git-svn-id: https://svn.d4science.research-infrastructures.eu/gcube/trunk/data-analysis/EcologicalEngine@71257 82a268e6-3cf1-43bd-a215-b396298e98cf

src/main/java/org/gcube/contentmanagement/lexicalmatcher/analysis/run/CategoryGuesser.java

@@ -212,6 +212,11 @@ public class CategoryGuesser {
 		col = null;
 	}
 
+	public ArrayList<SingleResult> getLastResults(){
+		return lastResults;
+	}
+	
+	ArrayList<SingleResult> lastResults;
 	public void runGuesser(String seriesName, String columnName, LexicalEngineConfiguration externalConfig, String CategoryFilter, String ColumnFilter, String SingletonString) throws Exception {
 		SessionFactory dbSession = null;
 		String cfgFileCompletePath = cfgFile;
@@ -304,6 +309,7 @@ public class CategoryGuesser {
 		AnalysisLogger.getLogger().warn("...End Calculation in " + t1 + "ms");
 
 		triesCounter = 0;
+		lastResults=checkingResults;
 		// close session if not more necessary
 		if (oneshotMode)
 			dbSession.close();
@@ -478,7 +484,8 @@ public class CategoryGuesser {
 			}
 		}
 		
-		return deleteDuplicates(results);
+//		return deleteDuplicates(results);
+		return results;
 	}
 
 	

src/main/java/org/gcube/contentmanagement/lexicalmatcher/analysis/test/ExampleGuesser.java

@@ -37,11 +37,12 @@ public class ExampleGuesser {
 			conf.setDatabaseDialect("org.hibernate.dialect.PostgreSQLDialect");
 			
 			guesser.runGuesser(seriesName, column, conf);
-			ArrayList<SingleResult> classification = guesser.getClassification();
+			ArrayList<SingleResult> classification = guesser.getLastResults();
 			
 			//only for debug
 			guesser.showResults(classification);
-
+			
+			guesser.shutdown();
 		} catch (Exception e) {
 			e.printStackTrace();
 		}

src/main/java/org/gcube/contentmanagement/lexicalmatcher/analysis/test/ExampleLexicalMatcher.java

@@ -0,0 +1,58 @@
+package org.gcube.contentmanagement.lexicalmatcher.analysis.test;
+
+import java.util.ArrayList;
+
+import org.gcube.contentmanagement.lexicalmatcher.analysis.core.LexicalEngineConfiguration;
+import org.gcube.contentmanagement.lexicalmatcher.analysis.guesser.data.Category;
+import org.gcube.contentmanagement.lexicalmatcher.analysis.guesser.data.SingleResult;
+import org.gcube.contentmanagement.lexicalmatcher.analysis.run.CategoryGuesser;
+import org.gcube.contentmanagement.lexicalmatcher.utils.AnalysisLogger;
+
+public class ExampleLexicalMatcher {
+
+	public static void main(String[] args) {
+
+		try {
+
+			String singleton = "Faroe Island";
+			String family = "COUNTRY_OLD";
+			String column = "field6";
+			
+			CategoryGuesser guesser = new CategoryGuesser();
+			LexicalEngineConfiguration conf = new LexicalEngineConfiguration();
+			
+			ArrayList<Category> categories = new ArrayList<Category>();
+			//human name, index, table name, description
+			categories.add(new Category("COUNTRY_OLD","39c98800-dd3c-11e0-b8d1-d1e2e7ba4f9d","rdf39c98800dd3c11e0b8d1d1e2e7ba4f9d","country"));
+			categories.add(new Category("CONTINENT_OLD","1d5d51f0-dd42-11e0-b8d3-d1e2e7ba4f9d","rdf1d5d51f0dd4211e0b8d3d1e2e7ba4f9d","continent reference data"));
+			categories.add(new Category("SPECIES_OLD","0a7fb500-dd3d-11e0-b8d1-d1e2e7ba4f9d","rdf0a7fb500dd3d11e0b8d1d1e2e7ba4f9d","species"));
+			categories.add(new Category("CodeListCountry","4c8d93a0-edc2-11e0-93e4-f6a9821baa29","rdf4c8d93a0edc211e093e4f6a9821baa29","Country"));
+			categories.add(new Category("CL_DIVISION","1140bdf0-dd2c-11e0-9220-ae17b3db32b7","rdf1140bdf0dd2c11e09220ae17b3db32b7","undefined"));
+			categories.add(new Category("CL_ASFIS_TAX","f87360f0-d9f9-11e0-ba05-d9adb0db767c","rdff87360f0d9f911e0ba05d9adb0db767c","undefined"));
+			conf.setCategories(categories);
+			
+			//CHANGE THIS TO ENHANCE THE RECALL
+			conf.setEntryAcceptanceThreshold(30);
+			conf.setReferenceChunksToTake(-1);
+			conf.setTimeSeriesChunksToTake(-1);
+			conf.setUseSimpleDistance(false);
+			
+			//database Parameters
+			conf.setDatabaseUserName("gcube");
+			conf.setDatabasePassword("d4science2");
+			conf.setDatabaseURL("jdbc:postgresql://localhost/testdb");
+			conf.setDatabaseDialect("org.hibernate.dialect.PostgreSQLDialect");
+			
+			guesser.runGuesser(singleton, conf, family,column );
+			ArrayList<SingleResult> detailedResults = guesser.getDetailedMatches();
+			AnalysisLogger.getLogger().warn("Detailed Match on Name :"+singleton);
+			
+			CategoryGuesser.showResults(detailedResults);
+			
+		} catch (Exception e) {
+			e.printStackTrace();
+		}
+
+	}
+
+}

src/main/resources/lexicalguesser/ALog.properties

@@ -8,9 +8,9 @@ log4j.appender.stdout.layout.ConversionPattern=%d{dd/MM/yyyy HH:mm:ss} %p %t %c
 
 
 #### Second appender writes to a file
-log4j.logger.AnalysisLogger=info,AR
+log4j.logger.AnalysisLogger=trace,AR,stdout
 log4j.appender.AR=org.apache.log4j.RollingFileAppender
-log4j.appender.AR.Threshold=info
+log4j.appender.AR.Threshold=trace
 log4j.appender.AR.File=${GLOBUS_LOCATION}/logs/Analysis.log
 log4j.appender.AR.MaxFileSize=1024KB
 log4j.appender.AR.MaxBackupIndex=2

src/main/resources/lexicalguesser/lexicalGuesser.properties

@@ -1,13 +1,13 @@
 #Percentage threshold for discarding a category
 categoryDiscardThreshold=0
 #Percentage threshold for accepting similarity between a single Time series entry and a reference entry
-entryAcceptanceThreshold=80
+entryAcceptanceThreshold=50
 #Size of a comparison chunk
-chunkSize=25
+chunkSize=50
 #Number of chunks to take from Time series for performing comparison respect to reference data; if set to -1 all chunks will be analyzed
-timeSeriesChunksToTake=2
+timeSeriesChunksToTake=1
 #Number of chunks to take from Reference for performing comparison Time Series Elements; if set to -1 all chunks will be analyzed 
-referenceChunksToTake =50
+referenceChunksToTake=5
 #Use random choice for chunks selection = true |false
 randomTake=true
 #Use Simple String Match as distance calculation
@@ -15,6 +15,6 @@ useSimpleDistance=false
 #Number Of Threads  to use
 numberOfThreadsToUse=5
 #if two final scores differ for more than this percentage, prune the lower result
-categoryDiscardDifferencialThreshold = 2
+categoryDiscardDifferencialThreshold = 50
 #maximum difference between a result and the best result
 singleEntryRecognitionMaxDeviation = 40