Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.
Python asraudiochineseemotion-recognitionfunasrmcp-servermultilingual-asropenai-compatible-apiparaformerpunctuationpytorchreal-time-asrspeaker-diarizationspeech-recognitionspeech-to-textstreaming-asrtranscriptionvllmvoice-activity-detectionwhisper-alternative
fix(runtime): close online 2pass sentence state
dio724a08585523
diff --git a/runtime/websocket/bin/websocket-server-2pass.cpp b/runtime/websocket/bin/websocket-server-2pass.cppindex 45c016c19..ddfe91ce0 100644--- a/runtime/websocket/bin/websocket-server-2pass.cpp+++ b/runtime/websocket/bin/websocket-server-2pass.cpp@@ -65,7 +65,7 @@ context_ptr WebSocketServer::on_tls_init(tls_mode mode,return ctx;}-nlohmann::json handle_result(FUNASR_RESULT result, websocketpp::connection_hdl& hdl, std::map<websocketpp::connection_hdl, std::shared_ptr<FUNASR_MESSAGE>,std::owner_less<websocketpp::connection_hdl>>& data_map) {+nlohmann::json handle_result(FUNASR_RESULT result, websocketpp::connection_hdl& hdl, std::map<websocketpp::connection_hdl, std::shared_ptr<FUNASR_MESSAGE>,std::owner_less<websocketpp::connection_hdl>>& data_map, const std::string& modetype, bool is_final) {std::shared_ptr<FUNASR_MESSAGE> data_msg = nullptr;auto it = data_map.find(hdl);if (it != data_map.end()) {@@ -76,8 +76,9 @@ nlohmann::json handle_result(FUNASR_RESULT result, websocketpp::connection_hdl&nlohmann::json jsonresult;jsonresult["text"] = "";- std::string tmp_online_msg = FunASRGetResult(result, 0);- if (tmp_online_msg != "") {+ const char* online_result = FunASRGetResult(result, 0);+ if (online_result != nullptr && online_result[0] != '\0') {+ std::string tmp_online_msg = online_result;LOG(INFO) << "wav_name: " << data_msg->msg["wav_name"].get<std::string>() << " | online_res :" << tmp_online_msg;jsonresult["text"] = tmp_online_msg;jsonresult["mode"] = "2pass-online";@@ -95,18 +96,26 @@ nlohmann::json handle_result(FUNASR_RESULT result, websocketpp::connection_hdl&data_msg->end_time = FunASRGetTpassEnd(result); // 记录句子的结束时间jsonresult["timestamp"] = data_msg->timestamp;- std::string tmp_tpass_msg = FunASRGetTpassResult(result, 0);- if (tmp_tpass_msg != "") {- LOG(INFO) << "wav_name: " << data_msg->msg["wav_name"].get<std::string>() << " | offline results : " << tmp_tpass_msg;- jsonresult["text"] = tmp_tpass_msg;- jsonresult["mode"] = "2pass-offline";-- // 句子结束,记录结束时间- jsonresult["start_time"] = data_msg->start_time;- jsonresult["end_time"] = data_msg->end_time;- jsonresult["slice_type"] = 2;- jsonresult["index"] = data_msg->index;-+ if (modetype != "online") {+ const char* tpass_result = FunASRGetTpassResult(result, 0);+ if (tpass_result != nullptr && tpass_result[0] != '\0') {+ std::string tmp_tpass_msg = tpass_result;+ LOG(INFO) << "wav_name: " << data_msg->msg["wav_name"].get<std::string>() << " | offline results : " << tmp_tpass_msg;+ jsonresult["text"] = tmp_tpass_msg;+ jsonresult["mode"] = "2pass-offline";++ // 句子结束,记录结束时间+ jsonresult["start_time"] = data_msg->start_time;+ jsonresult["end_time"] = data_msg->end_time;+ jsonresult["slice_type"] = 2;+ jsonresult["index"] = data_msg->index;++ data_msg->index++; //句子序号+ data_msg->is_sentence_started = false; // 重置句子状态+ }+ } else if (is_final && data_msg->is_sentence_started) {+ // online mode must not send 2pass-offline text, but final chunks still close+ // the sentence state so the next utterance starts with slice_type=0.data_msg->index++; //句子序号data_msg->is_sentence_started = false; // 重置句子状态}@@ -195,7 +204,7 @@ void WebSocketServer::do_decoder(}if (Result) {websocketpp::lib::error_code ec;- nlohmann::json jsonresult = handle_result(Result, hdl, data_map);+ nlohmann::json jsonresult = handle_result(Result, hdl, data_map, modetype, false);jsonresult["wav_name"] = wav_name;jsonresult["is_final"] = false;if (jsonresult["text"] != "") {@@ -238,7 +247,7 @@ void WebSocketServer::do_decoder(}if (Result) {websocketpp::lib::error_code ec;- nlohmann::json jsonresult = handle_result(Result, hdl, data_map);+ nlohmann::json jsonresult = handle_result(Result, hdl, data_map, modetype, true);jsonresult["wav_name"] = wav_name;jsonresult["is_final"] = true;//LOG(INFO) << "jsonresult: " << jsonresult.dump(4);