readme file in directory SQLFeatureExtraction is updated

SQLFeatureExtraction/README.md

@@ -3,3 +3,31 @@ In order to produce the executable file *SQLComparison.jar* from Java source cod
 	mvn clean install
 
 This command will install the maven project and create a executable jar file named *SQLComparison.jar* in the directory *SQLFeatureExtraction/target*
+
+By running *SQLComparison.jar* with no options, user can reproduce feature vectors generated by:
+
+(1) Each of the three similarity metrics (aligon, makiyama and aouiche); 
+
+(2) Under three different SQL log-like inputs (UB student exam, Bombay IIT and pocketdata-google+);
+
+(3) With no further feature engineering;
+
+(4) With each feature engineering module activated individually;
+
+(5) With all feature engineering modules activated.
+
+Options:
+
+(1) User can specify a single input data set by using -input option. The available option parameters are: (1)ub; (2)bombay; (3)googleplus. (e.g. Java -jar SQLComparison.jar -input ub)
+
+(2) User can specify a single similarity metric by using -metric option. The available option parameters are: (1)aligon; (2)makiyama; (3)aouiche. (e.g. Java -jar SQLComparison.jar -metric aligon)
+
+(3) User can specify regularization modules to be activated by using -modules option. The available modules are: ID=1, Naming; ID=2, Expression Regularization; ID=3, From-Nesting Flattening; ID=4, UNION pull-out. User should specify corresponding IDs and use symbol "&" for combining multiple selected modules, if any. (e.g. Java -jar SQLComparison.jar -modules 1&2&3&4)
+
+Result of running *SQLComparison.jar* will be saved to directory *SQLFeatureExtraction/data* and the result file naming format is: [input]+"_"+[metric]+ ".csv". If regularization is applied, naming format is: [input]+"_"+[metric]+"_regularization_"+[module IDs]+".csv"; If all modules of regularization are applied, naming format is simplified as: [input]+"_"+[metric]+"_regularization.csv".
+
+
+
+
+
+

SQLFeatureExtraction/src/main/java/Main.java

@@ -25,7 +25,7 @@ public static void main(String[] args) {
 			File f=new File(current);
 			String parent=f.getParent();
 			datapath=parent+"/data/";
-			System.out.println(datapath);
+			System.out.println("data path : "+datapath);
 		} catch (IOException e) {
 			e.printStackTrace();
 		}
@@ -79,7 +79,7 @@ else if (name.contains("makiyama"))
 		ArrayList<ArrayList<String> > queryLists=new ArrayList<ArrayList<String>> ();
 		System.out.println("---begin query retrieval------");
 		System.out.println();
-		long start=System.nanoTime();
+		//long start=System.nanoTime();
 
 		for(int i=0;i<data.length;i++){
 			System.out.println("data set using is "+data[i]);
@@ -95,7 +95,7 @@ else if (name.contains("makiyama"))
 			queryLists.add(queryList);	
 			System.out.println();
 		}
-		long end=System.nanoTime();
+		//long end=System.nanoTime();
 		System.out.println("query parsing finished.");
 		System.out.println();
 
@@ -150,26 +150,26 @@ public static void queryComparison(String data,String method,ArrayList<String> q
 		//do a pass of regularization	
 		if(regu){
 			System.out.println("begin regularization");		
-			long start=System.nanoTime();
+			//long start=System.nanoTime();
 			for(int k=0;k<statementList1.size();k++){		
 				Select s=(Select) statementList1.get(k);
 				statementList.add(CombinedRegularizer.regularize(s,modules));
 			}
-			long end=System.nanoTime();
+			//long end=System.nanoTime();
 			System.out.println("regularization ended.");
 			System.out.println();
 		}
 
 		System.out.println("begin query comparison and clustering. ");
-		long start=System.nanoTime();
+		//long start=System.nanoTime();
 		// method name can be either aouiche, makiyama or aligon
 		double[][] matrix;
 		if(regu)
 			matrix = Utility.createDistanceMatrix(method, statementList);
 		else
 			matrix = Utility.createDistanceMatrix(method, statementList1);	
 
-		long end=System.nanoTime();
+		//long end=System.nanoTime();
 		// write distance matrix to file
 		String outputpath;
 		if(!regu)