dashscope 1.8.0__py3-none-any.whl → 1.25.6__py3-none-any.whl

dashscope/audio/asr/translation_recognizer.py

@@ -0,0 +1,1006 @@
+# Copyright (c) Alibaba, Inc. and its affiliates.
+
+import json
+import os
+import threading
+import time
+import uuid
+from http import HTTPStatus
+from queue import Queue
+from threading import Timer
+from typing import Any, Dict, List
+
+from dashscope.client.base_api import BaseApi
+from dashscope.common.constants import ApiProtocol
+from dashscope.common.error import (InputDataRequired, InputRequired,
+                                    InvalidParameter, InvalidTask,
+                                    ModelRequired)
+from dashscope.common.logging import logger
+from dashscope.common.utils import _get_task_group_and_task
+from dashscope.protocol.websocket import WebsocketStreamingMode
+
+DASHSCOPE_TRANSLATION_KEY = 'translations'
+DASHSCOPE_TRANSCRIPTION_KEY = 'transcription'
+
+
+class ThreadSafeBool:
+    def __init__(self, initial_value=False):
+        self._value = initial_value
+        self._lock = threading.Lock()
+
+    def set(self, value):
+        with self._lock:
+            self._value = value
+
+    def get(self):
+        with self._lock:
+            return self._value
+
+
+class WordObj():
+    def __init__(self, ) -> None:
+        self.text: str = None
+        self.begin_time: int = None
+        self.end_time: int = None
+        self.fixed: bool = False
+        self._raw_data = None
+
+    @staticmethod
+    def from_json(json_data: Dict[str, Any]):
+        """Create a Word object from a JSON dictionary.
+        """
+        word = WordObj()
+        word.text = json_data['text']
+        word.begin_time = json_data['begin_time']
+        word.end_time = json_data['end_time']
+        word.fixed = json_data['fixed']
+        word._raw_data = json_data
+        return word
+
+    def __str__(self) -> str:
+        return 'Word: ' + json.dumps(self._raw_data, ensure_ascii=False)
+
+    def __repr__(self):
+        return self.__str__()
+
+
+class SentenceBaseObj():
+    def __init__(self, ) -> None:
+        self.sentence_id: int = -1
+        self.text: str = None
+        self.begin_time: int = None
+        self.end_time: int = None
+        self.words: List[WordObj] = []
+        self._raw_data = None
+
+    @staticmethod
+    def from_json(json_data: Dict[str, Any]):
+        """Create a SentenceBase object from a JSON dictionary.
+        """
+        sentence = SentenceBaseObj()
+        sentence.sentence_id = json_data['sentence_id']
+        sentence.text = json_data['text']
+        sentence.begin_time = json_data['begin_time']
+        if json_data.get('end_time') is not None:
+            sentence.end_time = json_data['end_time']
+        else:
+            sentence.end_time = json_data['current_time']
+        sentence.words = [
+            WordObj.from_json(word) for word in json_data['words']
+        ]
+        sentence._raw_data = json_data
+        return sentence
+
+    def __str__(self) -> str:
+        return json.dumps(self._raw_data, ensure_ascii=False)
+
+    def __repr__(self):
+        return self.__str__()
+
+
+class TranscriptionResult(SentenceBaseObj):
+    def __init__(self, ) -> None:
+        self.stash: SentenceBaseObj = None
+        self.is_sentence_end = False
+        # vad related
+        self.vad_pre_end: bool = False
+        self.pre_end_failed: bool = False
+        self.pre_end_timemillis: int = -1
+        self.pre_end_start_time: int = -1
+        self.pre_end_end_time: int = -1
+        self._raw_data = None
+
+    @staticmethod
+    def from_json(json_data: Dict[str, Any]):
+        """Create a TranscriptionResult object from a JSON dictionary.
+        """
+        transcription = TranscriptionResult()
+        transcription.sentence_id = json_data['sentence_id']
+        transcription.text = json_data['text']
+        transcription.begin_time = json_data['begin_time']
+        if json_data.get('end_time') is not None:
+            transcription.end_time = json_data['end_time']
+        else:
+            transcription.end_time = json_data['current_time']
+        transcription.words = [
+            WordObj.from_json(word) for word in json_data['words']
+        ]
+        transcription._raw_data = json_data
+        transcription.is_sentence_end = json_data.get('sentence_end')
+        if 'stash' in json_data:
+            transcription.stash = SentenceBaseObj.from_json(json_data['stash'])
+        if 'vad_pre_end' in json_data:
+            transcription.vad_pre_end = json_data['vad_pre_end']
+        if 'pre_end_failed' in json_data:
+            transcription.pre_end_failed = json_data['pre_end_failed']
+        if 'pre_end_start_time' in json_data:
+            transcription.pre_end_start_time = json_data['pre_end_start_time']
+        if 'pre_end_end_time' in json_data:
+            transcription.pre_end_end_time = json_data['pre_end_end_time']
+        transcription._raw_data = json_data
+        return transcription
+
+    def __str__(self) -> str:
+        return 'Transcriptions: ' + json.dumps(self._raw_data,
+                                               ensure_ascii=False)
+
+    def __repr__(self):
+        return self.__str__()
+
+
+class Translation(SentenceBaseObj):
+    def __init__(self, ) -> None:
+        self.language: str = None
+        self.stash: SentenceBaseObj = None
+        self.is_sentence_end = False
+        # vad related
+        self.vad_pre_end: bool = False
+        self.pre_end_failed: bool = False
+        self.pre_end_timemillis: int = -1
+        self.pre_end_start_time: int = -1
+        self.pre_end_end_time: int = -1
+        self._raw_data = None
+
+    @staticmethod
+    def from_json(json_data: Dict[str, Any]):
+        """Create a Translation object from a JSON dictionary.
+        """
+        translation = Translation()
+        translation.sentence_id = json_data['sentence_id']
+        translation.text = json_data['text']
+        translation.begin_time = json_data['begin_time']
+        if json_data.get('end_time') is not None:
+            translation.end_time = json_data['end_time']
+        else:
+            translation.end_time = json_data['current_time']
+        translation.words = [
+            WordObj.from_json(word) for word in json_data['words']
+        ]
+        translation._raw_data = json_data
+
+        translation.language = json_data['lang']
+        translation.is_sentence_end = json_data.get('sentence_end')
+        if 'stash' in json_data:
+            translation.stash = SentenceBaseObj.from_json(json_data['stash'])
+        if 'vad_pre_end' in json_data:
+            translation.vad_pre_end = json_data['vad_pre_end']
+        if 'pre_end_failed' in json_data:
+            translation.pre_end_failed = json_data['pre_end_failed']
+        if 'pre_end_start_time' in json_data:
+            translation.pre_end_start_time = json_data['pre_end_start_time']
+        if 'pre_end_end_time' in json_data:
+            translation.pre_end_end_time = json_data['pre_end_end_time']
+        translation._raw_data = json_data
+        return translation
+
+    def __str__(self) -> str:
+        return 'Translation: ' + json.dumps(self._raw_data, ensure_ascii=False)
+
+    def __repr__(self):
+        return self.__str__()
+
+
+class TranslationResult():
+    def __init__(self, ) -> None:
+        self.translations: Dict[str:Translation] = {}
+        self.is_sentence_end = False
+        self._raw_data = None
+
+    def get_translation(self, language) -> Translation:
+        if self.translations is None:
+            return None
+        return self.translations.get(language)
+
+    def get_language_list(self, ) -> List[str]:
+        if self.translations is None:
+            return None
+        return list(self.translations.keys())
+
+    @staticmethod
+    def from_json(json_data: List):
+        """Create a TranslationResult object from a JSON dictionary.
+        """
+        result = TranslationResult()
+        result._raw_data = json_data
+        for translation_json in json_data:
+            if not isinstance(translation_json, dict):
+                raise InvalidParameter(
+                    f'Invalid translation json data: {translation_json}')
+            else:
+                translation = Translation.from_json(translation_json)
+                result.translations[translation.language] = translation
+                if translation.is_sentence_end:
+                    result.is_sentence_end = True
+        return result
+
+    def __str__(self) -> str:
+        return 'TranslationList: ' + json.dumps(self._raw_data,
+                                                ensure_ascii=False)
+
+    def __repr__(self):
+        return self.__str__()
+
+
+class TranslationRecognizerResultPack():
+    def __init__(self) -> None:
+        self.transcription_result_list: List[TranscriptionResult] = []
+        self.translation_result_list: List[TranslationResult] = []
+        self.usage_list: List = []
+        self.request_id: str = None
+        self.error_message = None
+
+
+class TranslationRecognizerCallback():
+    """An interface that defines callback methods for getting translation recognizer results. # noqa E501
+       Derive from this class and implement its function to provide your own data.
+    """
+    def on_open(self) -> None:
+        pass
+
+    def on_complete(self) -> None:
+        pass
+
+    def on_error(self, message) -> None:
+        pass
+
+    def on_close(self) -> None:
+        pass
+
+    def on_event(self, request_id, transcription_result: TranscriptionResult,
+                 translation_result: TranslationResult, usage) -> None:
+        pass
+
+
+class TranslationRecognizerRealtime(BaseApi):
+    """TranslationRecognizerRealtime interface.
+
+    Args:
+        model (str): The requested model_id.
+        callback (TranslationRecognizerRealtime): A callback that returns
+            TranslationRecognizerRealtime results.
+        format (str): The input audio format.
+        sample_rate (int): The input audio sample rate.
+        workspace (str): The dashscope workspace id.
+
+        **kwargs:
+            phrase_id (list, `optional`): The ID of phrase.
+            disfluency_removal_enabled(bool, `optional`): Filter mood words,
+                turned off by default.
+            diarization_enabled (bool, `optional`): Speech auto diarization,
+                turned off by default.
+            speaker_count (int, `optional`): The number of speakers.
+            timestamp_alignment_enabled (bool, `optional`): Timestamp-alignment
+                calibration, turned off by default.
+            special_word_filter(str, `optional`): Sensitive word filter.
+            audio_event_detection_enabled(bool, `optional`):
+                Audio event detection, turned off by default.
+
+    Raises:
+        InputRequired: Input is required.
+    """
+
+    SILENCE_TIMEOUT_S = 23
+
+    def __init__(self,
+                 model: str,
+                 callback: TranslationRecognizerCallback,
+                 format: str,
+                 sample_rate: int,
+                 transcription_enabled: bool = True,
+                 source_language: str = None,
+                 translation_enabled: bool = False,
+                 workspace: str = None,
+                 **kwargs):
+        if model is None:
+            raise ModelRequired('Model is required!')
+        if format is None:
+            raise InputRequired('format is required!')
+        if sample_rate is None:
+            raise InputRequired('sample_rate is required!')
+
+        self.model = model
+        self.format = format
+        self.sample_rate = sample_rate
+        self.source_language = source_language
+        self.transcription_enabled = transcription_enabled
+        self.translation_enabled = translation_enabled
+        # continuous recognition with start() or once recognition with call()
+        self._recognition_once = False
+        self._callback = callback
+        self._running = False
+        self._stream_data = Queue()
+        self._worker = None
+        self._silence_timer = None
+        self._kwargs = kwargs
+        self._workspace = workspace
+        self._start_stream_timestamp = -1
+        self._first_package_timestamp = -1
+        self._stop_stream_timestamp = -1
+        self._on_complete_timestamp = -1
+        self.request_id_confirmed = False
+        self.last_request_id = uuid.uuid4().hex
+
+    def __del__(self):
+        if self._running:
+            self._running = False
+            self._stream_data = Queue()
+            if self._worker is not None and self._worker.is_alive():
+                self._worker.join()
+            if self._silence_timer is not None and self._silence_timer.is_alive(  # noqa E501
+            ):
+                self._silence_timer.cancel()
+                self._silence_timer = None
+            if self._callback:
+                self._callback.on_close()
+
+    def __receive_worker(self):
+        """Asynchronously, initiate a real-time transltion recognizer request and
+           obtain the result for parsing.
+        """
+        responses = self.__launch_request()
+        for part in responses:
+            if part.status_code == HTTPStatus.OK:
+                logger.debug('Received response request_id: {} {}'.format(
+                    part.request_id, part.output))
+                if len(part.output) == 0:
+                    self._on_complete_timestamp = time.time() * 1000
+                    logger.debug('last package delay {}'.format(
+                        self.get_last_package_delay()))
+                    self._callback.on_complete()
+                else:
+                    usage = None
+                    transcription = None
+                    translations = None
+                    if DASHSCOPE_TRANSCRIPTION_KEY in part.output:
+                        transcription = TranscriptionResult.from_json(
+                            part.output[DASHSCOPE_TRANSCRIPTION_KEY])
+                    if DASHSCOPE_TRANSLATION_KEY in part.output:
+                        translations = TranslationResult.from_json(
+                            part.output[DASHSCOPE_TRANSLATION_KEY])
+                    if transcription is not None or translations is not None:
+                        if (self._first_package_timestamp < 0):
+                            self._first_package_timestamp = time.time() * 1000
+                            logger.debug('first package delay {}'.format(
+                                self.get_first_package_delay()))
+
+                    if part.usage is not None:
+                        usage = part.usage
+                    if self.request_id_confirmed is False and part.request_id is not None:
+                        self.last_request_id = part.request_id
+                        self.request_id_confirmed = True
+                    self._callback.on_event(part.request_id, transcription,
+                                            translations, usage)
+            else:
+                self._running = False
+                self._stream_data = Queue()
+                self._callback.on_error(part)
+                self._callback.on_close()
+                break
+
+    def __launch_request(self):
+        """Initiate real-time translation recognizer requests.
+        """
+
+        self._tidy_kwargs()
+        task_name, _ = _get_task_group_and_task(__name__)
+        responses = super().call(
+            model=self.model,
+            task_group='audio',
+            task=task_name,
+            function='recognition',
+            input=self._input_stream_cycle(),
+            api_protocol=ApiProtocol.WEBSOCKET,
+            ws_stream_mode=WebsocketStreamingMode.DUPLEX,
+            is_binary_input=True,
+            sample_rate=self.sample_rate,
+            format=self.format,
+            stream=True,
+            source_language=self.source_language,
+            transcription_enabled=self.transcription_enabled,
+            translation_enabled=self.translation_enabled,
+            workspace=self._workspace,
+            pre_task_id=self.last_request_id,
+            **self._kwargs)
+        return responses
+
+    def start(self, **kwargs):
+        """Real-time translation recognizer in asynchronous mode.
+           Please call 'stop()' after you have completed translation & recognition.
+
+        Args:
+            phrase_id (str, `optional`): The ID of phrase.
+
+            **kwargs:
+                disfluency_removal_enabled(bool, `optional`):
+                    Filter mood words, turned off by default.
+                diarization_enabled (bool, `optional`):
+                    Speech auto diarization, turned off by default.
+                speaker_count (int, `optional`): The number of speakers.
+                timestamp_alignment_enabled (bool, `optional`):
+                    Timestamp-alignment calibration, turned off by default.
+                special_word_filter(str, `optional`): Sensitive word filter.
+                audio_event_detection_enabled(bool, `optional`):
+                    Audio event detection, turned off by default.
+
+        Raises:
+            InvalidParameter: This interface cannot be called again
+                if it has already been started.
+            InvalidTask: Task create failed.
+        """
+        assert self._callback is not None, 'Please set the callback to get the translation & recognition result.'  # noqa E501
+
+        if self._running:
+            raise InvalidParameter(
+                'TranslationRecognizerRealtime has started.')
+
+        self._start_stream_timestamp = -1
+        self._first_package_timestamp = -1
+        self._stop_stream_timestamp = -1
+        self._on_complete_timestamp = -1
+        self._kwargs.update(**kwargs)
+        self._recognition_once = False
+        self._worker = threading.Thread(target=self.__receive_worker)
+        self._worker.start()
+        if self._worker.is_alive():
+            self._running = True
+            self._callback.on_open()
+
+            # If audio data is not received for 23 seconds, the timeout exits
+            self._silence_timer = Timer(
+                TranslationRecognizerRealtime.SILENCE_TIMEOUT_S,
+                self._silence_stop_timer)
+            self._silence_timer.start()
+        else:
+            self._running = False
+            raise InvalidTask('Invalid task, task create failed.')
+
+    def call(self,
+             file: str,
+             phrase_id: str = None,
+             **kwargs) -> TranslationRecognizerResultPack:
+        """TranslationRecognizerRealtime in synchronous mode.
+
+        Args:
+            file (str): The path to the local audio file.
+            phrase_id (str, `optional`): The ID of phrase.
+
+            **kwargs:
+                disfluency_removal_enabled(bool, `optional`):
+                    Filter mood words, turned off by default.
+                diarization_enabled (bool, `optional`):
+                    Speech auto diarization, turned off by default.
+                speaker_count (int, `optional`): The number of speakers.
+                timestamp_alignment_enabled (bool, `optional`):
+                    Timestamp-alignment calibration, turned off by default.
+                special_word_filter(str, `optional`): Sensitive word filter.
+                audio_event_detection_enabled(bool, `optional`):
+                    Audio event detection, turned off by default.
+
+        Raises:
+            InvalidParameter: This interface cannot be called again
+                if it has already been started.
+            InputDataRequired: The supplied file was empty.
+
+        Returns:
+            TranslationRecognizerResultPack: The result of speech translation & recognition.
+        """
+        self._start_stream_timestamp = time.time() * 1000
+        if self._running:
+            raise InvalidParameter(
+                'TranslationRecognizerRealtime has been called.')
+
+        if os.path.exists(file):
+            if os.path.isdir(file):
+                raise IsADirectoryError('Is a directory: ' + file)
+        else:
+            raise FileNotFoundError('No such file or directory: ' + file)
+
+        self._recognition_once = True
+        self._stream_data = Queue()
+        self._phrase = phrase_id
+        self._kwargs.update(**kwargs)
+        results = TranslationRecognizerResultPack()
+        error_message = None
+
+        try:
+            audio_data: bytes = None
+            f = open(file, 'rb')
+            if os.path.getsize(file):
+                while True:
+                    audio_data = f.read(12800)
+                    if not audio_data:
+                        break
+                    else:
+                        self._stream_data.put(audio_data)
+            else:
+                raise InputDataRequired(
+                    'The supplied file was empty (zero bytes long)')
+            f.close()
+            self._stop_stream_timestamp = time.time() * 1000
+        except Exception as e:
+            logger.error(e)
+            raise e
+
+        if not self._stream_data.empty():
+            self._running = True
+            responses = self.__launch_request()
+            for part in responses:
+                if part.status_code == HTTPStatus.OK:
+                    logger.debug('received data: {}'.format(part.output))
+                    # debug log cal fpd
+                    transcription = None
+                    translation = None
+                    usage = None
+                    if ('translation' in part.output) or ('transcription'
+                                                          in part.output):
+                        if (self._first_package_timestamp < 0):
+                            self._first_package_timestamp = time.time() * 1000
+                            logger.debug('first package delay {}'.format(
+                                self._first_package_timestamp -
+                                self._start_stream_timestamp))
+                        if part.usage is not None:
+                            usage = part.usage
+
+                    if DASHSCOPE_TRANSCRIPTION_KEY in part.output:
+                        transcription = TranscriptionResult.from_json(
+                            part.output[DASHSCOPE_TRANSCRIPTION_KEY])
+
+                    if DASHSCOPE_TRANSLATION_KEY in part.output:
+                        translation = TranslationResult.from_json(
+                            part.output[DASHSCOPE_TRANSLATION_KEY])
+
+                    if (transcription is not None
+                            and transcription.is_sentence_end) or (
+                                translation is not None
+                                and translation.is_sentence_end):
+                        results.request_id = part.request_id
+                        results.transcription_result_list.append(transcription)
+                        results.translation_result_list.append(translation)
+                        results.usage_list.append(usage)
+                else:
+                    error_message = part
+                    logger.error(error_message)
+                    break
+
+        self._on_complete_timestamp = time.time() * 1000
+        logger.debug('last package delay {}'.format(
+            self.get_last_package_delay()))
+
+        self._stream_data = Queue()
+        self._recognition_once = False
+        self._running = False
+        results.error_message = error_message
+        return results
+
+    def stop(self):
+        """End asynchronous TranslationRecognizerRealtime.
+
+        Raises:
+            InvalidParameter: Cannot stop an uninitiated TranslationRecognizerRealtime.
+        """
+        if self._running is False:
+            raise InvalidParameter(
+                'TranslationRecognizerRealtime has stopped.')
+
+        self._stop_stream_timestamp = time.time() * 1000
+
+        self._running = False
+        if self._worker is not None and self._worker.is_alive():
+            self._worker.join()
+        self._stream_data = Queue()
+        if self._silence_timer is not None and self._silence_timer.is_alive():
+            self._silence_timer.cancel()
+            self._silence_timer = None
+        if self._callback:
+            self._callback.on_close()
+
+    def send_audio_frame(self, buffer: bytes):
+        """Push audio to TranslationRecognizerRealtime.
+
+        Raises:
+            InvalidParameter: Cannot send data to an uninitiated TranslationRecognizerRealtime.
+        """
+        if self._running is False:
+            raise InvalidParameter(
+                'TranslationRecognizerRealtime has stopped.')
+
+        if (self._start_stream_timestamp < 0):
+            self._start_stream_timestamp = time.time() * 1000
+        logger.debug('send_audio_frame: {}'.format(len(buffer)))
+        self._stream_data.put(buffer)
+
+    def _tidy_kwargs(self):
+        for k in self._kwargs.copy():
+            if self._kwargs[k] is None:
+                self._kwargs.pop(k, None)
+
+    def _input_stream_cycle(self):
+        while self._running:
+            while self._stream_data.empty():
+                if self._running:
+                    time.sleep(0.01)
+                    continue
+                else:
+                    break
+
+            # Reset silence_timer when getting stream.
+            if self._silence_timer is not None and self._silence_timer.is_alive(  # noqa E501
+            ):
+                self._silence_timer.cancel()
+                self._silence_timer = Timer(
+                    TranslationRecognizerRealtime.SILENCE_TIMEOUT_S,
+                    self._silence_stop_timer)
+                self._silence_timer.start()
+
+            while not self._stream_data.empty():
+                frame = self._stream_data.get()
+                yield bytes(frame)
+
+            if self._recognition_once:
+                self._running = False
+
+        # drain all audio data when invoking stop().
+        if self._recognition_once is False:
+            while not self._stream_data.empty():
+                frame = self._stream_data.get()
+                yield bytes(frame)
+
+    def _silence_stop_timer(self):
+        """If audio data is not received for a long time, exit worker.
+        """
+        self._running = False
+        if self._silence_timer is not None and self._silence_timer.is_alive():
+            self._silence_timer.cancel()
+        self._silence_timer = None
+        if self._worker is not None and self._worker.is_alive():
+            self._worker.join()
+        self._stream_data = Queue()
+
+    def get_first_package_delay(self):
+        """First Package Delay is the time between start sending audio and receive first words package
+        """
+        return self._first_package_timestamp - self._start_stream_timestamp
+
+    def get_last_package_delay(self):
+        """Last Package Delay is the time between stop sending audio and receive last words package
+        """
+        return self._on_complete_timestamp - self._stop_stream_timestamp
+
+    # 获取上一个任务的taskId
+    def get_last_request_id(self):
+        return self.last_request_id
+
+
+class TranslationRecognizerChat(BaseApi):
+    """TranslationRecognizerChat interface.
+
+    Args:
+        model (str): The requested model_id.
+        callback (TranslationRecognizerChat): A callback that returns
+            TranslationRecognizerChat results.
+        format (str): The input audio format.
+        sample_rate (int): The input audio sample rate.
+        workspace (str): The dashscope workspace id.
+
+        **kwargs:
+            phrase_id (list, `optional`): The ID of phrase.
+            disfluency_removal_enabled(bool, `optional`): Filter mood words,
+                turned off by default.
+            diarization_enabled (bool, `optional`): Speech auto diarization,
+                turned off by default.
+            speaker_count (int, `optional`): The number of speakers.
+            timestamp_alignment_enabled (bool, `optional`): Timestamp-alignment
+                calibration, turned off by default.
+            special_word_filter(str, `optional`): Sensitive word filter.
+            audio_event_detection_enabled(bool, `optional`):
+                Audio event detection, turned off by default.
+
+    Raises:
+        InputRequired: Input is required.
+    """
+
+    SILENCE_TIMEOUT_S = 23
+
+    def __init__(self,
+                 model: str,
+                 callback: TranslationRecognizerCallback,
+                 format: str,
+                 sample_rate: int,
+                 transcription_enabled: bool = True,
+                 source_language: str = None,
+                 translation_enabled: bool = False,
+                 workspace: str = None,
+                 **kwargs):
+        if model is None:
+            raise ModelRequired('Model is required!')
+        if format is None:
+            raise InputRequired('format is required!')
+        if sample_rate is None:
+            raise InputRequired('sample_rate is required!')
+
+        self.model = model
+        self.format = format
+        self.sample_rate = sample_rate
+        self.source_language = source_language
+        self.transcription_enabled = transcription_enabled
+        self.translation_enabled = translation_enabled
+        # continuous recognition with start() or once recognition with call()
+        self._recognition_once = False
+        self._callback = callback
+        self._running = False
+        self._stream_data = Queue()
+        self._worker = None
+        self._silence_timer = None
+        self._kwargs = kwargs
+        self._workspace = workspace
+        self._start_stream_timestamp = -1
+        self._first_package_timestamp = -1
+        self._stop_stream_timestamp = -1
+        self._on_complete_timestamp = -1
+        self.request_id_confirmed = False
+        self.last_request_id = uuid.uuid4().hex
+        self._is_sentence_end = ThreadSafeBool(False)
+
+    def __del__(self):
+        if self._running:
+            self._running = False
+            self._stream_data = Queue()
+            if self._worker is not None and self._worker.is_alive():
+                self._worker.join()
+            if self._silence_timer is not None and self._silence_timer.is_alive(  # noqa E501
+            ):
+                self._silence_timer.cancel()
+                self._silence_timer = None
+            if self._callback:
+                self._callback.on_close()
+
+    def __receive_worker(self):
+        """Asynchronously, initiate a real-time transltion recognizer request and
+           obtain the result for parsing.
+        """
+        responses = self.__launch_request()
+        for part in responses:
+            if part.status_code == HTTPStatus.OK:
+                logger.debug('Received response request_id: {} {}'.format(
+                    part.request_id, part.output))
+                if len(part.output) == 0:
+                    self._on_complete_timestamp = time.time() * 1000
+                    logger.debug('last package delay {}'.format(
+                        self.get_last_package_delay()))
+                    self._callback.on_complete()
+                else:
+                    usage = None
+                    transcription = None
+                    translations = None
+                    if DASHSCOPE_TRANSCRIPTION_KEY in part.output:
+                        transcription = TranscriptionResult.from_json(
+                            part.output[DASHSCOPE_TRANSCRIPTION_KEY])
+                    if DASHSCOPE_TRANSLATION_KEY in part.output:
+                        translations = TranslationResult.from_json(
+                            part.output[DASHSCOPE_TRANSLATION_KEY])
+                    if transcription is not None or translations is not None:
+                        if (self._first_package_timestamp < 0):
+                            self._first_package_timestamp = time.time() * 1000
+                            logger.debug('first package delay {}'.format(
+                                self.get_first_package_delay()))
+
+                    if part.usage is not None:
+                        usage = part.usage
+                    if self.request_id_confirmed is False and part.request_id is not None:
+                        self.last_request_id = part.request_id
+                        self.request_id_confirmed = True
+                    if transcription is not None and transcription.is_sentence_end:
+                        logger.debug(
+                            '[Chat] recv sentence end in transcription, stop asr'
+                        )
+                        self._is_sentence_end.set(True)
+                    if translations is not None and translations.is_sentence_end:
+                        logger.debug(
+                            '[Chat] recv sentence end in translation, stop asr'
+                        )
+                        self._is_sentence_end.set(True)
+                    self._callback.on_event(part.request_id, transcription,
+                                            translations, usage)
+            else:
+                self._running = False
+                self._stream_data = Queue()
+                self._callback.on_error(part)
+                self._callback.on_close()
+                break
+
+    def __launch_request(self):
+        """Initiate real-time translation recognizer requests.
+        """
+
+        self._tidy_kwargs()
+        task_name, _ = _get_task_group_and_task(__name__)
+        responses = super().call(
+            model=self.model,
+            task_group='audio',
+            task=task_name,
+            function='recognition',
+            input=self._input_stream_cycle(),
+            api_protocol=ApiProtocol.WEBSOCKET,
+            ws_stream_mode=WebsocketStreamingMode.DUPLEX,
+            is_binary_input=True,
+            sample_rate=self.sample_rate,
+            format=self.format,
+            stream=True,
+            source_language=self.source_language,
+            transcription_enabled=self.transcription_enabled,
+            translation_enabled=self.translation_enabled,
+            workspace=self._workspace,
+            pre_task_id=self.last_request_id,
+            **self._kwargs)
+        return responses
+
+    def start(self, **kwargs):
+        """Real-time translation recognizer in asynchronous mode.
+           Please call 'stop()' after you have completed translation & recognition.
+
+        Args:
+            phrase_id (str, `optional`): The ID of phrase.
+
+            **kwargs:
+                disfluency_removal_enabled(bool, `optional`):
+                    Filter mood words, turned off by default.
+                diarization_enabled (bool, `optional`):
+                    Speech auto diarization, turned off by default.
+                speaker_count (int, `optional`): The number of speakers.
+                timestamp_alignment_enabled (bool, `optional`):
+                    Timestamp-alignment calibration, turned off by default.
+                special_word_filter(str, `optional`): Sensitive word filter.
+                audio_event_detection_enabled(bool, `optional`):
+                    Audio event detection, turned off by default.
+
+        Raises:
+            InvalidParameter: This interface cannot be called again
+                if it has already been started.
+            InvalidTask: Task create failed.
+        """
+        assert self._callback is not None, 'Please set the callback to get the translation & recognition result.'  # noqa E501
+
+        if self._running:
+            raise InvalidParameter('TranslationRecognizerChat has started.')
+
+        self._start_stream_timestamp = -1
+        self._first_package_timestamp = -1
+        self._stop_stream_timestamp = -1
+        self._on_complete_timestamp = -1
+        self._kwargs.update(**kwargs)
+        self._recognition_once = False
+        self._worker = threading.Thread(target=self.__receive_worker)
+        self._worker.start()
+        if self._worker.is_alive():
+            self._running = True
+            self._callback.on_open()
+
+            # If audio data is not received for 23 seconds, the timeout exits
+            self._silence_timer = Timer(
+                TranslationRecognizerChat.SILENCE_TIMEOUT_S,
+                self._silence_stop_timer)
+            self._silence_timer.start()
+        else:
+            self._running = False
+            raise InvalidTask('Invalid task, task create failed.')
+
+    def stop(self):
+        """End asynchronous TranslationRecognizerChat.
+
+        Raises:
+            InvalidParameter: Cannot stop an uninitiated TranslationRecognizerChat.
+        """
+        if self._running is False:
+            raise InvalidParameter('TranslationRecognizerChat has stopped.')
+
+        self._stop_stream_timestamp = time.time() * 1000
+        logger.debug('stop TranslationRecognizerChat')
+        self._running = False
+        if self._worker is not None and self._worker.is_alive():
+            self._worker.join()
+        self._stream_data = Queue()
+        if self._silence_timer is not None and self._silence_timer.is_alive():
+            self._silence_timer.cancel()
+            self._silence_timer = None
+        if self._callback:
+            self._callback.on_close()
+
+    def send_audio_frame(self, buffer: bytes) -> bool:
+        """Push audio to TranslationRecognizerChat.
+
+        Raises:
+            InvalidParameter: Cannot send data to an uninitiated TranslationRecognizerChat.
+        """
+        if self._is_sentence_end.get():
+            logger.debug('skip audio due to has sentence end.')
+            return False
+
+        if self._running is False:
+            raise InvalidParameter('TranslationRecognizerChat has stopped.')
+
+        if (self._start_stream_timestamp < 0):
+            self._start_stream_timestamp = time.time() * 1000
+        logger.debug('send_audio_frame: {}'.format(len(buffer)))
+        self._stream_data.put(buffer)
+        return True
+
+    def _tidy_kwargs(self):
+        for k in self._kwargs.copy():
+            if self._kwargs[k] is None:
+                self._kwargs.pop(k, None)
+
+    def _input_stream_cycle(self):
+        while self._running:
+            while self._stream_data.empty():
+                if self._running:
+                    time.sleep(0.01)
+                    continue
+                else:
+                    break
+
+            # Reset silence_timer when getting stream.
+            if self._silence_timer is not None and self._silence_timer.is_alive(  # noqa E501
+            ):
+                self._silence_timer.cancel()
+                self._silence_timer = Timer(
+                    TranslationRecognizerChat.SILENCE_TIMEOUT_S,
+                    self._silence_stop_timer)
+                self._silence_timer.start()
+
+            while not self._stream_data.empty():
+                frame = self._stream_data.get()
+                yield bytes(frame)
+
+            if self._recognition_once:
+                self._running = False
+
+        # drain all audio data when invoking stop().
+        if self._recognition_once is False:
+            while not self._stream_data.empty():
+                frame = self._stream_data.get()
+                yield bytes(frame)
+
+    def _silence_stop_timer(self):
+        """If audio data is not received for a long time, exit worker.
+        """
+        self._running = False
+        if self._silence_timer is not None and self._silence_timer.is_alive():
+            self._silence_timer.cancel()
+        self._silence_timer = None
+        if self._worker is not None and self._worker.is_alive():
+            self._worker.join()
+        self._stream_data = Queue()
+
+    def get_first_package_delay(self):
+        """First Package Delay is the time between start sending audio and receive first words package
+        """
+        return self._first_package_timestamp - self._start_stream_timestamp
+
+    def get_last_package_delay(self):
+        """Last Package Delay is the time between stop sending audio and receive last words package
+        """
+        return self._on_complete_timestamp - self._stop_stream_timestamp
+
+    # 获取上一个任务的taskId
+    def get_last_request_id(self):
+        return self.last_request_id