dashscope 1.20.1__py3-none-any.whl → 1.20.2__py3-none-any.whl

dashscope/aigc/image_synthesis.py

@@ -127,7 +127,7 @@ class ImageSynthesis(BaseAsyncApi):
         if sketch_image_url is not None and sketch_image_url:
             input['sketch_image_url'] = sketch_image_url
         if ref_img is not None and ref_img:
-            input['ref_img'] == ref_img
+            input['ref_img'] = ref_img
         if extra_input is not None and extra_input:
             input = {**input, **extra_input}
 

dashscope/audio/asr/recognition.py

@@ -1,5 +1,6 @@
 import json
 import os
+import time
 import threading
 from http import HTTPStatus
 from threading import Timer
@@ -157,6 +158,8 @@ class Recognition(BaseApi):
         self._silence_timer = None
         self._kwargs = kwargs
         self._workspace = workspace
+        self._start_stream_timestamp = -1
+        self._first_package_timestamp = -1
 
     def __del__(self):
         if self._running:
@@ -183,12 +186,18 @@ class Recognition(BaseApi):
                 else:
                     usage: Dict[str, Any] = None
                     useags: List[Any] = None
-                    if 'sentence' in part.output and part.usage is not None:
-                        usage = {
-                            'end_time': part.output['sentence']['end_time'],
-                            'usage': part.usage
-                        }
-                        useags = [usage]
+                    if 'sentence' 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))
+                        sentence = part.output['sentence']
+                        logger.debug('Recv Result :{}, isEnd: {}'.format(sentence, RecognitionResult.is_sentence_end(sentence)))                        
+                        if part.usage is not None:                        
+                            usage = {
+                                'end_time': part.output['sentence']['end_time'],
+                                'usage': part.usage
+                            }
+                            useags = [usage]
 
                     self._callback.on_event(
                         RecognitionResult(
@@ -260,6 +269,8 @@ class Recognition(BaseApi):
         if self._running:
             raise InvalidParameter('Speech recognition has started.')
 
+        self._start_stream_timestamp = -1
+        self._first_package_timestamp = -1
         self._phrase = phrase_id
         self._kwargs.update(**kwargs)
         self._recognition_once = False
@@ -307,6 +318,7 @@ class Recognition(BaseApi):
         Returns:
             RecognitionResult: The result of speech recognition.
         """
+        self._start_stream_timestamp = time.time()*1000
         if self._running:
             raise InvalidParameter('Speech recognition has been called.')
 
@@ -350,7 +362,11 @@ class Recognition(BaseApi):
             for part in responses:
                 if part.status_code == HTTPStatus.OK:
                     if 'sentence' 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))                        
                         sentence = part.output['sentence']
+                        logger.debug('Recv Result :{}, isEnd: {}'.format(sentence, RecognitionResult.is_sentence_end(sentence)))
                         if RecognitionResult.is_sentence_end(sentence):
                             sentences.append(sentence)
 
@@ -408,6 +424,9 @@ class Recognition(BaseApi):
         if self._running is False:
             raise InvalidParameter('Speech recognition 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 = self._stream_data + [buffer]
 
     def _tidy_kwargs(self):

dashscope/audio/tts_v2/speech_synthesizer.py

@@ -5,9 +5,8 @@ import time
 import uuid
 from enum import Enum, unique
 
-import websocket
-
 import dashscope
+import websocket
 from dashscope.common.error import InputRequired, InvalidTask, ModelRequired
 from dashscope.common.logging import logger
 from dashscope.protocol.websocket import (ACTION_KEY, EVENT_KEY, HEADER,
@@ -239,6 +238,14 @@ class SpeechSynthesizer:
         self.headers = headers
         self.workspace = workspace
         self.additional_params = additional_params
+        self.model = model
+        self.voice = voice
+        self.aformat = format.format
+        if (self.aformat == 'DEFAULT'):
+            self.aformat = 'mp3'
+        self.sample_rate = format.sample_rate
+        if (self.sample_rate == 0):
+            self.sample_rate = 22050
 
         self.request = Request(
             apikey=self.apikey,
@@ -265,12 +272,18 @@ class SpeechSynthesizer:
         # since dashscope sdk will send first text in run-task
         if not self.callback:
             self.async_call = False
+        self._start_stream_timestamp = -1
+        self._first_package_timestamp = -1
+        self._recv_audio_length = 0
 
     def __send_str(self, data: str):
         logger.debug('>>>send {}'.format(data))
         self.ws.send(data)
 
     def __start_stream(self, ):
+        self._start_stream_timestamp = time.time() * 1000
+        self._first_package_timestamp = -1
+        self._recv_audio_length = 0
         if self.callback is None:
             raise InputRequired('callback is required!')
         # reset inner params
@@ -332,7 +345,7 @@ class SpeechSynthesizer:
         self.__submit_text(text)
         return None
 
-    def streaming_complete(self, complete_timeout_millie=10000):
+    def streaming_complete(self, complete_timeout_millis=10000):
         """
         Synchronously stop the streaming input speech synthesis task.
         Wait for all remaining synthesized audio before returning
@@ -340,7 +353,9 @@ class SpeechSynthesizer:
         Parameters:
         -----------
         complete_timeout_millis: int
-            Throws TimeoutError exception if it times out.
+            Throws TimeoutError exception if it times out. If the timeout is not None
+            and greater than zero, it will wait for the corresponding number of
+            milliseconds; otherwise, it will wait indefinitely.
         """
         if not self._is_started:
             raise InvalidTask('speech synthesizer has not been started.')
@@ -348,27 +363,38 @@ class SpeechSynthesizer:
             raise InvalidTask('speech synthesizer task has stopped.')
         request = self.request.getFinishRequest()
         self.__send_str(request)
-        if not self.complete_event.wait(timeout=complete_timeout_millie):
-            raise TimeoutError(
-                'speech synthesizer wait for complete timeout {}ms'.format(
-                    complete_timeout_millie))
+        if complete_timeout_millis is not None and complete_timeout_millis > 0:
+            if not self.complete_event.wait(timeout=complete_timeout_millis /
+                                            1000):
+                raise TimeoutError(
+                    'speech synthesizer wait for complete timeout {}ms'.format(
+                        complete_timeout_millis))
+        else:
+            self.complete_event.wait()
         self.close()
         self._stopped.set()
         self._is_started = False
 
-    def __waiting_for_complete(self):
-        if not self.complete_event.wait(timeout=10000):
+    def __waiting_for_complete(self, timeout):
+        if not self.complete_event.wait(timeout=timeout / 1000):
             raise TimeoutError(
                 'speech synthesizer wait for complete timeout 10000ms')
         self.close()
         self._stopped.set()
         self._is_started = False
 
-    def async_streaming_complete(self):
+    def async_streaming_complete(self, complete_timeout_millis=10000):
         """
         Asynchronously stop the streaming input speech synthesis task, returns immediately.
         You need to listen and handle the STREAM_INPUT_TTS_EVENT_SYNTHESIS_COMPLETE event in the on_event callback.
         Do not destroy the object and callback before this event.
+
+        Parameters:
+        -----------
+        complete_timeout_millis: int
+            Throws TimeoutError exception if it times out. If the timeout is not None
+            and greater than zero, it will wait for the corresponding number of
+            milliseconds; otherwise, it will wait indefinitely.
         """
 
         if not self._is_started:
@@ -377,7 +403,8 @@ class SpeechSynthesizer:
             raise InvalidTask('speech synthesizer task has stopped.')
         request = self.request.getFinishRequest()
         self.__send_str(request)
-        thread = threading.Thread(target=self.__waiting_for_complete)
+        thread = threading.Thread(target=self.__waiting_for_complete,
+                                  args=(complete_timeout_millis, ))
         thread.start()
 
     def streaming_cancel(self):
@@ -428,6 +455,18 @@ class SpeechSynthesizer:
         elif isinstance(message, (bytes, bytearray)):
             # 如果失败，认为是二进制消息
             logger.debug('<<<recv binary {}'.format(len(message)))
+            if (self._recv_audio_length == 0):
+                self._first_package_timestamp = time.time() * 1000
+                logger.debug('first package delay {}'.format(
+                    self._first_package_timestamp -
+                    self._start_stream_timestamp))
+            self._recv_audio_length += len(message) / (2 * self.sample_rate /
+                                                       1000)
+            current = time.time() * 1000
+            current_rtf = (current - self._first_package_timestamp
+                           ) / self._recv_audio_length
+            logger.debug('total audio {} ms, current_rtf: {}'.format(
+                self._recv_audio_length, current_rtf))
             # 只有在非异步调用的时候保存音频
             if not self.async_call:
                 if self._audio_data is None:
@@ -437,7 +476,7 @@ class SpeechSynthesizer:
             if self.callback:
                 self.callback.on_data(message)
 
-    def call(self, text: str):
+    def call(self, text: str, timeout_millis=None):
         """
         Speech synthesis.
         If callback is set, the audio will be returned in real-time through the on_event interface.
@@ -447,9 +486,14 @@ class SpeechSynthesizer:
         -----------
         text: str
             utf-8 encoded text
+        timeoutMillis:
+            Integer or None
         return: bytes
             If a callback is not set during initialization, the complete audio is returned
             as the function's return value. Otherwise, the return value is null.
+            If the timeout is set to a value greater than zero and not None,
+            it will wait for the corresponding number of milliseconds;
+            otherwise, it will wait indefinitely.
         """
         # print('还不支持非流式语音合成sdk调用大模型，使用流式模拟')
         if not self.callback:
@@ -457,10 +501,10 @@ class SpeechSynthesizer:
         self.__start_stream()
         self.__submit_text(text)
         if self.async_call:
-            self.async_streaming_complete()
+            self.async_streaming_complete(timeout_millis)
             return None
         else:
-            self.streaming_complete()
+            self.streaming_complete(timeout_millis)
             return self._audio_data
 
     # WebSocket关闭的回调函数
@@ -478,4 +522,4 @@ class SpeechSynthesizer:
 
     # 获取上一个任务的taskId
     def get_last_request_id(self):
-        return self.last_request_id
+        return self.last_request_id

dashscope/version.py

@@ -1 +1 @@
-__version__ = '1.20.1'
+__version__ = '1.20.2'

{dashscope-1.20.1.dist-info → dashscope-1.20.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: dashscope
-Version: 1.20.1
+Version: 1.20.2
 Summary: dashscope client sdk library
 Home-page: https://dashscope.aliyun.com/
 Author: Alibaba Cloud

{dashscope-1.20.1.dist-info → dashscope-1.20.2.dist-info}/RECORD

@@ -6,12 +6,12 @@ dashscope/files.py,sha256=QgJjwhtn9F548nCA8jD8OvE6aQEj-20hZqJgYXsUdQU,3930
 dashscope/finetune.py,sha256=_tflDUvu0KagSoCzLaf0hofpG_P8NU6PylL8CPjVhrA,6243
 dashscope/model.py,sha256=UPOn1qMYFhX-ovXi3BMxZEBk8qOK7WLJOYHMbPZwYBo,1440
 dashscope/models.py,sha256=1-bc-Ue68zurgu_y6RhfFr9uzeQMF5AZq-C32lJGMGU,1224
-dashscope/version.py,sha256=50ZydhgZIM_dwezjbX2voC6nDV_wx5kxvUuz3Nl2Eig,23
+dashscope/version.py,sha256=6KIFuOk_uHObRUtO_hf7gfIcH-9norzWjEkjzb1Dclg,23
 dashscope/aigc/__init__.py,sha256=s-MCA87KYiVumYtKtJi5IMN7xelSF6TqEU3s3_7RF-Y,327
 dashscope/aigc/code_generation.py,sha256=KAJVrGp6tiNFBBg64Ovs9RfcP5SrIhrbW3wdA89NKso,10885
 dashscope/aigc/conversation.py,sha256=xRoJlCR-IXHjSdkDrK74a9ut1FJg0FZhTNXZAJC18MA,14231
 dashscope/aigc/generation.py,sha256=53oMCmN5ZbqeqAsKxmdunXlRh-XP8ZtnA7hB2id4Koo,17897
-dashscope/aigc/image_synthesis.py,sha256=Hg2r6H7Vj4MsXwm62lHf2lTpUb6nA3xWGEYX2o-2TLQ,10419
+dashscope/aigc/image_synthesis.py,sha256=_TGkh4L_yBNQNmoJUKlfXfljKfxX-SSyxHHQBDm1AC8,10418
 dashscope/aigc/multimodal_conversation.py,sha256=SlNnnsUPV19gdx8fYJAtsMFWPNGY6vhk5IGHZ5ZczpI,5369
 dashscope/api_entities/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 dashscope/api_entities/aiohttp_request.py,sha256=aE3AeWba8Ig_xHMYjrAdkq0N61l_L2VFTG6HYh912X0,10229
@@ -31,12 +31,12 @@ dashscope/assistants/files.py,sha256=pwLVJ_pjpRFWyfI_MRvhH7Si7FzGDj4ChzZgWTpLOhg
 dashscope/audio/__init__.py,sha256=-ZRxrK-gV4QsUtlThIT-XwqB6vmyEsnhxIxdLmhCUuc,61
 dashscope/audio/asr/__init__.py,sha256=-s180qWn_JPSpCo1q0aDJJ5HQ3zTzD4z5yUwsRqH4aU,275
 dashscope/audio/asr/asr_phrase_manager.py,sha256=EjtbI3zz9UQGS1qv6Yb4zzEMj4OJJVXmwkqZyIrzvEA,7642
-dashscope/audio/asr/recognition.py,sha256=F2iz6hyXg16Z6DGlPwGpKfRNcAZIIsqXnNPtaZp4Fzo,17369
+dashscope/audio/asr/recognition.py,sha256=cEooE3wGf8kKfJIVbaXEytl5X6F0hMsLe8g4Bj9Fn4w,18768
 dashscope/audio/asr/transcription.py,sha256=1WAg9WH89antVzRYEKXb5LQP9xylZmX4YKp7v5oMYjY,8931
 dashscope/audio/tts/__init__.py,sha256=fbnieZX9yNFNh5BsxLpLXb63jlxzxrdCJakV3ignjlQ,194
 dashscope/audio/tts/speech_synthesizer.py,sha256=dnKx9FDDdO_ETHAjhK8zaMVaH6SfoTtN5YxXXqgY1JA,7571
 dashscope/audio/tts_v2/__init__.py,sha256=ve5a81qTbWDcRaSuritZtJBzryOIol2_dxzfqqdCw-k,345
-dashscope/audio/tts_v2/speech_synthesizer.py,sha256=sv5f4vi17rFcYDif5bSS4UuX_2eVUL2q5rTnXK1EoAg,16650
+dashscope/audio/tts_v2/speech_synthesizer.py,sha256=RenOvxp6BOs_-E_K5akMEBFw9G0HXoLnvHw1IOTUgeU,18991
 dashscope/client/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 dashscope/client/base_api.py,sha256=rXN97XGyDhCCaD_dz_clpFDjOJfpGjqiH7yX3LaD-GE,41233
 dashscope/common/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -82,9 +82,9 @@ dashscope/tokenizers/tokenizer.py,sha256=y6P91qTCYo__pEx_0VHAcj9YECfbUdRqZU1fdGT
 dashscope/tokenizers/tokenizer_base.py,sha256=REDhzRyDT13iequ61-a6_KcTy0GFKlihQve5HkyoyRs,656
 dashscope/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 dashscope/utils/oss_utils.py,sha256=fi8-PPsN-iR-iv5k2NS5Z8nlWkpgUhr56FRWm4BDh4A,6984
-dashscope-1.20.1.dist-info/LICENSE,sha256=Izp5L1DF1Mbza6qojkqNNWlE_mYLnr4rmzx2EBF8YFw,11413
-dashscope-1.20.1.dist-info/METADATA,sha256=pcY6Qv8KtqGlNAiqRr0JEfAzXBH8N56KlUD3o3expoA,6641
-dashscope-1.20.1.dist-info/WHEEL,sha256=G16H4A3IeoQmnOrYV4ueZGKSjhipXx8zc8nu9FGlvMA,92
-dashscope-1.20.1.dist-info/entry_points.txt,sha256=raEp5dOuj8whJ7yqZlDM8WQ5p2RfnGrGNo0QLQEnatY,50
-dashscope-1.20.1.dist-info/top_level.txt,sha256=woqavFJK9zas5xTqynmALqOtlafghjsk63Xk86powTU,10
-dashscope-1.20.1.dist-info/RECORD,,
+dashscope-1.20.2.dist-info/LICENSE,sha256=Izp5L1DF1Mbza6qojkqNNWlE_mYLnr4rmzx2EBF8YFw,11413
+dashscope-1.20.2.dist-info/METADATA,sha256=I52DKGYRGq77FO4pFbQwUHN-2g4qLLNz-jMO0DMQ008,6641
+dashscope-1.20.2.dist-info/WHEEL,sha256=G16H4A3IeoQmnOrYV4ueZGKSjhipXx8zc8nu9FGlvMA,92
+dashscope-1.20.2.dist-info/entry_points.txt,sha256=raEp5dOuj8whJ7yqZlDM8WQ5p2RfnGrGNo0QLQEnatY,50
+dashscope-1.20.2.dist-info/top_level.txt,sha256=woqavFJK9zas5xTqynmALqOtlafghjsk63Xk86powTU,10
+dashscope-1.20.2.dist-info/RECORD,,