Housekeeping

- Split out README.md into standalone.md, for all the bits that are standalone-specific
- Removed all automatically generated java files (generated by SWIG)
- Updated .gitignore to reflect the change above
- Updated DeepSpeechModel.java to be API-compatible with how it was, whilst also allowing the standalone to function (allowing you to load the libraries yourself, if wanted)

Author: TheDutchMC

native_client/java/.gitignore

@@ -18,4 +18,9 @@ jni/*.cpp
 jni/*.o
 
 #Setup.sh output
-build/
+build/
+
+#Autogenerated java files by SWIG
+libdeepspeech/src/main/java/org/deepspeech/libdeepspeech/*.java
+!libdeepspeech/src/main/java/org/deepspeech/libdeepspeech/DeepSpeechModel.java
+!libdeepspeech/src/main/java/org/deepspeech/libdeepspeech/DeepSpeechStreamingState.java=

native_client/java/README.md

@@ -17,199 +17,4 @@ For use with Android
 ## Standalone Java Bindings for DeepSpeech
 For use with standalone Java
 
-### Setup
-1. Get `libdeepspeech.so` by compiling or (from here)[], and copy it into `./libdeepspeech/libs/`
-2. Run `setup.sh` as root or with `sudo`
-3. You can now find the files needed for your project at `./build/`
-
-### Adding DeepSpeech to your Java project
->Note: You can do this on your own way too, but this works with the example usage code provided below.
-1. Copy `libdeepspeech.jar` from `./build/` into `{YOUR PROJECT ROOT}/libs/`
-2. Copy '*.so' from `./build/` into `{YOUR PROJECT ROOT}/src/main/resources/jni/x86_64/`
-3. Modify your `build.gradle` file to include:
-```groovy
-plugins {
-   	id 'com.github.johnrengelman.shadow' version '5.0.0' 
-}
-
-repositories {
-    flatDir {
-			dirs 'libs'
-    }
-}
-
-dependencies {
-    compile ':libdeepspeech'
-}
-```
-4. You can now compile your project with `./gradlew shadowJar`.
-
-### Notes
-- At this moment, this will only run on x86-64 Linux.
-
-### Example usage
-`LibDeepSpeech.java` example
-```java
-import java.io.File;
-import java.io.IOException;
-import java.io.InputStream;
-import java.net.URL;
-import java.nio.file.Files;
-
-import org.deepspeech.libdeepspeech.DeepSpeechModel;
-
-//This class facilitates the loading of the native libraries.
-//The implementation in here works with the Gradle guide provided in the README.md document.
-//However you can modify this to whatever you wish.
-public class LibDeepSpeech extends DeepSpeechModel {
-	
-	public LibDeepSpeech(String modelPath) {		
-		loadNativeLibs();
-		super.loadModel(modelPath);
-	}
-	
-	@Override
-	public void loadNativeLibs() {
-        String jniName = "libdeepspeech-jni.so";
-        String libName = "libdeepspeech.so";
-        
-        System.out.println("Setting up DeepSpeech...");
-        
-        URL jniUrl = DeepSpeechModel.class.getResource("/jni/x86_64/" + jniName);
-        URL libUrl = DeepSpeechModel.class.getResource("/jni/x86_64/" + libName);
-        File tmpDir = null;
-		try {
-			tmpDir = Files.createTempDirectory("libdeepspeech").toFile();
-		} catch (IOException e) {
-			e.printStackTrace();
-		}
-        tmpDir.deleteOnExit();
-    	
-        File jniTmpFile = new File(tmpDir, jniName);
-        jniTmpFile.deleteOnExit();
-        File libTmpFile = new File(tmpDir, libName);
-        libTmpFile.deleteOnExit();
-        
-        try (
-        		InputStream jniIn = jniUrl.openStream();
-        		InputStream libIn = libUrl.openStream();
-        ) {
-            Files.copy(jniIn, jniTmpFile.toPath());
-            Files.copy(libIn, libTmpFile.toPath());
-        } catch (IOException e) {
-			e.printStackTrace();
-		}
-                
-        System.load(jniTmpFile.getAbsolutePath());
-        System.load(libTmpFile.getAbsolutePath());
-	}
-}
-```
-
-`DeepSpeechExample.java`
-```java
-import java.io.FileNotFoundException;
-import java.io.IOException;
-import java.io.RandomAccessFile;
-import java.nio.ByteBuffer;
-import java.nio.ByteOrder;
-
-public class DeepSpeechExample {
-    private LibDeepSpeech deepSpeechModel = null
-
-    private final String TF_MODEL = "PATH_TO_YOUR_TENSORFLOW_MODEL";
-    private final int BEAM_WIDTH = 50;
-
-    public void createModel(String tensorFlowModel) {
-        if(this.deepSpeechModel == null) {
-            this.deepSpeechModel = new LibDeepSpeech(tensorFlowModel);
-            this.deepSpeechModel.setBeamWidth(BEAM_WIDTH);
-        }
-    }
-
-    //The format we want is a 16Khz 16 bit mono .WAV file!
-    public void doInfer(String audioFile) {
-        long inferenceExecTime = 0L;
-
-        //Create a new model to use during the inference
-    	this.newModel(this.TF_MODEL);
-    	
-    	System.out.println("Extracting audio features...");
-    	
-    	try {
-            RandomAccessFile wave = new RandomAccessFile(audioFile, "r");
-
-            //Assert that the audio format is PCM
-            wave.seek(20);
-            char audioFormat = this.readLEChar(wave);
-            assert (audioFormat == 1); // 1 is PCM
-
-            //Assert that the amount of channels is 1, meaning mono audio
-            wave.seek(22);
-            char numChannels = this.readLEChar(wave);
-            assert (numChannels == 1); // MONO
-
-            //Assert that the sample rate is the sample rate expected by the model
-            //This can vary per model!
-            wave.seek(24);
-            int sampleRate = this.readLEInt(wave);
-            assert (sampleRate == this.deepSpeechModel.sampleRate()); // desired sample rate
-
-            //Assert that the bits per sample is 16
-            wave.seek(34);
-            char bitsPerSample = this.readLEChar(wave);
-            assert (bitsPerSample == 16); // 16 bits per sample
-            
-            //Assert that the buffer size is more than 0
-            wave.seek(40);
-            int bufferSize = this.readLEInt(wave);
-            assert (bufferSize > 0);
-
-            //Read the actual contents of the audio
-            wave.seek(44);
-            byte[] bytes = new byte[bufferSize];
-            wave.readFully(bytes);
-
-            //Turn the byte[] into a short[] and set the correct byte order
-            short[] shorts = new short[bytes.length/2];
-            ByteBuffer.wrap(bytes).order(ByteOrder.LITTLE_ENDIAN).asShortBuffer().get(shorts);
-            
-            System.out.println("Running inference...");
-
-            //current time. Used later on to calculate the time the inference took
-            long inferenceStartTime = System.currentTimeMillis();
-
-            //This is where we actually do the inference
-            String decoded = this.deepSpeechModel.stt(shorts, shorts.length);
-
-            //Calculate how long it took to run the inference
-            inferenceExecTime = System.currentTimeMillis() - inferenceStartTime;
-
-            System.out.println(decoded);
-
-        } catch (FileNotFoundException e) {
-        	e.printStackTrace();
-        } catch (IOException e) {
-        	e.printStackTrace();
-        }
-    	
-    	System.out.println("Finished! Took " + inferenceExecTime);
-    }
-
-    //Helper function to read a char from a RandomAccessFile in little endian
-    private char readLEChar(RandomAccessFile f) throws IOException {
-        byte b1 = f.readByte();
-        byte b2 = f.readByte();
-        return (char)((b2 << 8) | b1);
-    }
-    
-    //Helper function to read an integer from a RandomAccessFile in little endian
-    private int readLEInt(RandomAccessFile f) throws IOException {
-        byte b1 = f.readByte();
-        byte b2 = f.readByte();
-        byte b3 = f.readByte();
-        byte b4 = f.readByte();
-        return (int)((b1 & 0xFF) | (b2 & 0xFF) << 8 | (b3 & 0xFF) << 16 | (b4 & 0xFF) << 24);
-    }
-}
-```
+See [standalone.md](standalone.md)

native_client/java/libdeepspeech/src/main/java/org/deepspeech/libdeepspeech/CandidateTranscript.java

@@ -9,9 +9,28 @@
 /**
  * @brief Exposes a DeepSpeech model in Java
  **/
-public abstract class DeepSpeechModel {
+public class DeepSpeechModel {
 
-    public abstract void loadNativeLibs();
+    /**
+    * Create a new DeepSpeechModel object and load the native libraries.<br>
+    * Use on: Android
+    */
+    public DeepSpeechModel() {
+        System.loadLibrary("deepspeech-jni");
+        System.loadLibrary("deepspeech");
+    }
+
+    /**
+    * Create a new DeepSpeechModel object.<br>
+    * Use on: Android, Standalone
+    * @param loadNativeLibrariesManually if true, the native libraries will not be loaded. The developer must do this manually then.
+    */
+    public DeepSpeechModel(boolean loadNativeLibrariesManually) {
+        if(!loadNativeLibrariesManually) {
+            System.loadLibrary("deepspeech-jni");
+            System.loadLibrary("deepspeech");
+        }
+    }
 
     // FIXME: We should have something better than those SWIGTYPE_*
     private SWIGTYPE_p_p_ModelState _mspp;