Add new unit test PCATest using smile machine learning library

src/test/java/org/mastodon/mamut/feature/dimensionalityreduction/pca/PCATest.java

@@ -0,0 +1,43 @@
+package org.mastodon.mamut.feature.dimensionalityreduction.pca;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+import static org.junit.jupiter.api.Assertions.assertTrue;
+
+import java.lang.invoke.MethodHandles;
+
+import org.junit.jupiter.api.Test;
+import org.mastodon.mamut.feature.dimensionalityreduction.RandomDataTools;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import smile.data.DataFrame;
+import smile.feature.extraction.PCA;
+
+class PCATest
+{
+	private static final Logger logger = LoggerFactory.getLogger( MethodHandles.lookup().lookupClass() );
+
+	@Test
+	void test()
+	{
+		int numCluster1 = 50;
+		int numCluster2 = 100;
+		double[][] inputData = RandomDataTools.generateSampleData( numCluster1, numCluster2 );
+		logger.debug( "dimensions rows: {}, columns:{}", inputData.length, inputData[ 0 ].length );
+
+		int targetDimensions = 2;
+
+		DataFrame dataFrame = DataFrame.of( inputData );
+		PCA pca = PCA.fit( dataFrame ).getProjection( targetDimensions );
+		double[][] pcaResult = pca.apply( inputData );
+
+		assertEquals( pcaResult.length, inputData.length );
+		assertEquals( targetDimensions, pcaResult[ 0 ].length );
+
+		for ( int i = 0; i < numCluster1; i++ )
+			assertTrue( pcaResult[ i ][ 0 ] < 0 );
+		for ( int i = numCluster1; i < numCluster1 + numCluster2; i++ )
+			assertTrue( pcaResult[ i ][ 0 ] > 0 );
+
+	}
+}