add vad+sense voice example for C API (#1291)

@@ -99,6 +99,45 @@ jobs:
           ./run.sh
           rm -rf sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20
 
+      - name: Test vad + sense-voice
+        shell: bash
+        run: |
+          gcc -o vad-sense-voice-c-api ./c-api-examples/vad-sense-voice-c-api.c \
+            -I ./build/install/include \
+            -L ./build/install/lib/ \
+            -l sherpa-onnx-c-api \
+            -l onnxruntime
+
+          ls -lh vad-sense-voice-c-api
+
+          if [[ ${{ matrix.os }} == ubuntu-latest ]]; then
+            ldd ./vad-sense-voice-c-api
+            echo "----"
+            readelf -d ./vad-sense-voice-c-api
+          fi
+
+          # Now download models
+          #
+          curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/silero_vad.onnx
+          curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/lei-jun-test.wav
+
+          curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2
+          tar xvf sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2
+          rm sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2
+
+          ls -lh sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17
+          echo "---"
+          ls -lh sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17/test_wavs
+
+          export LD_LIBRARY_PATH=$PWD/build/install/lib:$LD_LIBRARY_PATH
+          export DYLD_LIBRARY_PATH=$PWD/build/install/lib:$DYLD_LIBRARY_PATH
+
+          ./vad-sense-voice-c-api
+
+          rm -rf sherpa-onnx-sense-voice-*
+          rm -rf *.onnx
+          rm *.wav
+
       - name: Test sense-voice
         shell: bash
         run: |

@@ -45,6 +45,9 @@ target_link_libraries(streaming-paraformer-c-api sherpa-onnx-c-api)
 add_executable(telespeech-c-api telespeech-c-api.c)
 target_link_libraries(telespeech-c-api sherpa-onnx-c-api)
 
+add_executable(vad-sense-voice-c-api vad-sense-voice-c-api.c)
+target_link_libraries(vad-sense-voice-c-api sherpa-onnx-c-api)
+
 if(SHERPA_ONNX_HAS_ALSA)
   add_subdirectory(./asr-microphone-example)
 elseif((UNIX AND NOT APPLE) OR LINUX)

@@ -3,7 +3,8 @@
 // Copyright (c)  2024  Xiaomi Corporation
 
 //
-// This file demonstrates how to use non-streaming Paraformer with sherpa-onnx's C API.
+// This file demonstrates how to use non-streaming Paraformer with sherpa-onnx's
+// C API.
 // clang-format off
 //
 // wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-paraformer-zh-small-2024-03-09.tar.bz2
@@ -19,13 +20,14 @@
 #include "sherpa-onnx/c-api/c-api.h"
 
 int32_t main() {
-
-  const char *wav_filename = "sherpa-onnx-paraformer-zh-small-2024-03-09/test_wavs/0.wav";
-  const char *model_filename = "sherpa-onnx-paraformer-zh-small-2024-03-09/model.int8.onnx";
-  const char *tokens_filename = "sherpa-onnx-paraformer-zh-small-2024-03-09/tokens.txt";
+  const char *wav_filename =
+      "sherpa-onnx-paraformer-zh-small-2024-03-09/test_wavs/0.wav";
+  const char *model_filename =
+      "sherpa-onnx-paraformer-zh-small-2024-03-09/model.int8.onnx";
+  const char *tokens_filename =
+      "sherpa-onnx-paraformer-zh-small-2024-03-09/tokens.txt";
   const char *provider = "cpu";
 
-
   const SherpaOnnxWave *wave = SherpaOnnxReadWave(wav_filename);
   if (wave == NULL) {
     fprintf(stderr, "Failed to read %s\n", wav_filename);

@@ -19,8 +19,6 @@
 #include "sherpa-onnx/c-api/c-api.h"
 
 int32_t main() {
-  // You can find more test waves from
-  // https://hf-mirror.com/spaces/k2-fsa/spoken-language-identification/tree/main/test_wavs
   const char *wav_filename =
       "./sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17/test_wavs/en.wav";
   const char *model_filename =

@@ -3,7 +3,8 @@
 // Copyright (c)  2024  Xiaomi Corporation
 
 //
-// This file demonstrates how to use streaming Paraformer with sherpa-onnx's C API.
+// This file demonstrates how to use streaming Paraformer with sherpa-onnx's C
+// API.
 // clang-format off
 // 
 // wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-streaming-paraformer-bilingual-zh-en.tar.bz2
@@ -19,14 +20,16 @@
 #include "sherpa-onnx/c-api/c-api.h"
 
 int32_t main() {
-
-  const char *wav_filename = "sherpa-onnx-streaming-paraformer-bilingual-zh-en/test_wavs/0.wav";
-  const char *encoder_filename = "sherpa-onnx-streaming-paraformer-bilingual-zh-en/encoder.int8.onnx";
-  const char *decoder_filename = "sherpa-onnx-streaming-paraformer-bilingual-zh-en/decoder.int8.onnx";
-  const char *tokens_filename = "sherpa-onnx-streaming-paraformer-bilingual-zh-en/tokens.txt";
+  const char *wav_filename =
+      "sherpa-onnx-streaming-paraformer-bilingual-zh-en/test_wavs/0.wav";
+  const char *encoder_filename =
+      "sherpa-onnx-streaming-paraformer-bilingual-zh-en/encoder.int8.onnx";
+  const char *decoder_filename =
+      "sherpa-onnx-streaming-paraformer-bilingual-zh-en/decoder.int8.onnx";
+  const char *tokens_filename =
+      "sherpa-onnx-streaming-paraformer-bilingual-zh-en/tokens.txt";
   const char *provider = "cpu";
 
-
   const SherpaOnnxWave *wave = SherpaOnnxReadWave(wav_filename);
   if (wave == NULL) {
     fprintf(stderr, "Failed to read %s\n", wav_filename);
@@ -39,7 +42,6 @@ int32_t main() {
   paraformer_config.encoder = encoder_filename;
   paraformer_config.decoder = decoder_filename;
 
-
   // Online model config
   SherpaOnnxOnlineModelConfig online_model_config;
   memset(&online_model_config, 0, sizeof(online_model_config));

@@ -3,7 +3,8 @@
 // Copyright (c)  2024  Xiaomi Corporation
 
 //
-// This file demonstrates how to use streaming Zipformer with sherpa-onnx's C API.
+// This file demonstrates how to use streaming Zipformer with sherpa-onnx's C
+// API.
 // clang-format off
 // 
 // wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-streaming-zipformer-en-20M-2023-02-17.tar.bz2
@@ -19,15 +20,21 @@
 #include "sherpa-onnx/c-api/c-api.h"
 
 int32_t main() {
-
-  const char *wav_filename = "sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/test_wavs/0.wav";
-  const char *encoder_filename = "sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/encoder-epoch-99-avg-1.onnx";
-  const char *decoder_filename = "sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/decoder-epoch-99-avg-1.onnx";
-  const char *joiner_filename = "sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/joiner-epoch-99-avg-1.onnx";
-  const char *tokens_filename = "sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/tokens.txt";
+  const char *wav_filename =
+      "sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/test_wavs/0.wav";
+  const char *encoder_filename =
+      "sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/"
+      "encoder-epoch-99-avg-1.onnx";
+  const char *decoder_filename =
+      "sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/"
+      "decoder-epoch-99-avg-1.onnx";
+  const char *joiner_filename =
+      "sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/"
+      "joiner-epoch-99-avg-1.onnx";
+  const char *tokens_filename =
+      "sherpa-onnx-streaming-zipformer-en-20M-2023-02-17/tokens.txt";
   const char *provider = "cpu";
 
-
   const SherpaOnnxWave *wave = SherpaOnnxReadWave(wav_filename);
   if (wave == NULL) {
     fprintf(stderr, "Failed to read %s\n", wav_filename);
@@ -41,7 +48,6 @@ int32_t main() {
   zipformer_config.decoder = decoder_filename;
   zipformer_config.joiner = joiner_filename;
 
-
   // Online model config
   SherpaOnnxOnlineModelConfig online_model_config;
   memset(&online_model_config, 0, sizeof(online_model_config));

@@ -3,7 +3,8 @@
 // Copyright (c)  2024  Xiaomi Corporation
 
 //
-// This file demonstrates how to use TeleSpeech-ASR CTC model with sherpa-onnx's C API.
+// This file demonstrates how to use TeleSpeech-ASR CTC model with sherpa-onnx's
+// C API.
 // clang-format off
 //
 // wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04.tar.bz2
@@ -19,10 +20,12 @@
 #include "sherpa-onnx/c-api/c-api.h"
 
 int32_t main() {
-
-  const char *wav_filename = "sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04/test_wavs/3-sichuan.wav";
-  const char *model_filename = "sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04/model.int8.onnx";
-  const char *tokens_filename = "sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04/tokens.txt";
+  const char *wav_filename =
+      "sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04/test_wavs/3-sichuan.wav";
+  const char *model_filename =
+      "sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04/model.int8.onnx";
+  const char *tokens_filename =
+      "sherpa-onnx-telespeech-ctc-int8-zh-2024-06-04/tokens.txt";
   const char *provider = "cpu";
 
   const SherpaOnnxWave *wave = SherpaOnnxReadWave(wav_filename);

+// c-api-examples/vad-sense-voice-c-api.c
+//
+// Copyright (c)  2024  Xiaomi Corporation
+
+//
+// This file demonstrates how to use VAD + SenseVoice with sherpa-onnx's C API.
+// clang-format off
+//
+// wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/silero_vad.onnx
+// wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/lei-jun-test.wav
+//
+// wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2
+// tar xvf sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2
+// rm sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2
+//
+// clang-format on
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "sherpa-onnx/c-api/c-api.h"
+
+int32_t main() {
+  const char *wav_filename = "./lei-jun-test.wav";
+  const char *vad_filename = "./silero_vad.onnx";
+  const char *model_filename =
+      "./sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17/model.int8.onnx";
+  const char *tokens_filename =
+      "./sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17/tokens.txt";
+  const char *language = "auto";
+  const char *provider = "cpu";
+  int32_t use_inverse_text_normalization = 1;
+
+  const SherpaOnnxWave *wave = SherpaOnnxReadWave(wav_filename);
+  if (wave == NULL) {
+    fprintf(stderr, "Failed to read %s\n", wav_filename);
+    return -1;
+  }
+
+  if (wave->sample_rate != 16000) {
+    fprintf(stderr, "Expect the sample rate to be 16000. Given: %d\n",
+            wave->sample_rate);
+    SherpaOnnxFreeWave(wave);
+    return -1;
+  }
+
+  SherpaOnnxOfflineSenseVoiceModelConfig sense_voice_config;
+  memset(&sense_voice_config, 0, sizeof(sense_voice_config));
+  sense_voice_config.model = model_filename;
+  sense_voice_config.language = language;
+  sense_voice_config.use_itn = use_inverse_text_normalization;
+
+  // Offline model config
+  SherpaOnnxOfflineModelConfig offline_model_config;
+  memset(&offline_model_config, 0, sizeof(offline_model_config));
+  offline_model_config.debug = 0;
+  offline_model_config.num_threads = 1;
+  offline_model_config.provider = provider;
+  offline_model_config.tokens = tokens_filename;
+  offline_model_config.sense_voice = sense_voice_config;
+
+  // Recognizer config
+  SherpaOnnxOfflineRecognizerConfig recognizer_config;
+  memset(&recognizer_config, 0, sizeof(recognizer_config));
+  recognizer_config.decoding_method = "greedy_search";
+  recognizer_config.model_config = offline_model_config;
+
+  SherpaOnnxOfflineRecognizer *recognizer =
+      SherpaOnnxCreateOfflineRecognizer(&recognizer_config);
+
+  if (recognizer == NULL) {
+    fprintf(stderr, "Please check your recognizer config!\n");
+    SherpaOnnxFreeWave(wave);
+    return -1;
+  }
+
+  SherpaOnnxVadModelConfig vadConfig;
+  memset(&vadConfig, 0, sizeof(vadConfig));
+  vadConfig.silero_vad.model = vad_filename;
+  vadConfig.silero_vad.threshold = 0.5;
+  vadConfig.silero_vad.min_silence_duration = 0.5;
+  vadConfig.silero_vad.min_speech_duration = 0.5;
+  vadConfig.silero_vad.window_size = 512;
+  vadConfig.sample_rate = 16000;
+  vadConfig.num_threads = 1;
+  vadConfig.debug = 1;
+
+  SherpaOnnxVoiceActivityDetector *vad =
+      SherpaOnnxCreateVoiceActivityDetector(&vadConfig, 30);
+
+  if (vad == NULL) {
+    fprintf(stderr, "Please check your recognizer config!\n");
+    SherpaOnnxFreeWave(wave);
+    SherpaOnnxDestroyOfflineRecognizer(recognizer);
+    return -1;
+  }
+
+  int32_t window_size = vadConfig.silero_vad.window_size;
+  int32_t i = 0;
+
+  while (i + window_size < wave->num_samples) {
+    SherpaOnnxVoiceActivityDetectorAcceptWaveform(vad, wave->samples + i,
+                                                  window_size);
+    i += window_size;
+
+    while (!SherpaOnnxVoiceActivityDetectorEmpty(vad)) {
+      const SherpaOnnxSpeechSegment *segment =
+          SherpaOnnxVoiceActivityDetectorFront(vad);
+
+      SherpaOnnxOfflineStream *stream =
+          SherpaOnnxCreateOfflineStream(recognizer);
+      SherpaOnnxAcceptWaveformOffline(stream, wave->sample_rate,
+                                      segment->samples, segment->n);
+
+      SherpaOnnxDecodeOfflineStream(recognizer, stream);
+
+      const SherpaOnnxOfflineRecognizerResult *result =
+          SherpaOnnxGetOfflineStreamResult(stream);
+
+      float start = segment->start / 16000.0f;
+      float duration = segment->n / 16000.0f;
+      float stop = start + duration;
+
+      fprintf(stderr, "%.3f -- %.3f: %s\n", start, stop, result->text);
+
+      SherpaOnnxDestroyOfflineRecognizerResult(result);
+      SherpaOnnxDestroyOfflineStream(stream);
+
+      SherpaOnnxDestroySpeechSegment(segment);
+      SherpaOnnxVoiceActivityDetectorPop(vad);
+    }
+  }
+
+  SherpaOnnxVoiceActivityDetectorFlush(vad);
+
+  while (!SherpaOnnxVoiceActivityDetectorEmpty(vad)) {
+    const SherpaOnnxSpeechSegment *segment =
+        SherpaOnnxVoiceActivityDetectorFront(vad);
+
+    SherpaOnnxOfflineStream *stream = SherpaOnnxCreateOfflineStream(recognizer);
+    SherpaOnnxAcceptWaveformOffline(stream, wave->sample_rate, segment->samples,
+                                    segment->n);
+
+    SherpaOnnxDecodeOfflineStream(recognizer, stream);
+
+    const SherpaOnnxOfflineRecognizerResult *result =
+        SherpaOnnxGetOfflineStreamResult(stream);
+
+    float start = segment->start / 16000.0f;
+    float duration = segment->n / 16000.0f;
+    float stop = start + duration;
+
+    fprintf(stderr, "%.3f -- %.3f: %s\n", start, stop, result->text);
+
+    SherpaOnnxDestroyOfflineRecognizerResult(result);
+    SherpaOnnxDestroyOfflineStream(stream);
+
+    SherpaOnnxDestroySpeechSegment(segment);
+    SherpaOnnxVoiceActivityDetectorPop(vad);
+  }
+
+  SherpaOnnxDestroyOfflineRecognizer(recognizer);
+  SherpaOnnxDestroyVoiceActivityDetector(vad);
+  SherpaOnnxFreeWave(wave);
+
+  return 0;
+}

@@ -3,7 +3,8 @@
 // Copyright (c)  2024  Xiaomi Corporation
 
 //
-// This file demonstrates how to use non-streaming Zipformer with sherpa-onnx's C API.
+// This file demonstrates how to use non-streaming Zipformer with sherpa-onnx's
+// C API.
 // clang-format off
 //
 // wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-zipformer-small-en-2023-06-26.tar.bz2
@@ -19,15 +20,18 @@
 #include "sherpa-onnx/c-api/c-api.h"
 
 int32_t main() {
-
-  const char *wav_filename = "sherpa-onnx-zipformer-small-en-2023-06-26/test_wavs/0.wav";
-  const char *encoder_filename = "sherpa-onnx-zipformer-small-en-2023-06-26/encoder-epoch-99-avg-1.onnx";
-  const char *decoder_filename = "sherpa-onnx-zipformer-small-en-2023-06-26/decoder-epoch-99-avg-1.onnx";
-  const char *joiner_filename = "sherpa-onnx-zipformer-small-en-2023-06-26/joiner-epoch-99-avg-1.onnx";
-  const char *tokens_filename = "sherpa-onnx-zipformer-small-en-2023-06-26/tokens.txt";
+  const char *wav_filename =
+      "sherpa-onnx-zipformer-small-en-2023-06-26/test_wavs/0.wav";
+  const char *encoder_filename =
+      "sherpa-onnx-zipformer-small-en-2023-06-26/encoder-epoch-99-avg-1.onnx";
+  const char *decoder_filename =
+      "sherpa-onnx-zipformer-small-en-2023-06-26/decoder-epoch-99-avg-1.onnx";
+  const char *joiner_filename =
+      "sherpa-onnx-zipformer-small-en-2023-06-26/joiner-epoch-99-avg-1.onnx";
+  const char *tokens_filename =
+      "sherpa-onnx-zipformer-small-en-2023-06-26/tokens.txt";
   const char *provider = "cpu";
 
-
   const SherpaOnnxWave *wave = SherpaOnnxReadWave(wav_filename);
   if (wave == NULL) {
     fprintf(stderr, "Failed to read %s\n", wav_filename);
@@ -41,7 +45,6 @@ int32_t main() {
   zipformer_config.decoder = decoder_filename;
   zipformer_config.joiner = joiner_filename;
 
-
   // Offline model config
   SherpaOnnxOfflineModelConfig offline_model_config;
   memset(&offline_model_config, 0, sizeof(offline_model_config));