Revision: 14466
http://gate.svn.sourceforge.net/gate/?rev=14466&view=rev
Author: valyt
Date: 2011-10-31 17:22:03 +0000 (Mon, 31 Oct 2011)
Log Message:
-----------
- test data: moved the document metadata to document features;
- default testing index config now uses the new document metadata functioanlity;
Document features seem to work when using the H2 plugin. Next: do the same for
the Sesame one.
Modified Paths:
--------------
mimir/trunk/mimir-test/data/gatexml-output.zip
mimir/trunk/mimir-test/src/gate/mimir/test/QueryTests.java
mimir/trunk/mimir-test/src/gate/mimir/test/TestUtils.java
Added Paths:
-----------
mimir/trunk/mimir-test/src/gate/mimir/test/Scratch.java
Modified: mimir/trunk/mimir-test/data/gatexml-output.zip
===================================================================
(Binary files differ)
Modified: mimir/trunk/mimir-test/src/gate/mimir/test/QueryTests.java
===================================================================
--- mimir/trunk/mimir-test/src/gate/mimir/test/QueryTests.java 2011-10-31
17:19:36 UTC (rev 14465)
+++ mimir/trunk/mimir-test/src/gate/mimir/test/QueryTests.java 2011-10-31
17:22:03 UTC (rev 14466)
@@ -23,6 +23,7 @@
import static org.junit.Assert.fail;
import gate.Document;
import gate.Gate;
+import gate.mimir.AbstractSemanticAnnotationHelper;
import gate.mimir.IndexConfig;
import gate.mimir.SemanticAnnotationHelper;
import gate.mimir.index.IndexException;
@@ -79,6 +80,7 @@
@BeforeClass
public static void oneTimeSetUp() throws Exception {
Gate.setGateHome(new File("gate-home"));
+ Gate.setUserConfigFile(new File("gate-home/user-gate.xml"));
Gate.init();
// load the tokeniser plugin
Gate.getCreoleRegister().registerDirectories(new
File("gate-home/plugins/ANNIE-tokeniser").toURI().toURL());
@@ -93,7 +95,7 @@
// would use: "gate.mimir.ordi.ORDISemanticAnnotationHelper")
IndexConfig indexConfig = TestUtils.getTestIndexConfig(indexDir,
Class.forName("gate.mimir.db.DBSemanticAnnotationHelper", true,
-
Gate.getClassLoader()).asSubclass(SemanticAnnotationHelper.class));
+
Gate.getClassLoader()).asSubclass(AbstractSemanticAnnotationHelper.class));
// now start indexing the documents
Indexer indexer = new Indexer(indexConfig);
String pathToZipFile = "data/gatexml-output.zip";
@@ -125,9 +127,9 @@
engine.close();
engine = null;
// recursively delete index dir
- if(!TestUtils.deleteDir(indexDir)) {
- System.err.println("Could not delete index directory " + indexDir);
- }
+// if(!TestUtils.deleteDir(indexDir)) {
+// System.err.println("Could not delete index directory " + indexDir);
+// }
}
}
Added: mimir/trunk/mimir-test/src/gate/mimir/test/Scratch.java
===================================================================
--- mimir/trunk/mimir-test/src/gate/mimir/test/Scratch.java
(rev 0)
+++ mimir/trunk/mimir-test/src/gate/mimir/test/Scratch.java 2011-10-31
17:22:03 UTC (rev 14466)
@@ -0,0 +1,64 @@
+package gate.mimir.test;
+
+import it.unimi.dsi.mg4j.search.score.BM25FScorer;
+import it.unimi.dsi.mg4j.search.score.BM25Scorer;
+import it.unimi.dsi.mg4j.search.score.CountScorer;
+import it.unimi.dsi.mg4j.search.score.DelegatingScorer;
+import it.unimi.dsi.mg4j.search.score.Scorer;
+import it.unimi.dsi.mg4j.search.score.TfIdfScorer;
+
+import java.io.File;
+import java.text.NumberFormat;
+
+import gate.Gate;
+import gate.mimir.search.QueryEngine;
+import gate.mimir.search.query.QueryExecutor;
+import gate.mimir.search.query.QueryNode;
+import gate.mimir.search.query.parser.QueryParser;
+
+public class Scratch {
+
+ public static void main(String[] args) throws Exception {
+ Gate.setGateHome(new File("gate-home"));
+ Gate.setUserConfigFile(new File("gate-home/user-gate.xml"));
+ Gate.init();
+ // load the tokeniser plugin
+ Gate.getCreoleRegister().registerDirectories(
+ new File("gate-home/plugins/ANNIE-tokeniser").toURI().toURL());
+ // load the DB plugin
+ Gate.getCreoleRegister().registerDirectories(
+ new File("../plugins/db-h2").toURI().toURL());
+ // load the measurements plugin
+ Gate.getCreoleRegister().registerDirectories(
+ new File("../plugins/measurements").toURI().toURL());
+ Gate.getCreoleRegister().registerDirectories(
+ new File("../plugins/sparql").toURI().toURL());
+
+ QueryEngine qEngine = new QueryEngine(new File(args[0]));
+// String query = "the";
+ String query = "{Document}";
+ QueryNode qNode = QueryParser.parse(query);
+ long start = System.currentTimeMillis();
+ NumberFormat nf = NumberFormat.getNumberInstance();
+ long startLocal = System.currentTimeMillis();
+ QueryExecutor qExecutor = qNode.getQueryExecutor(qEngine);
+ int latestDoc = qExecutor.nextDocument(-1);
+ int totalHitCount = 0;
+ int docCount = 0;
+ while(latestDoc >= 0) {
+ docCount++;
+ int hitCount = 0;
+ while(qExecutor.nextHit() != null) hitCount++;
+ totalHitCount += hitCount;
+ System.out.println("Doc " + latestDoc + ", hits: " + hitCount);
+ latestDoc = qExecutor.nextDocument(-1);
+ }
+ System.out.println("Found " + nf.format(totalHitCount) + " hits in " +
+ nf.format(docCount) + " documents, in " +
+ nf.format(System.currentTimeMillis() - startLocal) + " ms.");
+ qExecutor.close();
+ System.out.println("Total time " +
+ nf.format(System.currentTimeMillis() - start) + " ms.");
+ qEngine.close();
+ }
+}
Property changes on: mimir/trunk/mimir-test/src/gate/mimir/test/Scratch.java
___________________________________________________________________
Added: svn:keywords
+ Id
Added: svn:eol-style
+ native
Modified: mimir/trunk/mimir-test/src/gate/mimir/test/TestUtils.java
===================================================================
--- mimir/trunk/mimir-test/src/gate/mimir/test/TestUtils.java 2011-10-31
17:19:36 UTC (rev 14465)
+++ mimir/trunk/mimir-test/src/gate/mimir/test/TestUtils.java 2011-10-31
17:22:03 UTC (rev 14466)
@@ -20,6 +20,7 @@
import gate.Gate;
import gate.creole.ANNIEConstants;
import gate.mimir.index.*;
+import gate.mimir.AbstractSemanticAnnotationHelper;
import gate.mimir.DocumentMetadataHelper;
import gate.mimir.IndexConfig;
import gate.mimir.SemanticAnnotationHelper;
@@ -44,11 +45,11 @@
public class TestUtils {
public static IndexConfig getTestIndexConfig(File indexDir,
- Class<? extends SemanticAnnotationHelper> helperClass)
+ Class<? extends AbstractSemanticAnnotationHelper>
helperClass)
throws IllegalArgumentException, InstantiationException,
IllegalAccessException, InvocationTargetException, SecurityException,
NoSuchMethodException, ClassNotFoundException {
- Constructor<? extends SemanticAnnotationHelper> helperMaker =
+ Constructor<? extends AbstractSemanticAnnotationHelper> helperMaker =
helperClass.getConstructor(String.class, String[].class,
String[].class,
String[].class, String[].class, String[].class);
Class<? extends SemanticAnnotationHelper> measurementsHelperClass =
@@ -61,8 +62,7 @@
new HashSet<String>(Arrays.asList(
new String[] {
"b", "i", "li", "ol", "p", "sup", "sub", "u", "ul"})));
- // index configuration based on the original SAM one but without the
- // sam-specific classes.
+ // index configuration used for testing.
return new IndexConfig(
indexDir,
"mimir",
@@ -101,14 +101,14 @@
new SemanticIndexerConfig(
new String[]{"Abstract", "Assignee",
"ClassificationIPCR", "InventionTitle",
- "Inventor", "PatentDocument",
"PriorityClaim"},
+ "Inventor", "Document", "PriorityClaim"},
new SemanticAnnotationHelper[] {
helperMaker.newInstance("Abstract", new
String[]{"lang"}, null, null, null, null),
helperMaker.newInstance("Assignee", null,
null, null, null, null),
helperMaker.newInstance("ClassificationIPCR",
new String[]{"status"}, null, null, null, null),
helperMaker.newInstance("InventionTitle", new
String[]{"lang", "status"}, null, null, null, null),
helperMaker.newInstance("Inventor", new
String[]{"format", "status"}, null, null, null, null),
- helperMaker.newInstance("PatentDocument",
null, new String[]{"date"}, null, new String[]{"ucid"}, null),
+ helperMaker.newInstance("Document", null, new
String[]{"date"}, null, new String[]{"ucid"}, null).asDocumentHelper(),
helperMaker.newInstance("PriorityClaim", null,
null, null, new String[]{"ucid"}, null)})
},
new DocumentMetadataHelper[] {docHelper},
This was sent by the SourceForge.net collaborative development platform, the
world's largest Open Source development site.
------------------------------------------------------------------------------
Get your Android app more play: Bring it to the BlackBerry PlayBook
in minutes. BlackBerry App World™ now supports Android™ Apps
for the BlackBerry® PlayBook™. Discover just how easy and simple
it is! http://p.sf.net/sfu/android-dev2dev
_______________________________________________
GATE-cvs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/gate-cvs
