update wbsocket for sensevoice & onnx models

雾聪

2024-09-25 3e44172c8b927ffc69b585d4fd80b458cb18ba97

update wbsocket for sensevoice & onnx models

 runtime/websocket/bin/funasr-wss-client.cpp

@@ -115,7 +115,7 @@

    // This method will block until the connection is complete  
    void run(const std::string& uri, const std::vector<string>& wav_list, const std::vector<string>& wav_ids, 
             int audio_fs, const std::unordered_map<std::string, int>& hws_map, int use_itn=1) {
             int audio_fs, const std::unordered_map<std::string, int>& hws_map, int use_itn=1, int svs_itn=1) {
        // Create a new connection to the given URI
        websocketpp::lib::error_code ec;
        typename websocketpp::client<T>::connection_ptr con =
@@ -147,7 +147,7 @@
                cv.wait(lock);
            }
            total_send += 1;
            send_wav_data(wav_list[i], wav_ids[i], audio_fs, hws_map, send_hotword, use_itn);
            send_wav_data(wav_list[i], wav_ids[i], audio_fs, hws_map, send_hotword, use_itn, svs_itn);
            if(send_hotword){
                send_hotword = false;
            }
@@ -186,7 +186,7 @@
    // send wav to server
    void send_wav_data(string wav_path, string wav_id, int audio_fs,
        const std::unordered_map<std::string, int>& hws_map, 
        bool send_hotword, bool use_itn) {
        bool send_hotword, bool use_itn, bool svs_itn) {
        uint64_t count = 0;
        std::stringstream val;

@@ -239,8 +239,12 @@
        jsonbegin["wav_format"] = wav_format;
        jsonbegin["audio_fs"] = sampling_rate;
        jsonbegin["itn"] = true;
        jsonbegin["svs_itn"] = true;
        if(use_itn == 0){
            jsonbegin["itn"] = false;
        }
        if(svs_itn == 0){
            jsonbegin["svs_itn"] = false;
        }
        jsonbegin["is_speaking"] = true;
        if(send_hotword){
@@ -368,6 +372,9 @@
    TCLAP::ValueArg<int> use_itn_(
        "", "use-itn",
        "use-itn is 1 means use itn, 0 means not use itn", false, 1, "int");
    TCLAP::ValueArg<int> svs_itn_(
        "", "svs-itn",
        "svs-itn is 1 means use itn and punc, 0 means not use", false, 1, "int");
    TCLAP::ValueArg<std::string> hotword_("", HOTWORD,
        "the hotword file, one hotword perline, Format: Hotword Weight (could be: 阿里巴巴 20)", false, "", "string");

@@ -378,6 +385,7 @@
    cmd.add(thread_num_);
    cmd.add(is_ssl_);
    cmd.add(use_itn_);
    cmd.add(svs_itn_);
    cmd.add(hotword_);
    cmd.parse(argc, argv);

@@ -387,6 +395,7 @@
    int threads_num = thread_num_.getValue();
    int is_ssl = is_ssl_.getValue();
    int use_itn = use_itn_.getValue();
    int svs_itn = svs_itn_.getValue();

    std::vector<websocketpp::lib::thread> client_threads;
    std::string uri = "";
@@ -431,17 +440,17 @@
    
    int audio_fs = audio_fs_.getValue();
    for (size_t i = 0; i < threads_num; i++) {
        client_threads.emplace_back([uri, wav_list, wav_ids, audio_fs, is_ssl, hws_map, use_itn]() {
        client_threads.emplace_back([uri, wav_list, wav_ids, audio_fs, is_ssl, hws_map, use_itn, svs_itn]() {
          if (is_ssl == 1) {
            WebsocketClient<websocketpp::config::asio_tls_client> c(is_ssl);

            c.m_client.set_tls_init_handler(bind(&OnTlsInit, ::_1));

            c.run(uri, wav_list, wav_ids, audio_fs, hws_map, use_itn);
            c.run(uri, wav_list, wav_ids, audio_fs, hws_map, use_itn, svs_itn);
          } else {
            WebsocketClient<websocketpp::config::asio_client> c(is_ssl);

            c.run(uri, wav_list, wav_ids, audio_fs, hws_map, use_itn);
            c.run(uri, wav_list, wav_ids, audio_fs, hws_map, use_itn, svs_itn);
          }
        });
    }

 runtime/websocket/bin/funasr-wss-server-2pass.cpp

@@ -55,11 +55,11 @@

    TCLAP::ValueArg<std::string> offline_model_revision(
        "", "offline-model-revision", "ASR offline model revision", false,
        "v2.0.4", "string");
        "v2.0.5", "string");

    TCLAP::ValueArg<std::string> online_model_revision(
        "", "online-model-revision", "ASR online model revision", false,
        "v2.0.4", "string");
        "v2.0.5", "string");

    TCLAP::ValueArg<std::string> quantize(
        "", QUANTIZE,
@@ -85,7 +85,7 @@
        "model_quant.onnx, punc.yaml",
        false, "damo/punc_ct-transformer_zh-cn-common-vad_realtime-vocab272727-onnx", "string");
    TCLAP::ValueArg<std::string> punc_revision(
        "", "punc-revision", "PUNC model revision", false, "v2.0.4", "string");
        "", "punc-revision", "PUNC model revision", false, "v2.0.5", "string");
    TCLAP::ValueArg<std::string> punc_quant(
        "", PUNC_QUANT,
        "true (Default), load the model of model_quant.onnx in punc_dir. If "
@@ -262,7 +262,7 @@

        size_t found = s_offline_asr_path.find("speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404");
        if (found != std::string::npos) {
            model_path["offline-model-revision"]="v2.0.4";
            model_path["offline-model-revision"]="v2.0.5";
        }

        found = s_offline_asr_path.find("speech_paraformer-large-contextual_asr_nat-zh-cn-16k-common-vocab8404");
@@ -272,7 +272,7 @@

        found = s_offline_asr_path.find("speech_paraformer-large_asr_nat-en-16k-common-vocab10020");
        if (found != std::string::npos) {
            model_path["model-revision"]="v2.0.4";
            model_path["model-revision"]="v2.0.5";
            s_itn_path="";
            s_lm_path="";
        }

 runtime/websocket/bin/funasr-wss-server.cpp

@@ -50,7 +50,7 @@
    TCLAP::ValueArg<std::string> model_revision(
        "", "model-revision",
        "ASR model revision",
        false, "v2.0.4", "string");
        false, "v2.0.5", "string");
    TCLAP::ValueArg<std::string> quantize(
        "", QUANTIZE,
        "true (Default), load the model of model_quant.onnx in model_dir. If set "
@@ -81,7 +81,7 @@
    TCLAP::ValueArg<std::string> punc_revision(
        "", "punc-revision",
        "PUNC model revision",
        false, "v2.0.4", "string");
        false, "v2.0.5", "string");
    TCLAP::ValueArg<std::string> punc_quant(
        "", PUNC_QUANT,
        "true (Default), load the model of model_quant.onnx in punc_dir. If set "
@@ -247,7 +247,7 @@
            // modify model-revision by model name
            size_t found = s_asr_path.find("speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404");
            if (found != std::string::npos) {
                model_path["model-revision"]="v2.0.4";
                model_path["model-revision"]="v2.0.5";
            }

            found = s_asr_path.find("speech_paraformer-large-contextual_asr_nat-zh-cn-16k-common-vocab8404");
@@ -257,11 +257,22 @@

            found = s_asr_path.find("speech_paraformer-large_asr_nat-en-16k-common-vocab10020");
            if (found != std::string::npos) {
                model_path["model-revision"]="v2.0.4";
                model_path["model-revision"]="v2.0.5";
                s_itn_path="";
                s_lm_path="";
            }

            found = s_asr_path.find(MODEL_SVS);
            if (found != std::string::npos) {
                model_path["model-revision"]="v2.0.5";
                s_itn_path="";
                model_path[ITN_DIR]="";
                s_lm_path="";
                model_path[LM_DIR]="";
                s_punc_path="";
                model_path[PUNC_DIR]="";
            }

            if (use_gpu_){
                model_type = "torchscript";
                if (s_blade=="true" || s_blade=="True" || s_blade=="TRUE"){

 runtime/websocket/bin/websocket-server.cpp

@@ -67,7 +67,9 @@
                                 bool itn,
                                 int audio_fs,
                                 std::string wav_format,
                                 FUNASR_DEC_HANDLE& decoder_handle) {
                                 FUNASR_DEC_HANDLE& decoder_handle,
                                 std::string svs_lang,
                                 bool sys_itn) {
  try {
    int num_samples = buffer.size();  // the size of the buf

@@ -78,7 +80,8 @@
      try{
        FUNASR_RESULT Result = FunOfflineInferBuffer(
            asr_handle, buffer.data(), buffer.size(), RASR_NONE, nullptr, 
            hotwords_embedding, audio_fs, wav_format, itn, decoder_handle);
            hotwords_embedding, audio_fs, wav_format, itn, decoder_handle,
            svs_lang, sys_itn);
        if (Result != nullptr){
          asr_result = FunASRGetResult(Result, 0);  // get decode result
          stamp_res = FunASRGetStamp(Result);
@@ -162,6 +165,8 @@
  data_msg->msg["audio_fs"] = 16000; // default is 16k
  data_msg->msg["access_num"] = 0; // the number of access for this object, when it is 0, we can free it saftly
  data_msg->msg["is_eof"]=false;
  data_msg->msg["svs_lang"]="auto";
  data_msg->msg["svs_itn"]=true;
  FUNASR_DEC_HANDLE decoder_handle =
    FunASRWfstDecoderInit(asr_handle, ASR_OFFLINE, global_beam_, lattice_beam_, am_scale_);
  data_msg->decoder_handle = decoder_handle;
@@ -357,6 +362,12 @@
      if (jsonresult.contains("itn")) {
        msg_data->msg["itn"] = jsonresult["itn"];
      }
      if (jsonresult.contains("svs_lang")) {
        msg_data->msg["svs_lang"] = jsonresult["svs_lang"];
      }
      if (jsonresult.contains("svs_itn")) {
        msg_data->msg["svs_itn"] = jsonresult["svs_itn"];
      }
      if ((jsonresult["is_speaking"] == false ||
          jsonresult["is_finished"] == true) && 
          msg_data->msg["is_eof"] != true && 
@@ -375,7 +386,9 @@
                              msg_data->msg["itn"],
                              msg_data->msg["audio_fs"],
                              msg_data->msg["wav_format"],
                              std::ref(msg_data->decoder_handle)));
                              std::ref(msg_data->decoder_handle),
                              msg_data->msg["svs_lang"],
                              msg_data->msg["svs_itn"]));
        msg_data->msg["access_num"]=(int)(msg_data->msg["access_num"])+1;
      }
      break;

 runtime/websocket/bin/websocket-server.h

@@ -122,7 +122,9 @@
                  bool itn,
                  int audio_fs,
                  std::string wav_format,
                  FUNASR_DEC_HANDLE& decoder_handle);
                  FUNASR_DEC_HANDLE& decoder_handle,
                  std::string svs_lang,
                  bool sys_itn);

  void initAsr(std::map<std::string, std::string>& model_path, int thread_num, bool use_gpu=false, int batch_size=1);
  void on_message(websocketpp::connection_hdl hdl, message_ptr msg);