dashscope 1.24.3__py3-none-any.whl → 1.24.5__py3-none-any.whl

dashscope/__init__.py

@@ -24,7 +24,7 @@ from dashscope.embeddings.batch_text_embedding_response import \
     BatchTextEmbeddingResponse
 from dashscope.embeddings.multimodal_embedding import (
     MultiModalEmbedding, MultiModalEmbeddingItemAudio,
-    MultiModalEmbeddingItemImage, MultiModalEmbeddingItemText)
+    MultiModalEmbeddingItemImage, MultiModalEmbeddingItemText, AioMultiModalEmbedding)
 from dashscope.embeddings.text_embedding import TextEmbedding
 from dashscope.files import Files
 from dashscope.models import Models
@@ -55,6 +55,7 @@ __all__ = [
     Models,
     TextEmbedding,
     MultiModalEmbedding,
+    AioMultiModalEmbedding,
     MultiModalEmbeddingItemAudio,
     MultiModalEmbeddingItemImage,
     MultiModalEmbeddingItemText,

dashscope/aigc/__init__.py

@@ -1,18 +1,20 @@
 # Copyright (c) Alibaba, Inc. and its affiliates.
-
 from .conversation import Conversation, History, HistoryItem
-from .generation import Generation
-from .image_synthesis import ImageSynthesis
+from .generation import Generation, AioGeneration
+from .image_synthesis import ImageSynthesis, AioImageSynthesis
 from .multimodal_conversation import MultiModalConversation, AioMultiModalConversation
-from .video_synthesis import VideoSynthesis
+from .video_synthesis import VideoSynthesis, AioVideoSynthesis
 
 __all__ = [
     Generation,
+    AioGeneration,
     Conversation,
     HistoryItem,
     History,
     ImageSynthesis,
+    AioImageSynthesis,
     MultiModalConversation,
     AioMultiModalConversation,
     VideoSynthesis,
+    AioVideoSynthesis,
 ]

dashscope/aigc/multimodal_conversation.py

@@ -24,9 +24,10 @@ class MultiModalConversation(BaseApi):
     def call(
         cls,
         model: str,
-        messages: List,
+        messages: List = None,
         api_key: str = None,
         workspace: str = None,
+        text: str = None,
         **kwargs
     ) -> Union[MultiModalConversationResponse, Generator[
             MultiModalConversationResponse, None, None]]:
@@ -55,6 +56,7 @@ class MultiModalConversation(BaseApi):
                 if None, will retrieve by rule [1].
                 [1]: https://help.aliyun.com/zh/dashscope/developer-reference/api-key-settings. # noqa E501
             workspace (str): The dashscope workspace id.
+            text (str): The text to generate.
             **kwargs:
                 stream(bool, `optional`): Enable server-sent events
                     (ref: https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events/Using_server-sent_events)  # noqa E501
@@ -68,8 +70,11 @@ class MultiModalConversation(BaseApi):
                     tokens with top_p probability mass. So 0.1 means only
                     the tokens comprising the top 10% probability mass are
                     considered[qwen-turbo,bailian-v1].
+                voice(string, `optional`): The voice name of qwen tts, include 'Cherry'/'Ethan'/'Sunny'/'Dylan' and so on,
+                    you can get the total voice list : https://help.aliyun.com/zh/model-studio/qwen-tts.
                 top_k(float, `optional`):
 
+
         Raises:
             InvalidInput: The history and auto_history are mutually exclusive.
 
@@ -78,18 +83,24 @@ class MultiModalConversation(BaseApi):
                   Generator[MultiModalConversationResponse, None, None]]: If
             stream is True, return Generator, otherwise MultiModalConversationResponse.
         """
-        if (messages is None or not messages):
-            raise InputRequired('prompt or messages is required!')
         if model is None or not model:
             raise ModelRequired('Model is required!')
         task_group, _ = _get_task_group_and_task(__name__)
-        msg_copy = copy.deepcopy(messages)
-        has_upload = cls._preprocess_messages(model, msg_copy, api_key)
-        if has_upload:
-            headers = kwargs.pop('headers', {})
-            headers['X-DashScope-OssResourceResolve'] = 'enable'
-            kwargs['headers'] = headers
-        input = {'messages': msg_copy}
+        input = {}
+        msg_copy = None
+
+        if messages is not None and messages:
+            msg_copy = copy.deepcopy(messages)
+            has_upload = cls._preprocess_messages(model, msg_copy, api_key)
+            if has_upload:
+                headers = kwargs.pop('headers', {})
+                headers['X-DashScope-OssResourceResolve'] = 'enable'
+                kwargs['headers'] = headers
+
+        if text is not None and text:
+            input.update({'text': text})
+        if msg_copy is not None:
+            input.update({'messages': msg_copy})
         response = super().call(model=model,
                                 task_group=task_group,
                                 task=MultiModalConversation.task,
@@ -145,9 +156,10 @@ class AioMultiModalConversation(BaseAioApi):
     async def call(
         cls,
         model: str,
-        messages: List,
+        messages: List = None,
         api_key: str = None,
         workspace: str = None,
+        text: str = None,
         **kwargs
     ) -> Union[MultiModalConversationResponse, Generator[
             MultiModalConversationResponse, None, None]]:
@@ -176,6 +188,7 @@ class AioMultiModalConversation(BaseAioApi):
                 if None, will retrieve by rule [1].
                 [1]: https://help.aliyun.com/zh/dashscope/developer-reference/api-key-settings. # noqa E501
             workspace (str): The dashscope workspace id.
+            text (str): The text to generate.
             **kwargs:
                 stream(bool, `optional`): Enable server-sent events
                     (ref: https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events/Using_server-sent_events)  # noqa E501
@@ -189,6 +202,8 @@ class AioMultiModalConversation(BaseAioApi):
                     tokens with top_p probability mass. So 0.1 means only
                     the tokens comprising the top 10% probability mass are
                     considered[qwen-turbo,bailian-v1].
+                voice(string, `optional`): The voice name of qwen tts, include 'Cherry'/'Ethan'/'Sunny'/'Dylan' and so on,
+                    you can get the total voice list : https://help.aliyun.com/zh/model-studio/qwen-tts.
                 top_k(float, `optional`):
 
         Raises:
@@ -199,18 +214,24 @@ class AioMultiModalConversation(BaseAioApi):
                   Generator[MultiModalConversationResponse, None, None]]: If
             stream is True, return Generator, otherwise MultiModalConversationResponse.
         """
-        if (messages is None or not messages):
-            raise InputRequired('prompt or messages is required!')
         if model is None or not model:
             raise ModelRequired('Model is required!')
         task_group, _ = _get_task_group_and_task(__name__)
-        msg_copy = copy.deepcopy(messages)
-        has_upload = cls._preprocess_messages(model, msg_copy, api_key)
-        if has_upload:
-            headers = kwargs.pop('headers', {})
-            headers['X-DashScope-OssResourceResolve'] = 'enable'
-            kwargs['headers'] = headers
-        input = {'messages': msg_copy}
+        input = {}
+        msg_copy = None
+
+        if messages is not None and messages:
+            msg_copy = copy.deepcopy(messages)
+            has_upload = cls._preprocess_messages(model, msg_copy, api_key)
+            if has_upload:
+                headers = kwargs.pop('headers', {})
+                headers['X-DashScope-OssResourceResolve'] = 'enable'
+                kwargs['headers'] = headers
+
+        if text is not None and text:
+            input.update({'text': text})
+        if msg_copy is not None:
+            input.update({'messages': msg_copy})
         response = await super().call(model=model,
                                       task_group=task_group,
                                       task=AioMultiModalConversation.task,

dashscope/api_entities/dashscope_response.py

@@ -152,6 +152,26 @@ class Choice(DictMixin):
                          **kwargs)
 
 
+@dataclass(init=False)
+class Audio(DictMixin):
+    data: str
+    url: str
+    id: str
+    expires_at: int
+
+    def __init__(self,
+                 data: str = None,
+                 url: str = None,
+                 id: str = None,
+                 expires_at: int = None,
+                 **kwargs):
+        super().__init__(data=data,
+                         url=url,
+                         id=id,
+                         expires_at=expires_at,
+                         **kwargs)
+
+
 @dataclass(init=False)
 class GenerationOutput(DictMixin):
     text: str
@@ -217,20 +237,25 @@ class GenerationResponse(DashScopeAPIResponse):
 @dataclass(init=False)
 class MultiModalConversationOutput(DictMixin):
     choices: List[Choice]
+    audio: Audio
 
     def __init__(self,
                  text: str = None,
                  finish_reason: str = None,
                  choices: List[Choice] = None,
+                 audio: Audio = None,
                  **kwargs):
         chs = None
         if choices is not None:
             chs = []
             for choice in choices:
                 chs.append(Choice(**choice))
+        if audio is not None:
+            audio = Audio(**audio)
         super().__init__(text=text,
                          finish_reason=finish_reason,
                          choices=chs,
+                         audio=audio,
                          **kwargs)
 
 
@@ -238,15 +263,18 @@ class MultiModalConversationOutput(DictMixin):
 class MultiModalConversationUsage(DictMixin):
     input_tokens: int
     output_tokens: int
+    characters: int
 
     # TODO add image usage info.
 
     def __init__(self,
                  input_tokens: int = 0,
                  output_tokens: int = 0,
+                 characters: int = 0,
                  **kwargs):
         super().__init__(input_tokens=input_tokens,
                          output_tokens=output_tokens,
+                         characters=characters,
                          **kwargs)
 
 
@@ -378,7 +406,7 @@ class RecognitionResponse(DashScopeAPIResponse):
         """
         result = False
         if sentence is not None and 'end_time' in sentence and sentence[
-                'end_time'] is not None:
+            'end_time'] is not None:
             result = True
         return result
 
@@ -445,8 +473,8 @@ class ImageSynthesisOutput(DictMixin):
     results: List[ImageSynthesisResult]
 
     def __init__(self,
-                 task_id: str =  None,
-                 task_status: str =  None,
+                 task_id: str = None,
+                 task_status: str = None,
                  results: List[ImageSynthesisResult] = [],
                  **kwargs):
         res = []

dashscope/audio/tts_v2/speech_synthesizer.py

@@ -98,6 +98,10 @@ class Request:
         volume=50,
         speech_rate=1.0,
         pitch_rate=1.0,
+        seed=0,
+        synthesis_type=0,
+        instruction=None,
+        language_hints: list = None,
     ):
         self.task_id = self.genUid()
         self.apikey = apikey
@@ -109,6 +113,10 @@ class Request:
         self.volume = volume
         self.speech_rate = speech_rate
         self.pitch_rate = pitch_rate
+        self.seed = seed
+        self.synthesis_type = synthesis_type
+        self.instruction = instruction
+        self.language_hints = language_hints
 
     def genUid(self):
         # 生成随机UUID
@@ -156,6 +164,8 @@ class Request:
                     'rate': self.speech_rate,
                     'format': self.format,
                     'pitch': self.pitch_rate,
+                    'seed': self.seed,
+                    'type': self.synthesis_type
                 },
             },
         }
@@ -163,6 +173,10 @@ class Request:
             cmd['payload']['parameters']['bit_rate'] = self.bit_rate
         if additional_params:
             cmd['payload']['parameters'].update(additional_params)
+        if self.instruction is not None:
+            cmd['payload']['parameters']['instruction'] = self.instruction
+        if self.language_hints is not None:
+            cmd['payload']['parameters']['language_hints'] = self.language_hints
         return json.dumps(cmd)
 
     def getContinueRequest(self, text):
@@ -207,6 +221,10 @@ class SpeechSynthesizer:
         volume=50,
         speech_rate=1.0,
         pitch_rate=1.0,
+        seed=0,
+        synthesis_type=0,
+        instruction=None,
+        language_hints: list = None,
         headers=None,
         callback: ResultCallback = None,
         workspace=None,
@@ -237,6 +255,14 @@ class SpeechSynthesizer:
             Dashscope workspace ID.
         url: str
             Dashscope WebSocket URL.
+        seed: int
+            The seed of the synthesizer, with a range from 0 to 65535. Default is 0.
+        synthesis_type: int
+            The type of the synthesizer, Default is 0.
+        instruction: str
+            The instruction of the synthesizer, max length is 128.
+        language_hints: list
+            The language hints of the synthesizer. supported language: zh, en.
         additional_params: Dict
             Additional parameters for the Dashscope API.
         """
@@ -271,6 +297,10 @@ class SpeechSynthesizer:
             volume=volume,
             speech_rate=speech_rate,
             pitch_rate=pitch_rate,
+            seed=seed,
+            synthesis_type=synthesis_type,
+            instruction=instruction,
+            language_hints=language_hints
         )
         self.last_request_id = self.request.task_id
         self.start_event = threading.Event()

dashscope/embeddings/multimodal_embedding.py

@@ -5,7 +5,7 @@ from typing import List
 
 from dashscope.api_entities.dashscope_response import (DashScopeAPIResponse,
                                                        DictMixin)
-from dashscope.client.base_api import BaseApi
+from dashscope.client.base_api import BaseApi, BaseAioApi
 from dashscope.common.error import InputRequired, ModelRequired
 from dashscope.common.utils import _get_task_group_and_task
 from dashscope.utils.oss_utils import preprocess_message_element
@@ -111,3 +111,72 @@ class MultiModalEmbedding(BaseApi):
                 if is_upload and not has_upload:
                     has_upload = True
         return has_upload
+
+
+class AioMultiModalEmbedding(BaseAioApi):
+    task = 'multimodal-embedding'
+
+    class Models:
+        multimodal_embedding_one_peace_v1 = 'multimodal-embedding-one-peace-v1'
+
+    @classmethod
+    async def call(cls,
+         model: str,
+         input: List[MultiModalEmbeddingItemBase],
+         api_key: str = None,
+         workspace: str = None,
+         **kwargs) -> DashScopeAPIResponse:
+        """Get embedding multimodal contents..
+
+        Args:
+            model (str): The embedding model name.
+            input (List[MultiModalEmbeddingElement]): The embedding elements,
+                every element include data, modal, factor field.
+            workspace (str): The dashscope workspace id.
+            **kwargs:
+                auto_truncation(bool, `optional`): Automatically truncate
+                audio longer than 15 seconds or text longer than 70 words.
+                Default to false(Too long input will result in failure).
+
+        Returns:
+            DashScopeAPIResponse: The embedding result.
+        """
+        if input is None or not input:
+            raise InputRequired('prompt is required!')
+        if model is None or not model:
+            raise ModelRequired('Model is required!')
+        embedding_input = {}
+        has_upload = cls._preprocess_message_inputs(model, input, api_key)
+        if has_upload:
+            headers = kwargs.pop('headers', {})
+            headers['X-DashScope-OssResourceResolve'] = 'enable'
+            kwargs['headers'] = headers
+        embedding_input['contents'] = input
+        kwargs.pop('stream', False)  # not support streaming output.
+        task_group, function = _get_task_group_and_task(__name__)
+        response = await super().call(
+            model=model,
+            input=embedding_input,
+            task_group=task_group,
+            task=MultiModalEmbedding.task,
+            function=function,
+            api_key=api_key,
+            workspace=workspace,
+            **kwargs)
+        return response
+
+    @classmethod
+    def _preprocess_message_inputs(cls, model: str, input: List[dict],
+                                   api_key: str):
+        """preprocess following inputs
+        input = [{'factor': 1, 'text': 'hello'},
+                {'factor': 2, 'audio': ''},
+                {'factor': 3, 'image': ''}]
+        """
+        has_upload = False
+        for elem in input:
+            if not isinstance(elem, (int, float, bool, str, bytes, bytearray)):
+                is_upload = preprocess_message_element(model, elem, api_key)
+                if is_upload and not has_upload:
+                    has_upload = True
+        return has_upload

dashscope/multimodal/__init__.py

@@ -0,0 +1,20 @@
+# Copyright (c) Alibaba, Inc. and its affiliates.
+
+from .tingwu import tingwu
+from .tingwu.tingwu import TingWu
+from .tingwu.tingwu_realtime import TingWuRealtime, TingWuRealtimeCallback
+
+from .multimodal_dialog import MultiModalDialog, MultiModalCallback
+from .dialog_state import DialogState
+from .multimodal_constants import *
+from .multimodal_request_params import *
+
+__all__ = [
+    'tingwu',
+    'TingWu',
+    'TingWuRealtime',
+    'TingWuRealtimeCallback',
+    'MultiModalDialog',
+    'MultiModalCallback',
+    'DialogState'
+]

dashscope/multimodal/tingwu/__init__.py

@@ -0,0 +1,10 @@
+# Copyright (c) Alibaba, Inc. and its affiliates.
+
+from .tingwu import TingWu
+from .tingwu_realtime import TingWuRealtime, TingWuRealtimeCallback
+
+__all__ = [
+    'TingWu',
+    'TingWuRealtime',
+    'TingWuRealtimeCallback'
+]

dashscope/multimodal/tingwu/tingwu_realtime.py

@@ -0,0 +1,579 @@
+# Copyright (c) Alibaba, Inc. and its affiliates.
+
+import json
+import platform
+import threading
+import time
+import uuid
+from dataclasses import dataclass, field
+from queue import Queue
+import dashscope
+from dashscope.client.base_api import BaseApi
+from dashscope.common.error import (InvalidParameter, ModelRequired)
+import websocket
+
+from dashscope.common.logging import logger
+from dashscope.protocol.websocket import ActionType
+
+
+class TingWuRealtimeCallback:
+    """An interface that defines callback methods for getting TingWu results.
+       Derive from this class and implement its function to provide your own data.
+    """
+
+    def on_open(self) -> None:
+        pass
+
+    def on_started(self, task_id: str) -> None:
+        pass
+
+    def on_speech_listen(self, result: dict):
+        pass
+
+    def on_recognize_result(self, result: dict):
+        pass
+
+    def on_ai_result(self, result: dict):
+        pass
+
+    def on_stopped(self) -> None:
+        pass
+
+    def on_error(self, error_code: str, error_msg: str) -> None:
+        pass
+
+    def on_close(self, close_status_code, close_msg):
+        """
+        callback when websocket connection is closed
+
+        :param close_status_code
+        :param close_msg
+        """
+        pass
+
+
+class TingWuRealtime(BaseApi):
+    """TingWuRealtime interface.
+
+    Args:
+        model (str): The requested model_id.
+        callback (TingWuRealtimeCallback): A callback that returns
+            speech recognition results.
+        app_id (str): The dashscope tingwu app id.
+        format (str): The input audio format for TingWu request.
+        sample_rate (int): The input audio sample rate.
+        terminology (str): The correct instruction set id.
+        workspace (str): The dashscope workspace id.
+
+        **kwargs:
+            max_end_silence (int): The maximum end silence time.
+            other_params (dict, `optional`): Other parameters.
+
+    Raises:
+        InputRequired: Input is required.
+    """
+
+    SILENCE_TIMEOUT_S = 60
+
+    def __init__(self,
+                 model: str,
+                 callback: TingWuRealtimeCallback,
+                 audio_format: str = "pcm",
+                 sample_rate: int = 16000,
+                 max_end_silence: int = None,
+                 app_id: str = None,
+                 terminology: str = None,
+                 workspace: str = None,
+                 api_key: str = None,
+                 base_address: str = None,
+                 data_id: str = None,
+                 **kwargs):
+        if api_key is None:
+            self.api_key = dashscope.api_key
+        else:
+            self.api_key = api_key
+        if base_address is None:
+            self.base_address = dashscope.base_websocket_api_url
+        else:
+            self.base_address = base_address
+
+        if model is None:
+            raise ModelRequired('Model is required!')
+
+        self.data_id = data_id
+        self.max_end_silence = max_end_silence
+        self.model = model
+        self.audio_format = audio_format
+        self.app_id = app_id
+        self.terminology = terminology
+        self.sample_rate = sample_rate
+        # 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.request = _Request()
+        self.response = _TingWuResponse(self._callback, self.close)  # 传递 self.close 作为回调
+
+    def _on_message(self, ws, message):
+        logger.debug(f"<<<<<<< Received message: {message}")
+        if isinstance(message, str):
+            self.response.handle_text_response(message)
+        elif isinstance(message, (bytes, bytearray)):
+            self.response.handle_binary_response(message)
+
+    def _on_error(self, ws, error):
+        logger.error(f"Error: {error}")
+        if self._callback:
+            error_code = ""  # 默认错误码
+            if "connection" in str(error).lower():
+                error_code = "1001"  # 连接错误
+            elif "timeout" in str(error).lower():
+                error_code = "1002"  # 超时错误
+            elif "authentication" in str(error).lower():
+                error_code = "1003"  # 认证错误
+            self._callback.on_error(error_code=error_code, error_msg=str(error))
+
+    def _on_close(self, ws, close_status_code, close_msg):
+        try:
+            logger.debug(
+                "WebSocket connection closed with status {} and message {}".format(close_status_code, close_msg))
+            if close_status_code is None:
+                close_status_code = 1000
+            if close_msg is None:
+                close_msg = "websocket is closed"
+            self._callback.on_close(close_status_code, close_msg)
+        except Exception as e:
+            logger.error(f"Error: {e}")
+
+    def _on_open(self, ws):
+        self._callback.on_open()
+        self._running = True
+
+    # def _on_pong(self):
+    #     logger.debug("on pong")
+
+    def start(self, **kwargs):
+        """
+        interface for starting TingWu connection
+        """
+        assert self._callback is not None, 'Please set the callback to get the TingWu result.'  # noqa E501
+
+        if self._running:
+            raise InvalidParameter('TingWu client has started.')
+
+        # self._start_stream_timestamp = -1
+        # self._first_package_timestamp = -1
+        # self._stop_stream_timestamp = -1
+        # self._on_complete_timestamp = -1
+        if self._kwargs is not None and len(self._kwargs) != 0:
+            self._kwargs.update(**kwargs)
+
+        self._connect(self.api_key)
+        logger.debug("connected with server.")
+        self._send_start_request()
+
+    def send_audio_data(self, speech_data: bytes):
+        """send audio data to server"""
+        if self._running:
+            self.__send_binary_frame(speech_data)
+
+    def stop(self):
+        if self.ws is None or not self.ws.sock or not self.ws.sock.connected:
+            self._callback.on_close(1001, "websocket is not connected")
+            return
+        _send_speech_json = self.request.generate_stop_request("stop")
+        self._send_text_frame(_send_speech_json)
+
+    """inner class"""
+
+    def _send_start_request(self):
+        """send start request"""
+        _start_json = self.request.generate_start_request(
+            workspace_id=self._workspace,
+            direction_name="start",
+            app_id=self.app_id,
+            model=self.model,
+            audio_format=self.audio_format,
+            sample_rate=self.sample_rate,
+            terminology=self.terminology,
+            max_end_silence=self.max_end_silence,
+            data_id=self.data_id,
+            **self._kwargs
+        )
+        # send start request
+        self._send_text_frame(_start_json)
+
+    def _run_forever(self):
+        self.ws.run_forever(ping_interval=5, ping_timeout=4)
+
+    def _connect(self, api_key: str):
+        """init websocket connection"""
+        self.ws = websocket.WebSocketApp(self.base_address, header=self.request.get_websocket_header(api_key),
+                                         on_open=self._on_open,
+                                         on_message=self._on_message,
+                                         on_error=self._on_error,
+                                         on_close=self._on_close)
+        self.thread = threading.Thread(target=self._run_forever)
+        # 统一心跳机制配置
+        self.ws.ping_interval = 5
+        self.ws.ping_timeout = 4
+        self.thread.daemon = True
+        self.thread.start()
+
+        self._wait_for_connection()
+
+    def close(self):
+        if self.ws is None or not self.ws.sock or not self.ws.sock.connected:
+            return
+        self.ws.close()
+
+    def _wait_for_connection(self):
+        """wait for connection using event instead of busy waiting"""
+        timeout = 5
+        start_time = time.time()
+        while not (self.ws.sock and self.ws.sock.connected) and (time.time() - start_time) < timeout:
+            time.sleep(0.1)  # 短暂休眠，避免密集轮询
+
+    def _send_text_frame(self, text: str):
+        # 避免在日志中记录敏感信息，如API密钥等
+        # 只记录非敏感信息
+        if '"Authorization"' not in text:
+            logger.info('>>>>>> send text frame : %s' % text)
+        else:
+            logger.info('>>>>>> send text frame with authorization header')
+        self.ws.send(text, websocket.ABNF.OPCODE_TEXT)
+
+    def __send_binary_frame(self, binary: bytes):
+        # _log.info('send binary frame length: %d' % len(binary))
+        self.ws.send(binary, websocket.ABNF.OPCODE_BINARY)
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        self.cleanup()
+        return False
+
+    def cleanup(self):
+        """cleanup resources"""
+        try:
+            if self.ws:
+                self.ws.close()
+            if self.thread and self.thread.is_alive():
+                # 设置标志位通知线程退出
+                self.thread.join(timeout=2)
+            # 清除引用
+            self.ws = None
+            self.thread = None
+            self._callback = None
+            self.response = None
+        except Exception as e:
+            logger.error(f"Error in cleanup: {e}")
+
+    def send_audio_frame(self, buffer: bytes):
+        """Push audio to server
+
+        Raises:
+            InvalidParameter: Cannot send data to an uninitiated recognition.
+        """
+        if self._running is False:
+            raise InvalidParameter('TingWu client has stopped.')
+
+        if self._start_stream_timestamp < 0:
+            self._start_stream_timestamp = time.time() * 1000
+        logger.debug('send_audio_frame: {}'.format(len(buffer)))
+        self.__send_binary_frame(buffer)
+
+
+class _Request:
+    def __init__(self):
+        # websocket header
+        self.ws_headers = None
+        # request body for voice chat
+        self.header = None
+        self.payload = None
+        # params
+        self.task_id = None
+        self.app_id = None
+        self.workspace_id = None
+
+    def get_websocket_header(self, api_key):
+        ua = 'dashscope/%s; python/%s; platform/%s; processor/%s' % (
+            '1.18.0',  # dashscope version
+            platform.python_version(),
+            platform.platform(),
+            platform.processor(),
+        )
+        self.ws_headers = {
+            "User-Agent": ua,
+            "Authorization": f"bearer {api_key}",
+            "Accept": "application/json"
+        }
+        logger.info('websocket header: {}'.format(self.ws_headers))
+        return self.ws_headers
+
+    def generate_start_request(self, direction_name: str,
+                               app_id: str,
+                               model: str = None,
+                               workspace_id: str = None,
+                               audio_format: str = None,
+                               sample_rate: int = None,
+                               terminology: str = None,
+                               max_end_silence: int = None,
+                               data_id: str = None,
+                               **kwargs
+                               ) -> str:
+        """
+        build start request.
+        :param app_id: web console app id
+        :param direction_name:
+        :param workspace_id: web console workspace id
+        :param model: model name
+        :param audio_format: audio format
+        :param sample_rate: sample rate
+        :param terminology:
+        :param max_end_silence:
+        :param data_id:
+        :return:
+        Args:
+            :
+        """
+        self._get_dash_request_header(ActionType.START)
+        parameters = self._get_start_parameters(audio_format=audio_format, sample_rate=sample_rate,
+                                                max_end_silence=max_end_silence,
+                                                terminology=terminology,
+                                                **kwargs)
+        self._get_dash_request_payload(direction_name=direction_name, app_id=app_id, workspace_id=workspace_id,
+                                       model=model,
+                                       data_id=data_id,
+                                       request_params=parameters)
+
+        cmd = {
+            "header": self.header,
+            "payload": self.payload
+        }
+        return json.dumps(cmd)
+
+    @staticmethod
+    def _get_start_parameters(audio_format: str = None,
+                              sample_rate: int = None,
+                              terminology: str = None,
+                              max_end_silence: int = None,
+                              **kwargs):
+        """
+        build start request parameters inner.
+        :param kwargs: parameters
+        :return
+        """
+        parameters = {}
+        if audio_format is not None:
+            parameters['format'] = audio_format
+        if sample_rate is not None:
+            parameters['sampleRate'] = sample_rate
+        if terminology is not None:
+            parameters['terminology'] = terminology
+        if max_end_silence is not None:
+            parameters['maxEndSilence'] = max_end_silence
+        if kwargs is not None and len(kwargs) != 0:
+            parameters.update(kwargs)
+        return parameters
+
+    def generate_stop_request(self, direction_name: str) -> str:
+        """
+        build stop request.
+        :param direction_name
+        :return
+        """
+        self._get_dash_request_header(ActionType.FINISHED)
+        self._get_dash_request_payload(direction_name, self.app_id)
+
+        cmd = {
+            "header": self.header,
+            "payload": self.payload
+        }
+        return json.dumps(cmd)
+
+    def _get_dash_request_header(self, action: str):
+        """
+        :param action: ActionType ：run-task, continue-task, finish-task
+        """
+        if self.task_id is None:
+            self.task_id = get_random_uuid()
+        self.header = DashHeader(action=action, task_id=self.task_id).to_dict()
+
+    def _get_dash_request_payload(self, direction_name: str,
+                                  app_id: str,
+                                  workspace_id: str = None,
+                                  custom_input=None,
+                                  model: str = None,
+                                  data_id: str = None,
+                                  request_params=None,
+                                  ):
+        """
+        build start request payload inner.
+        :param direction_name: inner direction name
+        :param app_id: web console app id
+        :param request_params: start direction body parameters
+        :param custom_input: user custom input
+        :param data_id: data id
+        :param model: model name
+        """
+        if custom_input is not None:
+            input = custom_input
+        else:
+            input = RequestBodyInput(
+                workspace_id=workspace_id,
+                app_id=app_id,
+                directive=direction_name,
+                data_id=data_id
+            )
+
+        self.payload = DashPayload(
+            model=model,
+            input=input.to_dict(),
+            parameters=request_params
+        ).to_dict()
+
+
+class _TingWuResponse:
+    def __init__(self, callback: TingWuRealtimeCallback, close_callback=None):
+        super().__init__()
+        self.task_id = None  # 对话ID.
+        self._callback = callback
+        self._close_callback = close_callback  # 保存关闭回调函数
+
+    def handle_text_response(self, response_json: str):
+        """
+        handle text response.
+        :param response_json: json format response from server
+        """
+        logger.info("<<<<<< server response: %s" % response_json)
+        try:
+            # try to parse response as json
+            json_data = json.loads(response_json)
+            header = json_data.get('header', {})
+            if header.get('event') == 'task-failed':
+                logger.error('Server returned invalid message: %s' % response_json)
+                if self._callback:
+                    self._callback.on_error(error_code=header.get('error_code'),
+                                            error_msg=header.get('error_message'))
+                return
+            if header.get('event') == "task-started":
+                self._handle_started(header.get('task_id'))
+                return
+
+            payload = json_data.get('payload', {})
+            output = payload.get('output', {})
+            if output is not None:
+                action = output.get('action')
+                logger.info("Server response action: %s" % action)
+                self._handle_tingwu_agent_text_response(action=action, response_json=json_data)
+
+        except json.JSONDecodeError:
+            logger.error("Failed to parse message as JSON.")
+
+    def handle_binary_response(self, response_binary: bytes):
+        """
+        handle binary response.
+        :param response_binary: server response binary。
+        """
+        logger.info("<<<<<< server response binary length: %d" % len(response_binary))
+
+    def _handle_tingwu_agent_text_response(self, action: str, response_json: dict):
+        payload = response_json.get('payload', {})
+        output = payload.get('output', {})
+        if action == "task-failed":
+            self._callback.on_error(error_code=output.get('errorCode'),
+                                    error_msg=output.get('errorMessage'))
+        elif action == "speech-listen":
+            self._callback.on_speech_listen(response_json)
+        elif action == "recognize-result":
+            self._callback.on_recognize_result(response_json)
+        elif action == "ai-result":
+            self._callback.on_ai_result(response_json)
+        elif action == "speech-end":  # ai-result事件永远会先于speech-end事件
+            self._callback.on_stopped()
+            if self._close_callback is not None:
+                self._close_callback()
+        else:
+            logger.info("Unknown response name:" + action)
+
+    def _handle_started(self, task_id: str):
+        self.task_id = task_id
+        self._callback.on_started(self.task_id)
+
+
+def get_random_uuid() -> str:
+    """generate random uuid."""
+    return uuid.uuid4().hex
+
+
+@dataclass
+class RequestBodyInput():
+    app_id: str
+    directive: str
+    data_id: str = field(default=None)
+    workspace_id: str = field(default=None)
+
+    def to_dict(self):
+        body_input = {
+            "appId": self.app_id,
+            "directive": self.directive,
+        }
+        if self.workspace_id is not None:
+            body_input["workspace_id"] = self.workspace_id
+        if self.data_id is not None:
+            body_input["dataId"] = self.data_id
+        return body_input
+
+
+@dataclass
+class DashHeader:
+    action: str
+    task_id: str = field(default=get_random_uuid())
+    streaming: str = field(default="duplex")  # 默认为 duplex
+
+    def to_dict(self):
+        return {
+            "action": self.action,
+            "task_id": self.task_id,
+            "request_id": self.task_id,
+            "streaming": self.streaming
+        }
+
+
+@dataclass
+class DashPayload:
+    task_group: str = field(default="aigc")
+    function: str = field(default="generation")
+    model: str = field(default="")
+    task: str = field(default="multimodal-generation")
+    parameters: dict = field(default=None)
+    input: dict = field(default=None)
+
+    def to_dict(self):
+        payload = {
+            "task_group": self.task_group,
+            "function": self.function,
+            "model": self.model,
+            "task": self.task,
+        }
+
+        if self.parameters is not None:
+            payload["parameters"] = self.parameters
+
+        if self.input is not None:
+            payload["input"] = self.input
+
+        return payload

dashscope/version.py

@@ -1,3 +1,3 @@
 # Copyright (c) Alibaba, Inc. and its affiliates.
 
-__version__ = '1.24.3'
+__version__ = '1.24.5'

{dashscope-1.24.3.dist-info → dashscope-1.24.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dashscope
-Version: 1.24.3
+Version: 1.24.5
 Summary: dashscope client sdk library
 Home-page: https://dashscope.aliyun.com/
 Author: Alibaba Cloud

{dashscope-1.24.3.dist-info → dashscope-1.24.5.dist-info}/RECORD

@@ -1,16 +1,16 @@
-dashscope/__init__.py,sha256=MJI4PJmnevfQA_GA30_L3hmsq49hooZpg25k6w39dts,3120
+dashscope/__init__.py,sha256=96J137Im9Ii9uxfVOOYkZDJNZXF1sEbcH4-QXFr4xEw,3172
 dashscope/cli.py,sha256=64oGkevgX0RHPPmMg0sevXDgaFLQNA_0vdtjQ7Z2pHM,26492
 dashscope/files.py,sha256=vRDQygm3lOqBZR73o7KNHs1iTBVuvLncuwJNxIYjzAU,3981
 dashscope/model.py,sha256=B5v_BtYLPqj6raClejBgdKg6WTGwhH_f-20pvsQqmsk,1491
 dashscope/models.py,sha256=dE4mzXkl85G343qVylSGpURPRdA5pZSqXlx6PcxqC_Q,1275
-dashscope/version.py,sha256=V0YmAKw8G9I8NSRumZjDtPh52jnAd9OGCwh9X-oeC_M,74
-dashscope/aigc/__init__.py,sha256=m51CHEKL3WPq-s14OF-G1Uk3rLj6B6KrU55bbCKU-Ak,500
+dashscope/version.py,sha256=2fvqw7bZLyWOIDvUb8DEkdi6y_VgyljhOeYdITEksWM,74
+dashscope/aigc/__init__.py,sha256=kYvYEoRK-NUHyMWpBDNQBz4fVA__uOhHRK2kDTBaWgk,617
 dashscope/aigc/chat_completion.py,sha256=ONlyyssIbfaKKcFo7cEKhHx5OCF2XX810HFzIExW1ho,14813
 dashscope/aigc/code_generation.py,sha256=p_mxDKJLQMW0IjFD46JRlZuEZCRESSVKEfLlAevBtqw,10936
 dashscope/aigc/conversation.py,sha256=95xEEY4ThZJysj5zy3aMw7ql9KLJVfD_1iHv9QZ17Ew,14282
 dashscope/aigc/generation.py,sha256=xMcMu16rICTdjZiD_sPqYV_Ltdp4ewGzzfC7JD9VApY,17948
 dashscope/aigc/image_synthesis.py,sha256=Itx9h5brEwC-d3Mj_ntDHGd4qaitqDg9DeGHMJouhMk,28178
-dashscope/aigc/multimodal_conversation.py,sha256=Kjg8Gtfhl_Ok8WVwD-AeT-VBN9hh6E74TfkCxkL5wbY,10821
+dashscope/aigc/multimodal_conversation.py,sha256=BXpUthyGxJHQs18-m_ZzAw6MI5nSM4_NVMUfTDSC1-k,11682
 dashscope/aigc/video_synthesis.py,sha256=RSPjar5-YiF9xclRmf9H7-5QbRxLcsNXO4zS7oTKi2I,24137
 dashscope/api_entities/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 dashscope/api_entities/aiohttp_request.py,sha256=1L7XdIJ9L65cQmX8x9JCR4t5hNIMDrbiWADfKKp9yfo,10280
@@ -18,7 +18,7 @@ dashscope/api_entities/api_request_data.py,sha256=04rpYPNK1HkT3iTPJmZpquH621xcBb
 dashscope/api_entities/api_request_factory.py,sha256=ynpbFmxSne4dJkv5m40Vlwt4hJSxQPprAuUgMSQIQDg,5639
 dashscope/api_entities/base_request.py,sha256=W2SzrSAGFS6V8DErfSrayQtSL0T4iO7BrC8flr7nt1w,977
 dashscope/api_entities/chat_completion_types.py,sha256=1WMWPszhM3HaJBVz-ZXx-El4D8-RfVUL3ym65xsDRLk,11435
-dashscope/api_entities/dashscope_response.py,sha256=slc7o9jNS5yzv2giEPzz9CDOW6X797nkAocgZ1r84aU,22089
+dashscope/api_entities/dashscope_response.py,sha256=31guU41ePkLyFsVVN-1WODXdOHiURzRyxxhrUmX9dGM,22835
 dashscope/api_entities/encryption.py,sha256=rUCZx3wwVvS5oyKXEeWgyWPxM8Y5d4AaVdgxLhizBqA,5517
 dashscope/api_entities/http_request.py,sha256=MTxYsbkK8oYWDp8ZPjrkdY9YbnQ9SEIy87riyJidMXo,16484
 dashscope/api_entities/websocket_request.py,sha256=PS0FU854-HjTbKa68f4GHa7-noFRMzKySJGfPkrrBjw,16146
@@ -46,7 +46,7 @@ dashscope/audio/tts/__init__.py,sha256=xYpMFseUZGgqgj_70zcX2VsLv-L7qxJ3d-bbdj_hO
 dashscope/audio/tts/speech_synthesizer.py,sha256=vD1xQV-rew8qAsIaAGH5amsNtB0SqdtNhVHhJHGQ-xk,7622
 dashscope/audio/tts_v2/__init__.py,sha256=me9a3_7KsHQxcJ8hx4SeKlY1e_ThHVvGMw7Yn0uoscM,333
 dashscope/audio/tts_v2/enrollment.py,sha256=-nrlywYSOP73Bm9ETTSxNnlp-B8ezJcUmd59mVvyvgk,6361
-dashscope/audio/tts_v2/speech_synthesizer.py,sha256=jbTwybwJjhbgUa6TgeLUYnkDBkmk-tjzxta1FtYeWAk,20824
+dashscope/audio/tts_v2/speech_synthesizer.py,sha256=p764P4TYwLkvvPCpA4VnFwlNbIJbuNbp2d9mxgni7Ws,22047
 dashscope/client/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 dashscope/client/base_api.py,sha256=znAJ65DeHiFw1H7FWK0YrkLz1CoNcyqUxF8EJ3gujeY,52523
 dashscope/common/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -65,17 +65,18 @@ dashscope/customize/finetunes.py,sha256=AL_kGTJXMvM2ej-EKsLLd1dUphPQdVTefFVCSVH-
 dashscope/embeddings/__init__.py,sha256=XQ7vKr8oZM2CmdOduE53BWy6_Qpn9xUPkma64yw8Gws,291
 dashscope/embeddings/batch_text_embedding.py,sha256=lVhvTS8McYfXuqt_8CmmhA6bPqD0nrGv965kjYG_j0E,8842
 dashscope/embeddings/batch_text_embedding_response.py,sha256=ZfkJMUq8GRsFA6XUTsiAsIySqGJH-VPi2P9Ba1KTU-s,2056
-dashscope/embeddings/multimodal_embedding.py,sha256=NwjQsdkKgUz51ozGjqFDzVlLcZjY0m1JNdH1EyAY0a4,4109
+dashscope/embeddings/multimodal_embedding.py,sha256=PEF7DmtE5cbrXw4k3WQcfmsBKaAY3CTIoei3SyhOl34,6774
 dashscope/embeddings/text_embedding.py,sha256=2MPEyMB99xueDbvFg9kKAe8bgHMDEaFLaFa6GzDWDHg,2108
 dashscope/io/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 dashscope/io/input_output.py,sha256=0aXrRJFo1ZqYm_AJWR_w88O4-Btn9np2zUhrrUdBdfw,3992
-dashscope/multimodal/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+dashscope/multimodal/__init__.py,sha256=fyqeolbDLWVn5wSpPZ3nAOnUBRF9k6mlsy6dCmgjPvI,533
 dashscope/multimodal/dialog_state.py,sha256=CtOdfGWhq0ePG3bc8-7inhespETtPD4QDli1513hd1A,1522
 dashscope/multimodal/multimodal_constants.py,sha256=z_QVq01E43FAqKQnDu9vdf89d1zuYlWyANewWTEXVJM,1282
 dashscope/multimodal/multimodal_dialog.py,sha256=HymlaQYp7SgJdoKbT27SNiviyRRoM91zklNBwTHmm1Q,23939
 dashscope/multimodal/multimodal_request_params.py,sha256=Lbxf_kLnFUkhty8AU9wL7ws9tYbmhHPVmsiXLdynlJg,8402
-dashscope/multimodal/tingwu/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+dashscope/multimodal/tingwu/__init__.py,sha256=Gi9GEM0bdeJlZpvyksSeHOc2--_tG5aF6QAx6TAS2fE,225
 dashscope/multimodal/tingwu/tingwu.py,sha256=01d-QOeuB1QmRhiZqbXJ8pHoGqT0C-xZTjIs_ZBXOyw,2613
+dashscope/multimodal/tingwu/tingwu_realtime.py,sha256=oBeqrZit3uBZHuyI7m9VILz2qaqJRMO0-Nm2eJ5Q63g,20215
 dashscope/nlp/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 dashscope/nlp/understanding.py,sha256=00ado-ibYEzBRT0DgKGd3bohQDNW73xnFhJ_1aa87lw,2880
 dashscope/protocol/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -99,9 +100,9 @@ dashscope/tokenizers/tokenizer.py,sha256=3FQVDvMNkCW9ccYeJdjrd_PIMMD3Xv7aNZkaYOE
 dashscope/tokenizers/tokenizer_base.py,sha256=5EJIFuizMWESEmLmbd38yJnfeHmPnzZPwsO4aOGjpl4,707
 dashscope/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 dashscope/utils/oss_utils.py,sha256=aZIHlMN2JOfVw6kp0SVrMw_N1MfoTcR_-wiRbJ7DgHw,7501
-dashscope-1.24.3.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-dashscope-1.24.3.dist-info/METADATA,sha256=fLj2JRtUdQhu6TCj2GdodS2iFXLay7WWgMCcEq15jm4,7146
-dashscope-1.24.3.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-dashscope-1.24.3.dist-info/entry_points.txt,sha256=e9C3sOf9zDYL0O5ROEGX6FT8w-QK_kaGRWmPZDHAFys,49
-dashscope-1.24.3.dist-info/top_level.txt,sha256=woqavFJK9zas5xTqynmALqOtlafghjsk63Xk86powTU,10
-dashscope-1.24.3.dist-info/RECORD,,
+dashscope-1.24.5.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+dashscope-1.24.5.dist-info/METADATA,sha256=eRxoK1TphMD4hn-vUM3aPMtaCsvCzstnDe-QPH9A4Q0,7146
+dashscope-1.24.5.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+dashscope-1.24.5.dist-info/entry_points.txt,sha256=e9C3sOf9zDYL0O5ROEGX6FT8w-QK_kaGRWmPZDHAFys,49
+dashscope-1.24.5.dist-info/top_level.txt,sha256=woqavFJK9zas5xTqynmALqOtlafghjsk63Xk86powTU,10
+dashscope-1.24.5.dist-info/RECORD,,