Add readR and readWeka methods

Author: AlkaSaliss

main.java

@@ -13,7 +13,7 @@
 
 public class main {
 
-	public static void main(String[] args) {
+	public static void main2(String[] args) {
 		// TODO Auto-generated method stub
 
 

methods/methods/R/methods.rdb

@@ -39,11 +39,15 @@ public RDecisionTree(String form, String method, int coly) {
 
 	}
 
-	public void fit(String dataPath) throws ScriptException {
+	public void fit(String dataPath, boolean hasRowname) throws ScriptException {
 		/*
 		 * loading the file
 		 */
+		if (hasRowname) {
 		engine.eval("data <- as.data.frame(read.csv("+dataPath+", header=T, row.names=1))");
+		} else {
+			engine.eval("data <- as.data.frame(read.csv("+dataPath+", header=T, row.names=NULL))");
+		}
 		/*
 		 * Spliting data into training and test sets
 		 */

methods/methods/R/methods.rdx

@@ -14,6 +14,7 @@
 import weka.core.converters.CSVLoader;
 
 import weka.filters.Filter;
+import weka.filters.unsupervised.attribute.NumericToNominal;
 import weka.filters.unsupervised.attribute.Remove;
 
 
@@ -29,7 +30,7 @@ public class WekaDecisionTree {
 
 
 
-	public WekaDecisionTree(String path, boolean hasRowName) throws Exception {
+	public WekaDecisionTree(String path, boolean hasRowName, boolean toNumeric) throws Exception {
 		this.datapath = path;
 
 		//BufferedReader reader = new BufferedReader(new FileReader(path));
@@ -55,6 +56,29 @@ public WekaDecisionTree(String path, boolean hasRowName) throws Exception {
 		}
 
 
+		
+		
+		/*
+		 * Convert numeric target variable to categories
+		 * */
+		if (toNumeric) {
+		
+		NumericToNominal convert= new NumericToNominal();
+        String[] options= new String[2];
+        options[0]="-R"; //remove option 
+        options[1]= String.valueOf(this.data.numAttributes() ) ;
+        
+        convert.setOptions(options);
+        convert.setInputFormat(this.data);
+        this.data =Filter.useFilter(this.data, convert);
+        
+		}
+		
+		
+	}
+	
+	public int getAttributeIndex(String label) {
+		return this.data.attribute(label).index();
 	}
 
 	public String getDatapath() {

src/main/java/fr/ensai/renjin_ml/RDecisionTree.java

@@ -17,28 +17,10 @@ public test() {
 	public static void main(String[] args) throws Exception {
 
 		//create the tree object
-		RDecisionTree tree = new RDecisionTree("Species~.", "class", 5);
-		tree.fit("\"src/main/resources/iris.csv\"");
+		RDecisionTree tree = new RDecisionTree("quality~.", "class", 12);
+		tree.fit("\"src/main/resources/winequality.csv\"", false);
 		tree.predict();
-		
-		
-		// create a script engine manager:
-	   /* RenjinScriptEngineFactory factory = new RenjinScriptEngineFactory();
-	    // create a Renjin engine:
-	    RenjinScriptEngine engine = factory.getScriptEngine();
-
-	    engine.eval("dat <- data.frame(x=1:10, y=(1:10)+rnorm(n=10))");
-	    engine.eval("print(typeof(dat))");*/
-	    
-	 /*   String chemin = "\"src/main/resources/iris.csv\"";
-	    engine.eval("data <- read.csv(" + chemin + ", header=T, row.names=1)");*/
-	    //engine.eval("print(data)");
-	  //  engine.eval("print(lm(y ~ x, df))");
-	   /* engine.eval("rg <- lm(y ~ x, dat)");
-	    
-	    //engine.eval("install.packages(\"rpart\")");
-		engine.eval("library(rpart)");
-	    
+			    
 	   // ListVector vect = (ListVector) engine.get("rg");
 	  // System.out.println(engine.get("rg$residuals"));
 	  // List<String> test = List
@@ -47,29 +29,24 @@ public static void main(String[] args) throws Exception {
 
 	   //System.out.println( df.class);
 
-		System.out.println( engine.eval("packageVersion(\"rpart\")"));*/
-		
-		/*engine.eval("n=2*floor(nrow(data)/3)");
-		engine.eval("train=data[1:n,]");
-		engine.eval("library(rpart)"); //load the package
-		engine.eval("fit <- rpart(Species ~ ., data=train)");*/
-		
+				
 		System.out.println("*************************************************\n");
 
-		WekaDecisionTree treeWeka = new WekaDecisionTree("src/main/resources/iris.csv", true);
+		WekaDecisionTree treeWeka = new WekaDecisionTree("src/main/resources/winequality.csv", false, true);
 		J48 model = treeWeka.fit((float)0.7);
 
-		treeWeka.predict(model);
+		//treeWeka.predict(model);
 
-		System.out.println("*************************************************\n");
-		//System.out.println(treeWeka.getTest());
+		System.out.println("**********************Position***************************\n");
+		//System.out.println(treeWeka.getData());
+		//System.out.println(treeWeka.getAttributeIndex("chlorides"));
 
 		System.out.println("*************************************************\n");
 		//System.out.println(tree.getEngine().eval("print(train)"));
 		//double a = ((DoubleArrayVector) tree.getEngine().eval("accuracy")).asReal();
-		ListVector a =  (ListVector) tree.getEngine().eval("pred");
+		//ListVector a =  (ListVector) tree.getEngine().eval("pred");
 
-		System.out.println(a.asReal());
+		//System.out.println(tree.getEngine().eval("print(pred)"));
 
 
 	  }

src/main/java/fr/ensai/renjin_ml/WekaDecisionTree.java

@@ -1,9 +1,84 @@
 package metier;
 
+import java.io.File;
+import java.io.IOException;
+
+import javax.script.ScriptException;
+
+import org.renjin.script.RenjinScriptEngine;
+
+import weka.core.Instances;
+import weka.core.converters.CSVLoader;
+import weka.filters.Filter;
+import weka.filters.unsupervised.attribute.NumericToNominal;
+import weka.filters.unsupervised.attribute.Remove;
+
 public class Data {
 	String path;
-	String header;
+	String header="true";
 	String targetname;
+	String hasRowNames="false"; //R
+	String sep = ","; //R
+	String dec = "."; //R
+	
+	String catTarget = "false"; //weka
+	String toNominal = "false"; //
+	
+	
+	
+	
+	public void readR(RenjinScriptEngine engine) throws ScriptException {
+		
+		String rownames = this.hasRowNames.toUpperCase()=="true" ? "1" : "NULL";
+		
+		engine.eval("\"data <- read.csv(\"+path+\", header="+this.header.toUpperCase() +"," + sep + ", row.names="+rownames+")\"");
+		
+	}
+	
+	public Instances readWeka() throws Exception {
+		
+		File f = new File(this.path);
+		CSVLoader cnv = new CSVLoader();
+		cnv.setSource(f);
+		Instances data = cnv.getDataSet();
+		
+		
+		/*
+		 * Retrieve the target column index and set this column as model dependant variable*/
+		int targetIndex = data.attribute(this.targetname).index(); // target variable index
+		data.setClassIndex(targetIndex);
+		
+		if (this.hasRowNames.toUpperCase() == "TRUE") {
+			String[] options = new String[2];
+			 options[0] = "-R";                                    // "range"
+			 options[1] = "1";                                     // first attribute
+			 Remove remove = new Remove();                         // new instance of filter
+			 remove.setOptions(options);                           // set options
+			 remove.setInputFormat(data);                          // inform filter about dataset **AFTER** setting options
+			 data = Filter.useFilter(data, remove);   // apply filter
+		}
+		
+		
+		
+		
+		/*
+		 * Convert numeric target variable to categories
+		 * */
+		if (toNominal.toUpperCase()=="TRUE") {
+		
+			NumericToNominal convert= new NumericToNominal();
+	        String[] options= new String[2];
+	        options[0]="-R";
+	        options[1]= String.valueOf(targetIndex) ;
+	        
+	        convert.setOptions(options);
+	        convert.setInputFormat(data);
+	        data =Filter.useFilter(data, convert);
+		}
+		
+		return data;
+		
+	}
 
 
 }