEARCH} as follows:
- * <p>
- * <pre>
- *
- * select ?s
- * where {
- * ?s bd:search "scale-out RDF triplestore" .
- * ?s bd:minRelevance "0.5"^^xsd:double .
- * }
- *
- * </pre>
+ * The default for {@link #MAX_HITS}.
*/
+ final int DEFAULT_MAX_HITS = Integer.MAX_VALUE;
+
+ /**
+ * Magic predicate used to query for free text search metadata. Use in
+ * conjunction with {@link #SEARCH} as follows:
+ * <p>
+ *
+ * <pre>
+ *
+ * select ?s
+ * where {
+ * ?s bd:search "scale-out RDF triplestore" .
+ * ?s bd:minRelevance "0.5"^^xsd:double .
+ * }
+ *
+ * </pre>
+ *
+ * The relevance scores are in [0.0:1.0]. You should NOT specify a minimum
+ * relevance of ZERO (0.0) as this can drag in way too many unrelated
+ * results. The default is {@value #DEFAULT_MIN_RELEVANCE}.
+ */
final URI MIN_RELEVANCE = new URIImpl(SEARCH_NAMESPACE+"minRelevance");
+ final double DEFAULT_MIN_RELEVANCE = 0.25d;
+
+ /**
+ * Magic predicate used to query for free text search metadata. Use in
+ * conjunction with {@link #SEARCH} as follows:
+ * <p>
+ *
+ * <pre>
+ *
+ * select ?s
+ * where {
+ * ?s bd:search "scale-out RDF triplestore" .
+ * ?s bd:maxRelevance "0.9"^^xsd:double .
+ * }
+ *
+ * </pre>
+ *
+ * The relevance scores are in [0.0:1.0]. The default maximum relevance is
+ * {@value #DEFAULT_MAX_RELEVANCE}.
+ */
+ final URI MAX_RELEVANCE = new URIImpl(SEARCH_NAMESPACE+"maxRelevance");
+
/**
- * Magic predicate used to query for free text search metadata. Use
- * in conjunction with {@link #SEARCH} as follows:
- * <p>
- * <pre>
- *
- * select ?s
- * where {
- * ?s bd:search "scale-out RDF triplestore" .
- * ?s bd:maxRelevance "0.9"^^xsd:double .
- * }
- *
- * </pre>
+ * The default value for {@link #MAX_RELEVANCE} unless overridden.
*/
- final URI MAX_RELEVANCE = new URIImpl(SEARCH_NAMESPACE+"maxRelevance");
+ final double DEFAULT_MAX_RELEVANCE = 1.0d;
/**
* Magic predicate used to query for free text search metadata. Use
@@ -216,7 +239,16 @@
*/
final URI MATCH_ALL_TERMS = new URIImpl(SEARCH_NAMESPACE+"matchAllTerms");
+ final boolean DEFAULT_MATCH_ALL_TERMS = false;
+
+ final boolean DEFAULT_PREFIX_MATCH = false;
+
/**
+ * The default timeout for a free text search (milliseconds).
+ */
+ final long DEFAULT_TIMEOUT = Long.MAX_VALUE;
+
+ /**
* Sesame has the notion of a "null" graph. Any time you insert a statement
* into a quad store and the context position is not specified, it is
* actually inserted into this "null" graph. If SPARQL <code>DATASET</code>
Modified: branches/QUADS_QUERY_BRANCH/bigdata-rdf/src/test/com/bigdata/rdf/lexicon/TestFullTextIndex.java
===================================================================
--- branches/QUADS_QUERY_BRANCH/bigdata-rdf/src/test/com/bigdata/rdf/lexicon/TestFullTextIndex.java 2011-05-20 00:12:12 UTC (rev 4529)
+++ branches/QUADS_QUERY_BRANCH/bigdata-rdf/src/test/com/bigdata/rdf/lexicon/TestFullTextIndex.java 2011-05-20 15:04:59 UTC (rev 4530)
@@ -136,8 +136,9 @@
minCosine, 1.0d/* maxCosine */,
Integer.MAX_VALUE/* maxRank */,
false/* matchAllTerms */,
- 2L/* timeout */,
- TimeUnit.SECONDS);
+ Long.MAX_VALUE,//2L/* timeout */,
+ TimeUnit.MILLISECONDS// TimeUnit.SECONDS
+ );
// assertEquals("#hits", (long) expected.length, itr.size());
Modified: branches/QUADS_QUERY_BRANCH/bigdata-sails/src/java/com/bigdata/rdf/sail/BigdataEvaluationStrategyImpl3.java
===================================================================
--- branches/QUADS_QUERY_BRANCH/bigdata-sails/src/java/com/bigdata/rdf/sail/BigdataEvaluationStrategyImpl3.java 2011-05-20 00:12:12 UTC (rev 4529)
+++ branches/QUADS_QUERY_BRANCH/bigdata-sails/src/java/com/bigdata/rdf/sail/BigdataEvaluationStrategyImpl3.java 2011-05-20 15:04:59 UTC (rev 4530)
@@ -30,7 +30,6 @@
import org.openrdf.query.algebra.BNodeGenerator;
import org.openrdf.query.algebra.Bound;
import org.openrdf.query.algebra.Compare;
-import org.openrdf.query.algebra.Compare.CompareOp;
import org.openrdf.query.algebra.CompareAll;
import org.openrdf.query.algebra.CompareAny;
import org.openrdf.query.algebra.Datatype;
@@ -64,7 +63,6 @@
import org.openrdf.query.algebra.Regex;
import org.openrdf.query.algebra.SameTerm;
import org.openrdf.query.algebra.StatementPattern;
-import org.openrdf.query.algebra.StatementPattern.Scope;
import org.openrdf.query.algebra.Str;
import org.openrdf.query.algebra.TupleExpr;
import org.openrdf.query.algebra.UnaryTupleOperator;
@@ -72,6 +70,8 @@
import org.openrdf.query.algebra.ValueConstant;
import org.openrdf.query.algebra.ValueExpr;
import org.openrdf.query.algebra.Var;
+import org.openrdf.query.algebra.Compare.CompareOp;
+import org.openrdf.query.algebra.StatementPattern.Scope;
import org.openrdf.query.algebra.evaluation.impl.EvaluationStrategyImpl;
import org.openrdf.query.algebra.evaluation.iterator.FilterIterator;
import org.openrdf.query.algebra.helpers.QueryModelVisitorBase;
@@ -83,12 +83,12 @@
import com.bigdata.bop.IConstant;
import com.bigdata.bop.IConstraint;
import com.bigdata.bop.IPredicate;
-import com.bigdata.bop.IPredicate.Annotations;
import com.bigdata.bop.IValueExpression;
import com.bigdata.bop.IVariable;
import com.bigdata.bop.IVariableOrConstant;
import com.bigdata.bop.NV;
import com.bigdata.bop.PipelineOp;
+import com.bigdata.bop.IPredicate.Annotations;
import com.bigdata.bop.ap.Predicate;
import com.bigdata.bop.bindingSet.ListBindingSet;
import com.bigdata.bop.constraint.INBinarySearch;
@@ -109,7 +109,6 @@
import com.bigdata.rdf.internal.constraints.IsLiteralBOp;
import com.bigdata.rdf.internal.constraints.IsURIBOp;
import com.bigdata.rdf.internal.constraints.MathBOp;
-import com.bigdata.rdf.internal.constraints.MathBOp.MathOp;
import com.bigdata.rdf.internal.constraints.NotBOp;
import com.bigdata.rdf.internal.constraints.OrBOp;
import com.bigdata.rdf.internal.constraints.RangeBOp;
@@ -117,15 +116,16 @@
import com.bigdata.rdf.internal.constraints.SPARQLConstraint;
import com.bigdata.rdf.internal.constraints.SameTermBOp;
import com.bigdata.rdf.internal.constraints.StrBOp;
+import com.bigdata.rdf.internal.constraints.MathBOp.MathOp;
import com.bigdata.rdf.lexicon.LexiconRelation;
import com.bigdata.rdf.model.BigdataValue;
import com.bigdata.rdf.sail.BigdataSail.Options;
import com.bigdata.rdf.sail.sop.SOp;
import com.bigdata.rdf.sail.sop.SOp2BOpUtility;
import com.bigdata.rdf.sail.sop.SOpTree;
-import com.bigdata.rdf.sail.sop.SOpTree.SOpGroup;
import com.bigdata.rdf.sail.sop.SOpTreeBuilder;
import com.bigdata.rdf.sail.sop.UnsupportedOperatorException;
+import com.bigdata.rdf.sail.sop.SOpTree.SOpGroup;
import com.bigdata.rdf.spo.DefaultGraphSolutionExpander;
import com.bigdata.rdf.spo.ExplicitSPOFilter;
import com.bigdata.rdf.spo.ISPO;
@@ -2181,12 +2181,12 @@
final Iterator<IHit> itr = (Iterator)database.getLexiconRelation()
.getSearchEngine().search(label, languageCode,
- false/* prefixMatch */,
- 0d/* minCosine */,
- 1.0d/* maxCosine */,
- Integer.MAX_VALUE/* maxRank */,
- false/* matchAllTerms */,
- 0L/* timeout */,
+ BD.DEFAULT_PREFIX_MATCH,//false/* prefixMatch */,
+ BD.DEFAULT_MIN_RELEVANCE,//0d/* minCosine */,
+ BD.DEFAULT_MAX_RELEVANCE,//1.0d/* maxCosine */,
+ BD.DEFAULT_MAX_HITS,//Integer.MAX_VALUE/* maxRank */,
+ BD.DEFAULT_MATCH_ALL_TERMS,//false/* matchAllTerms */,
+ BD.DEFAULT_TIMEOUT,//0/* timeout */,
TimeUnit.MILLISECONDS);
// ensure that named graphs are handled correctly for quads
Modified: branches/QUADS_QUERY_BRANCH/bigdata-sails/src/java/com/bigdata/rdf/sail/FreeTextSearchExpander.java
===================================================================
--- branches/QUADS_QUERY_BRANCH/bigdata-sails/src/java/com/bigdata/rdf/sail/FreeTextSearchExpander.java 2011-05-20 00:12:12 UTC (rev 4529)
+++ branches/QUADS_QUERY_BRANCH/bigdata-sails/src/java/com/bigdata/rdf/sail/FreeTextSearchExpander.java 2011-05-20 15:04:59 UTC (rev 4530)
@@ -169,23 +169,15 @@
prefixMatch = false;
}
- /*
- * FIXME This is using a constant (1000ms) for the timeout on
- * the free text search. That needs to be passed down from the
- * SAIL.
- *
- * @todo Rather than explicitly passing in all of these as
- * parameters to the constructor, why not pass them through as
- * annotations on the magic predicate?
- */
hiterator = textNdx.search(s,
query.getLanguage(),
prefixMatch,
- minRelevance == null ? 0d : minRelevance.doubleValue()/* minCosine */,
- maxRelevance == null ? 1.0d : maxRelevance.doubleValue()/* maxCosine */,
- maxHits == null ? Integer.MAX_VALUE : maxHits.intValue()+1/* maxRank */,
+ minRelevance == null ? BD.DEFAULT_MIN_RELEVANCE : minRelevance.doubleValue()/* minCosine */,
+ maxRelevance == null ? BD.DEFAULT_MAX_RELEVANCE : maxRelevance.doubleValue()/* maxCosine */,
+ maxHits == null ? BD.DEFAULT_MAX_HITS/*Integer.MAX_VALUE*/ : maxHits.intValue()+1/* maxRank */,
matchAllTerms,
- 0L/* timeout */, TimeUnit.MILLISECONDS);
+ BD.DEFAULT_TIMEOUT/*0L*//* timeout */,
+ TimeUnit.MILLISECONDS);
}
return hiterator;
Modified: branches/QUADS_QUERY_BRANCH/bigdata-sails/src/java/com/bigdata/rdf/sail/bench/NanoSparqlClient.java
===================================================================
--- branches/QUADS_QUERY_BRANCH/bigdata-sails/src/java/com/bigdata/rdf/sail/bench/NanoSparqlClient.java 2011-05-20 00:12:12 UTC (rev 4529)
+++ branches/QUADS_QUERY_BRANCH/bigdata-sails/src/java/com/bigdata/rdf/sail/bench/NanoSparqlClient.java 2011-05-20 15:04:59 UTC (rev 4530)
@@ -866,7 +866,7 @@
public final String queryStr;
/** Metadata about each query presentation. */
- public LinkedBlockingQueue<QueryTrial> trials = new LinkedBlockingQueue<QueryTrial>(/* unbounded */);
+ public final LinkedBlockingQueue<QueryTrial> trials = new LinkedBlockingQueue<QueryTrial>(/* unbounded */);
/**
* Total elapsed nanoseconds over all {@link QueryTrial}s for this
@@ -939,7 +939,7 @@
/**
* Order by increasing elapsed time (slowest queries are last).
*/
- public int compareTo(Score o) {
+ public int compareTo(final Score o) {
if (elapsedNanos < o.elapsedNanos)
return -1;
if (elapsedNanos > o.elapsedNanos)
Modified: branches/QUADS_QUERY_BRANCH/bigdata-sails/src/test/com/bigdata/rdf/sail/TestBigdataSailEvaluationStrategyImpl.java
===================================================================
--- branches/QUADS_QUERY_BRANCH/bigdata-sails/src/test/com/bigdata/rdf/sail/TestBigdataSailEvaluationStrategyImpl.java 2011-05-20 00:12:12 UTC (rev 4529)
+++ branches/QUADS_QUERY_BRANCH/bigdata-sails/src/test/com/bigdata/rdf/sail/TestBigdataSailEvaluationStrategyImpl.java 2011-05-20 15:04:59 UTC (rev 4530)
@@ -706,12 +706,14 @@
System.err.println("\"Mike\" = " + sail.getDatabase().getIV(new LiteralImpl("Mike")));
System.err.println("\"Jane\" = " + sail.getDatabase().getIV(new LiteralImpl("Jane")));
- String query =
+ final double minRelevance = 0d;
+ final String query =
"select ?s ?label " +
"where { " +
" ?s <"+RDF.TYPE+"> <"+person+"> . " + // [160, 8, 164], [156, 8, 164]
" ?s <"+RDFS.LABEL+"> ?label . " + // [160, 148, 174], [156, 148, 170]
" ?label <"+search+"> \"Mi*\" . " + // [174, 0, 0]
+ " ?label <"+BD.MIN_RELEVANCE+"> \""+minRelevance+"\" . " +
"}";
{ // evalute it once so i can see it
Modified: branches/QUADS_QUERY_BRANCH/bigdata-sails/src/test/com/bigdata/rdf/sail/TestNamedGraphs.java
===================================================================
--- branches/QUADS_QUERY_BRANCH/bigdata-sails/src/test/com/bigdata/rdf/sail/TestNamedGraphs.java 2011-05-20 00:12:12 UTC (rev 4529)
+++ branches/QUADS_QUERY_BRANCH/bigdata-sails/src/test/com/bigdata/rdf/sail/TestNamedGraphs.java 2011-05-20 15:04:59 UTC (rev 4530)
@@ -1593,6 +1593,7 @@
public void testSearchQuery() throws Exception {
+ final double minRelevance = 0d;
final BigdataSail sail = getSail();
try {
sail.initialize();
@@ -1638,6 +1639,7 @@
"where { " +
" ?y <"+ BD.SEARCH+"> \"Chris*\" . " +
" ?x <"+ RDFS.LABEL.stringValue() + "> ?y . " +
+ " ?y <"+BD.MIN_RELEVANCE+"> \""+minRelevance+"\" . " +
"}";
final TupleQuery tupleQuery = cxn.prepareTupleQuery(
@@ -1671,6 +1673,7 @@
" graph <http://example.org> { " +
" ?y <"+ BD.SEARCH+"> \"Chris*\" . " +
" ?x <"+ RDFS.LABEL.stringValue() + "> ?y ." +
+ " ?y <"+BD.MIN_RELEVANCE+"> \""+minRelevance+"\" . " +
" } . " +
"}";
Modified: branches/QUADS_QUERY_BRANCH/bigdata-sails/src/test/com/bigdata/rdf/sail/TestSearchQuery.java
===================================================================
--- branches/QUADS_QUERY_BRANCH/bigdata-sails/src/test/com/bigdata/rdf/sail/TestSearchQuery.java 2011-05-20 00:12:12 UTC (rev 4529)
+++ branches/QUADS_QUERY_BRANCH/bigdata-sails/src/test/com/bigdata/rdf/sail/TestSearchQuery.java 2011-05-20 15:04:59 UTC (rev 4530)
@@ -300,7 +300,9 @@
final Set<Value> expected = new HashSet<Value>();
- expected.add(new LiteralImpl("Yellow Rose"));
+ // Note: Whether or not this solution is present depends on the
+ // default value for minCosine.
+// expected.add(new LiteralImpl("Yellow Rose"));
expected.add(new LiteralImpl("Old Yellow House"));
@@ -316,11 +318,13 @@
final BindingSet solution = itr.next();
- System.out.println("solution[" + i + "] : " + solution);
+ if (log.isInfoEnabled())
+ log.info("solution[" + i + "] : " + solution);
final Value actual = solution.getValue("X");
- System.out.println("X[" + i + "] = " + actual + " ("
+ if (log.isInfoEnabled())
+ log.info("X[" + i + "] = " + actual + " ("
+ actual.getClass().getName() + ")");
assertTrue("Not expecting X=" + actual, expected
@@ -364,6 +368,15 @@
final URI ENTITY = new URIImpl("http://bigdata.com/system#Entity");
+ final String query = "construct {"+//
+ "?s <" + RDF.TYPE + "> <" + ENTITY + "> ."+//
+ " } " + "where { "+//
+ " ?s <" + RDF.TYPE + "> <" + ENTITY + "> ."+//
+ " ?s ?p ?lit ."+//
+ " ?lit <" + BD.SEARCH + "> \"systap\" ."+//
+ " ?lit <" + BD.MIN_RELEVANCE + "> \"0.0\"^^<http://www.w3.org/2001/XMLSchema#double> ."+//
+ " }";
+
// the ontology (nothing is indexed for full text search).
final Graph test_restart_1 = new GraphImpl(); {
@@ -446,10 +459,6 @@
}
{ // run the query (free text search)
- final String query = "construct { ?s <" + RDF.TYPE + "> <"
- + ENTITY + "> . } " + "where { ?s <" + RDF.TYPE
- + "> <" + ENTITY + "> . ?s ?p ?lit . ?lit <"
- + BD.SEARCH + "> \"systap\" . }";
final RepositoryConnection cxn = repo.getConnection();
try {
// silly construct queries, can't guarantee distinct
@@ -459,13 +468,9 @@
QueryLanguage.SPARQL, query);
graphQuery.evaluate(new StatementCollector(results));
for (Statement stmt : results) {
- log.info(stmt);
+ if(log.isInfoEnabled())
+ log.info(stmt);
}
- /*
- * @todo this test is failing : review with MikeP and
- * figure out if it is the test or the system under
- * test.
- */
assertTrue(results.contains(new StatementImpl(SYSTAP,
RDF.TYPE, ENTITY)));
} finally {
@@ -489,10 +494,10 @@
repo.initialize();
{ // run the query again
- final String query = "construct { ?s <" + RDF.TYPE + "> <"
- + ENTITY + "> . } " + "where { ?s <" + RDF.TYPE
- + "> <" + ENTITY + "> . ?s ?p ?lit . ?lit <"
- + BD.SEARCH + "> \"systap\" . }";
+// final String query = "construct { ?s <" + RDF.TYPE + "> <"
+// + ENTITY + "> . } " + "where { ?s <" + RDF.TYPE
+// + "> <" + ENTITY + "> . ?s ?p ?lit . ?lit <"
+// + BD.SEARCH + "> \"systap\" . }";
final RepositoryConnection cxn = repo.getConnection();
try {
// silly construct queries, can't guarantee distinct
@@ -502,7 +507,8 @@
QueryLanguage.SPARQL, query);
graphQuery.evaluate(new StatementCollector(results));
for (Statement stmt : results) {
- log.info(stmt);
+ if(log.isInfoEnabled())
+ log.info(stmt);
}
assertTrue("Lost commit?", results
.contains(new StatementImpl(SYSTAP, RDF.TYPE,
@@ -782,9 +788,10 @@
int i = 0;
while (result.hasNext()) {
- System.err.println(i++ + ": " + result.next().toString());
+ if(log.isInfoEnabled())
+ log.info(i++ + ": " + result.next().toString());
}
- assertTrue("wrong # of results", i == 7);
+ assertEquals("wrong # of results", 7, i);
result = tupleQuery.evaluate();
@@ -795,12 +802,12 @@
final Hiterator<IHit> hits =
search.search(searchQuery,
null, // languageCode
- false, // prefixMatch
- 0d, // minCosine
- 1.0d, // maxCosine
- 10000, // maxRank (=maxResults + 1)
- false, // matchAllTerms
- 1000L, // timeout
+ BD.DEFAULT_PREFIX_MATCH,//false, // prefixMatch
+ BD.DEFAULT_MIN_RELEVANCE,//0d, // minCosine
+ BD.DEFAULT_MAX_RELEVANCE,//1.0d, // maxCosine
+ BD.DEFAULT_MAX_HITS,//10000, // maxRank (=maxResults + 1)
+ BD.DEFAULT_MATCH_ALL_TERMS,//false, // matchAllTerms
+ BD.DEFAULT_TIMEOUT,//1000L, // timeout
TimeUnit.MILLISECONDS // unit
);
@@ -814,7 +821,8 @@
new BindingImpl("s", s),
new BindingImpl("o", o),
new BindingImpl("score", score));
- System.err.println(bs);
+ if(log.isInfoEnabled())
+ log.info(bs);
answer.add(bs);
}
@@ -845,9 +853,10 @@
int i = 0;
while (result.hasNext()) {
- System.err.println(i++ + ": " + result.next().toString());
+ if(log.isInfoEnabled())
+ log.info(i++ + ": " + result.next().toString());
}
- assertTrue("wrong # of results", i == 5);
+ assertEquals("wrong # of results", 5, i);
result = tupleQuery.evaluate();
@@ -858,12 +867,12 @@
final Hiterator<IHit> hits =
search.search(searchQuery,
null, // languageCode
- false, // prefixMatch
- 0d, // minCosine
- 1.0d, // maxCosine
+ BD.DEFAULT_PREFIX_MATCH,//false, // prefixMatch
+ BD.DEFAULT_MIN_RELEVANCE,//0d, // minCosine
+ BD.DEFAULT_MAX_RELEVANCE,//1.0d, // maxCosine
maxHits+1, // maxRank (=maxResults + 1)
- false, // matchAllTerms
- 1000L, // timeout
+ BD.DEFAULT_MATCH_ALL_TERMS,//false, // matchAllTerms
+ BD.DEFAULT_TIMEOUT,//1000L, // timeout
TimeUnit.MILLISECONDS // unit
);
@@ -877,7 +886,8 @@
new BindingImpl("s", s),
new BindingImpl("o", o),
new BindingImpl("score", score));
- System.err.println(bs);
+ if(log.isInfoEnabled())
+ log.info(bs);
answer.add(bs);
}
@@ -910,9 +920,10 @@
int i = 0;
while (result.hasNext()) {
- System.err.println(i++ + ": " + result.next().toString());
+ if(log.isInfoEnabled())
+ log.info(i++ + ": " + result.next().toString());
}
- assertTrue("wrong # of results", i == 2);
+ assertEquals("wrong # of results", 2, i);
result = tupleQuery.evaluate();
@@ -923,12 +934,12 @@
final Hiterator<IHit> hits =
search.search(searchQuery,
null, // languageCode
- false, // prefixMatch
+ BD.DEFAULT_PREFIX_MATCH,//false, // prefixMatch
minRelevance, // minCosine
maxRelevance, // maxCosine
- 10000, // maxRank (=maxResults + 1)
- false, // matchAllTerms
- 1000L, // timeout
+ BD.DEFAULT_MAX_HITS,//10000, // maxRank (=maxResults + 1)
+ BD.DEFAULT_MATCH_ALL_TERMS,//false, // matchAllTerms
+ BD.DEFAULT_TIMEOUT,//1000L, // timeout
TimeUnit.MILLISECONDS // unit
);
@@ -942,7 +953,8 @@
new BindingImpl("s", s),
new BindingImpl("o", o),
new BindingImpl("score", score));
- System.err.println(bs);
+ if(log.isInfoEnabled())
+ log.info(bs);
answer.add(bs);
}
@@ -969,7 +981,8 @@
"} " +
"order by desc(?score)";
- log.info("\n"+query);
+ if(log.isInfoEnabled())
+ log.info("\n"+query);
final TupleQuery tupleQuery =
cxn.prepareTupleQuery(QueryLanguage.SPARQL, query);
@@ -978,9 +991,10 @@
int i = 0;
while (result.hasNext()) {
- log.info(i++ + ": " + result.next().toString());
+ if(log.isInfoEnabled())
+ log.info(i++ + ": " + result.next().toString());
}
- assertTrue("wrong # of results: " + i, i == 2);
+ assertEquals("wrong # of results: " + i, 2, i);
result = tupleQuery.evaluate();
@@ -991,12 +1005,12 @@
final Hiterator<IHit> hits =
search.search(searchQuery,
null, // languageCode
- false, // prefixMatch
+ BD.DEFAULT_PREFIX_MATCH,//false, // prefixMatch
minRelevance, // minCosine
maxRelevance, // maxCosine
- 10000, // maxRank (=maxResults + 1)
- false, // matchAllTerms
- 1000L, // timeout
+ BD.DEFAULT_MAX_HITS,//10000, // maxRank (=maxResults + 1)
+ BD.DEFAULT_MATCH_ALL_TERMS,//false, // matchAllTerms
+ BD.DEFAULT_TIMEOUT,//1000L, // timeout
TimeUnit.MILLISECONDS // unit
);
@@ -1039,7 +1053,8 @@
"} " +
"order by desc(?score)";
- log.info("\n"+query);
+ if(log.isInfoEnabled())
+ log.info("\n"+query);
final TupleQuery tupleQuery =
cxn.prepareTupleQuery(QueryLanguage.SPARQL, query);
@@ -1048,9 +1063,10 @@
int i = 0;
while (result.hasNext()) {
- log.info(i++ + ": " + result.next().toString());
+ if(log.isInfoEnabled())
+ log.info(i++ + ": " + result.next().toString());
}
- assertTrue("wrong # of results: " + i, i == 3);
+ assertEquals("wrong # of results: " + i, 3, i);
result = tupleQuery.evaluate();
@@ -1064,9 +1080,9 @@
true, // prefixMatch
minRelevance, // minCosine
maxRelevance, // maxCosine
- 10000, // maxRank (=maxResults + 1)
- false, // matchAllTerms
- 1000L, // timeout
+ BD.DEFAULT_MAX_HITS,//10000, // maxRank (=maxResults + 1)
+ BD.DEFAULT_MATCH_ALL_TERMS,//false, // matchAllTerms
+ BD.DEFAULT_TIMEOUT,//1000L, // timeout
TimeUnit.MILLISECONDS // unit
);
@@ -1092,7 +1108,7 @@
final String searchQuery = "to*";
final double minRelevance = 0.0d;
- final double maxRelevance = 1.0d;
+ final double maxRelevance = BD.DEFAULT_MAX_RELEVANCE;//1.0d;
final String query =
"select ?s ?o ?score " +
@@ -1101,13 +1117,14 @@
" ?s <"+RDFS.LABEL+"> ?o . " +
" ?o <"+BD.SEARCH+"> \""+searchQuery+"\" . " +
" ?o <"+BD.RELEVANCE+"> ?score . " +
-// " ?o <"+BD.MIN_RELEVANCE+"> \""+minRelevance+"\" . " +
+ " ?o <"+BD.MIN_RELEVANCE+"> \""+minRelevance+"\" . " +
// " ?o <"+BD.MAX_HITS+"> \"5\" . " +
// " filter regex(?o, \""+searchQuery+"\") " +
"} " +
"order by desc(?score)";
- log.info("\n"+query);
+ if(log.isInfoEnabled())
+ log.info("\n"+query);
final TupleQuery tupleQuery =
cxn.prepareTupleQuery(QueryLanguage.SPARQL, query);
@@ -1116,9 +1133,10 @@
int i = 0;
while (result.hasNext()) {
- log.info(i++ + ": " + result.next().toString());
+ if(log.isInfoEnabled())
+ log.info(i++ + ": " + result.next().toString());
}
- assertTrue("wrong # of results: " + i, i == 1);
+ assertEquals("wrong # of results: " + i, 1, i);
result = tupleQuery.evaluate();
@@ -1132,9 +1150,9 @@
true, // prefixMatch
minRelevance, // minCosine
maxRelevance, // maxCosine
- 10000, // maxRank (=maxResults + 1)
- false, // matchAllTerms
- 1000L, // timeout
+ BD.DEFAULT_MAX_HITS,//10000, // maxRank (=maxResults + 1)
+ BD.DEFAULT_MATCH_ALL_TERMS,//false, // matchAllTerms
+ BD.DEFAULT_TIMEOUT,//1000L, // timeout
TimeUnit.MILLISECONDS // unit
);
@@ -1180,7 +1198,8 @@
int i = 0;
while (result.hasNext()) {
- log.info(i++ + ": " + result.next().toString());
+ if(log.isInfoEnabled())
+ log.info(i++ + ": " + result.next().toString());
}
// assertTrue("wrong # of results: " + i, i == 1);
@@ -1196,9 +1215,9 @@
true, // prefixMatch
minRelevance, // minCosine
maxRelevance, // maxCosine
- 10000, // maxRank (=maxResults + 1)
+ BD.DEFAULT_MAX_HITS,//10000, // maxRank (=maxResults + 1)
true, // matchAllTerms
- 1000L, // timeout
+ BD.DEFAULT_TIMEOUT,//1000L, // timeout
TimeUnit.MILLISECONDS // unit
);
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
