Add JNI support for spoken language identification (#782)

kotlin-api-examples/Main.kt

@@ -7,6 +7,7 @@ fun callback(samples: FloatArray): Unit {
 }
 
 fun main() {
+  testSpokenLanguageIdentifcation()
   testAudioTagging()
   testSpeakerRecognition()
   testTts()
@@ -14,6 +15,41 @@ fun main() {
   testAsr("zipformer2-ctc")
 }
 
+fun testSpokenLanguageIdentifcation() {
+  val config = SpokenLanguageIdentificationConfig(
+    whisper = SpokenLanguageIdentificationWhisperConfig(
+      encoder = "./sherpa-onnx-whisper-tiny/tiny-encoder.int8.onnx",
+      decoder = "./sherpa-onnx-whisper-tiny/tiny-decoder.int8.onnx",
+      tailPaddings = 33,
+    ),
+    numThreads=1,
+    debug=true,
+    provider="cpu",
+  )
+  val slid = SpokenLanguageIdentification(assetManager=null, config=config)
+
+  val testFiles = arrayOf(
+    "./spoken-language-identification-test-wavs/ar-arabic.wav",
+    "./spoken-language-identification-test-wavs/bg-bulgarian.wav",
+    "./spoken-language-identification-test-wavs/de-german.wav",
+  )
+
+  for (waveFilename in testFiles) {
+    val objArray = WaveReader.readWaveFromFile(
+        filename = waveFilename,
+    )
+    val samples: FloatArray = objArray[0] as FloatArray
+    val sampleRate: Int = objArray[1] as Int
+
+    val stream = slid.createStream()
+    stream.acceptWaveform(samples, sampleRate = sampleRate)
+    val lang = slid.compute(stream)
+    stream.release()
+    println(waveFilename)
+    println(lang)
+  }
+}
+
 fun testAudioTagging() {
   val config = AudioTaggingConfig(
       model=AudioTaggingModelConfig(

kotlin-api-examples/SpokenLanguageIdentification.kt

@@ -0,0 +1,71 @@
+package com.k2fsa.sherpa.onnx
+
+import android.content.res.AssetManager
+import android.util.Log
+
+private val TAG = "sherpa-onnx"
+
+data class SpokenLanguageIdentificationWhisperConfig (
+    var encoder: String,
+    var decoder: String,
+    var tailPaddings: Int = -1,
+)
+
+data class SpokenLanguageIdentificationConfig (
+    var whisper: SpokenLanguageIdentificationWhisperConfig,
+    var numThreads: Int = 1,
+    var debug: Boolean = false,
+    var provider: String = "cpu",
+)
+
+class SpokenLanguageIdentification (
+    assetManager: AssetManager? = null,
+    config: SpokenLanguageIdentificationConfig,
+) {
+    private var ptr: Long
+
+    init {
+        ptr = if (assetManager != null) {
+            newFromAsset(assetManager, config)
+        } else {
+            newFromFile(config)
+        }
+    }
+
+    protected fun finalize() {
+        if (ptr != 0L) {
+            delete(ptr)
+            ptr = 0
+        }
+    }
+
+    fun release() = finalize()
+
+    fun createStream(): OfflineStream {
+        val p = createStream(ptr)
+        return OfflineStream(p)
+    }
+
+    fun compute(stream: OfflineStream) =  compute(ptr, stream.ptr)
+
+    private external fun newFromAsset(
+        assetManager: AssetManager,
+        config: SpokenLanguageIdentificationConfig,
+    ): Long
+
+    private external fun newFromFile(
+        config: SpokenLanguageIdentificationConfig,
+    ): Long
+
+    private external fun delete(ptr: Long)
+
+    private external fun createStream(ptr: Long): Long
+
+    private external fun compute(ptr: Long, streamPtr: Long): String
+
+    companion object {
+        init {
+            System.loadLibrary("sherpa-onnx-jni")
+        }
+    }
+}

kotlin-api-examples/run.sh

@@ -30,19 +30,19 @@ cd ../kotlin-api-examples
 
 function testSpeakerEmbeddingExtractor() {
   if [ ! -f ./3dspeaker_speech_eres2net_large_sv_zh-cn_3dspeaker_16k.onnx ]; then
-    wget -q https://github.com/k2-fsa/sherpa-onnx/releases/download/speaker-recongition-models/3dspeaker_speech_eres2net_large_sv_zh-cn_3dspeaker_16k.onnx
+    curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/speaker-recongition-models/3dspeaker_speech_eres2net_large_sv_zh-cn_3dspeaker_16k.onnx
   fi
 
   if [ ! -f ./speaker1_a_cn_16k.wav ]; then
-    wget -q https://github.com/csukuangfj/sr-data/raw/main/test/3d-speaker/speaker1_a_cn_16k.wav
+    curl -SL -O https://github.com/csukuangfj/sr-data/raw/main/test/3d-speaker/speaker1_a_cn_16k.wav
   fi
 
   if [ ! -f ./speaker1_b_cn_16k.wav ]; then
-    wget -q https://github.com/csukuangfj/sr-data/raw/main/test/3d-speaker/speaker1_b_cn_16k.wav
+    curl -SL -O https://github.com/csukuangfj/sr-data/raw/main/test/3d-speaker/speaker1_b_cn_16k.wav
   fi
 
   if [ ! -f ./speaker2_a_cn_16k.wav ]; then
-    wget -q https://github.com/csukuangfj/sr-data/raw/main/test/3d-speaker/speaker2_a_cn_16k.wav
+    curl -SL -O https://github.com/csukuangfj/sr-data/raw/main/test/3d-speaker/speaker2_a_cn_16k.wav
   fi
 }
 
@@ -53,7 +53,7 @@ function testAsr() {
   fi
 
   if [ ! -d ./sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13 ]; then
-    wget -q https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13.tar.bz2
+    curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13.tar.bz2
     tar xvf sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13.tar.bz2
     rm sherpa-onnx-streaming-zipformer-ctc-multi-zh-hans-2023-12-13.tar.bz2
   fi
@@ -61,7 +61,7 @@ function testAsr() {
 
 function testTts() {
   if [ ! -f ./vits-piper-en_US-amy-low/en_US-amy-low.onnx ]; then
-    wget -q https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/vits-piper-en_US-amy-low.tar.bz2
+    curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/tts-models/vits-piper-en_US-amy-low.tar.bz2
     tar xf vits-piper-en_US-amy-low.tar.bz2
     rm vits-piper-en_US-amy-low.tar.bz2
   fi
@@ -75,7 +75,22 @@ function testAudioTagging() {
   fi
 }
 
+function testSpokenLanguageIdentification() {
+  if [ ! -f ./sherpa-onnx-whisper-tiny/tiny-encoder.int8.onnx ]; then
+    curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-whisper-tiny.tar.bz2
+    tar xvf sherpa-onnx-whisper-tiny.tar.bz2
+    rm sherpa-onnx-whisper-tiny.tar.bz2
+  fi
+
+  if [ ! -f ./spoken-language-identification-test-wavs/ar-arabic.wav ]; then
+    curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/spoken-language-identification-test-wavs.tar.bz2
+    tar xvf spoken-language-identification-test-wavs.tar.bz2
+    rm spoken-language-identification-test-wavs.tar.bz2
+  fi
+}
+
 function test() {
+  testSpokenLanguageIdentification
   testAudioTagging
   testSpeakerEmbeddingExtractor
   testAsr
@@ -90,6 +105,7 @@ kotlinc-jvm -include-runtime -d main.jar \
   OfflineStream.kt \
   SherpaOnnx.kt \
   Speaker.kt \
+  SpokenLanguageIdentification.kt \
   Tts.kt \
   WaveReader.kt \
   faked-asset-manager.kt \
@@ -101,13 +117,13 @@ java -Djava.library.path=../build/lib -jar main.jar
 
 function testTwoPass() {
   if [ ! -f ./sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/encoder-epoch-99-avg-1.int8.onnx ]; then
-    wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-streaming-zipformer-en-20M-2023-02-17.tar.bz2
+    curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-streaming-zipformer-en-20M-2023-02-17.tar.bz2
     tar xvf sherpa-onnx-streaming-zipformer-en-20M-2023-02-17.tar.bz2
     rm sherpa-onnx-streaming-zipformer-en-20M-2023-02-17.tar.bz2
   fi
 
   if [ ! -f ./sherpa-onnx-whisper-tiny.en/tiny.en-encoder.int8.onnx ]; then
-    wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-whisper-tiny.en.tar.bz2
+    curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-whisper-tiny.en.tar.bz2
     tar xvf sherpa-onnx-whisper-tiny.en.tar.bz2
     rm sherpa-onnx-whisper-tiny.en.tar.bz2
   fi