Adds FeatureExtractor and corresponding test class;

sphinx4-core/src/main/java/edu/cmu/sphinx/api/FeatureExtractor.java

@@ -0,0 +1,136 @@
+package edu.cmu.sphinx.api;
+
+import java.io.IOException;
+import java.io.InputStream;
+import java.util.ArrayList;
+import java.util.List;
+
+import edu.cmu.sphinx.frontend.Data;
+import edu.cmu.sphinx.frontend.DataEndSignal;
+import edu.cmu.sphinx.frontend.DataProcessingException;
+import edu.cmu.sphinx.frontend.DoubleData;
+import edu.cmu.sphinx.frontend.FloatData;
+import edu.cmu.sphinx.frontend.FrontEnd;
+
+/**
+ * Extracts features from input stream
+ * 
+ * @author Vladisav Jelisavcic
+ * 
+ */
+class FeatureExtractor {
+
+	private FrontEnd frontEnd;
+	private int featureLength = -1;
+
+	/**
+	 * Constructs a FeatureExtractor.
+	 * 
+	 * @param cm
+	 *            the configuration manager
+	 * @param frontEndName
+	 *            the name for the frontEnd to be used
+	 * @param inputStream
+	 *            data stream
+	 * 
+	 * @throws IOException
+	 *             if error occurred
+	 */
+	public FeatureExtractor(InputStream inputStream, int sampleRate)
+			throws IOException {
+		Configuration configuration = new Configuration();
+
+		configuration.setAcousticModelPath("resource:/edu/cmu/sphinx/models/en-us/en-us");
+		configuration.setDictionaryPath("resource:/edu/cmu/sphinx/models/en-us/cmudict-en-us.dict");
+		configuration.setLanguageModelPath("resource:/edu/cmu/sphinx/models/en-us/en-us.lm.bin");
+
+		Context ctx = new Context(configuration);
+		ctx.setSampleRate(sampleRate);
+		ctx.setSpeechSource(inputStream);
+
+		frontEnd = (FrontEnd) ctx.getInstance(FrontEnd.class);
+	}
+
+	/**
+	 * Extracts all features from the supplied InputStream.
+	 * 
+	 * @return float[][] when called first time, null otherwise
+	 * 
+	 * @throws DataProcessinException
+	 *             if error occurred
+	 */
+	public List<float[]> getAllFeatures() throws DataProcessingException {
+		List<float[]> featureList = new ArrayList<float[]>();
+
+		Data feature = frontEnd.getData();
+		if (feature == null)
+			return null;
+
+		while (!(feature instanceof DataEndSignal)) {
+			if (feature instanceof DoubleData) {
+				double[] featureData = ((DoubleData) feature).getValues();
+				if (featureLength < 0) {
+					featureLength = featureData.length;
+				}
+				float[] convertedData = new float[featureData.length];
+				for (int i = 0; i < featureData.length; i++) {
+					convertedData[i] = (float) featureData[i];
+				}
+				featureList.add(convertedData);
+			} else if (feature instanceof FloatData) {
+				float[] featureData = ((FloatData) feature).getValues();
+				if (featureLength < 0) {
+					featureLength = featureData.length;
+				}
+				featureList.add(featureData);
+			}
+			feature = frontEnd.getData();
+		}
+
+		return featureList;
+	}
+
+	/**
+	 * Extracts a single feature frame from the supplied InputStream.
+	 * 
+	 * @return float[] or null if end of stream reached
+	 * 
+	 * @throws DataProcessinException
+	 *             if error occurred
+	 */
+	public float[] getNextFeatureFrame() throws DataProcessingException {
+		Data feature = frontEnd.getData();
+		if (feature == null)
+			return null;
+
+		while (!(feature instanceof DoubleData || feature instanceof FloatData)) {
+			feature = frontEnd.getData();
+			if (feature == null)
+				return null;
+		}
+
+		if (feature instanceof DoubleData) {
+			double[] featureData = ((DoubleData) feature).getValues();
+			if (featureLength < 0) {
+				featureLength = featureData.length;
+			}
+			float[] convertedData = new float[featureData.length];
+			for (int i = 0; i < featureData.length; i++) {
+				convertedData[i] = (float) featureData[i];
+			}
+
+			return convertedData;
+		} else if (feature instanceof FloatData) {
+			float[] featureData = ((FloatData) feature).getValues();
+			if (featureLength < 0) {
+				featureLength = featureData.length;
+			}
+
+			return featureData;
+		}
+
+		return null;
+	}
+
+}
+

sphinx4-core/src/test/java/edu/cmu/sphinx/api/FeatureExtractorTest.java

@@ -0,0 +1,101 @@
+/**
+ * 
+ */
+package edu.cmu.sphinx.api;
+
+import static org.testng.AssertJUnit.assertEquals;
+
+import java.io.IOException;
+import java.io.InputStream;
+import java.util.List;
+import java.util.Scanner;
+
+import org.testng.annotations.AfterTest;
+import org.testng.annotations.BeforeTest;
+import org.testng.annotations.Test;
+
+/**
+ * @author Vladisav Jelisavcic
+ *
+ */
+public class FeatureExtractorTest {
+
+	static final String FRONTEND_NAME = "cepstraFrontEnd";
+	private float[] features;
+	final int sampleRate = 16000;
+
+	@BeforeTest
+	public void setUp() {
+		InputStream asciiStream = FeatureExtractorTest.class
+				.getResourceAsStream("/edu/cmu/sphinx/api/10001-90210-01803.features");
+		Scanner sc = new Scanner(asciiStream);
+		int numDataPoints = sc.nextInt();
+
+		features = new float[numDataPoints];
+		int i = 0;
+		while (sc.hasNextFloat()) {
+			features[i++] = sc.nextFloat();
+		}
+		sc.close();
+	}
+
+	@AfterTest
+	public void tearDown() {
+
+	}
+
+	/**
+	 * Test method for
+	 * {@link edu.cmu.sphinx.api.FeatureExtractor#getAllFeatures()}.
+	 * 
+	 * @throws IOException
+	 */
+	@Test
+	public void testGetAllFeatures() throws IOException {
+		InputStream audioStream = FeatureExtractorTest.class
+				.getResourceAsStream("/edu/cmu/sphinx/tools/bandwidth/10001-90210-01803.wav");
+
+		FeatureExtractor fe = new FeatureExtractor(audioStream, sampleRate);
+		List<float[]> data = fe.getAllFeatures();
+
+		int numSamples = data.size();
+		int numFeatures = data.get(0).length;
+
+		int numDataPoints = numSamples * numFeatures;
+
+		assertEquals(features.length, numDataPoints); // check if all data
+														// points are loaded
+		for (int i = 0; i < numSamples; i++) {
+			for (int j = 0; j < numFeatures; j++) {
+				assertEquals(features[i * numFeatures + j], data.get(i)[j]);
+			}
+		}
+	}
+
+	/**
+	 * Test method for
+	 * {@link edu.cmu.sphinx.api.FeatureExtractor#getNextFeatureFrame()}.
+	 * 
+	 * @throws IOException
+	 */
+	@Test
+	public void testGetNextFeatureFrame() throws IOException {
+		InputStream audioStream = FeatureExtractorTest.class
+				.getResourceAsStream("/edu/cmu/sphinx/tools/bandwidth/10001-90210-01803.wav");
+
+		FeatureExtractor fe = new FeatureExtractor(audioStream, sampleRate);
+
+		int numDataPoints = 0;
+		float[] data;
+		while ((data = fe.getNextFeatureFrame()) != null) {
+			for (int i = 0; i < data.length; i++) {
+				assertEquals(features[i + numDataPoints], data[i]);
+			}
+			numDataPoints += data.length;
+		}
+
+		assertEquals(features.length, numDataPoints); // check if all data points are loaded
+	}
+
+}
+