Skip to content

(model/cosyvoice): support flush api #124

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
xiong-binbin merged 1 commit into from
May 12, 2026
Merged

(model/cosyvoice): support flush api #124

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
59 changes: 46 additions & 13 deletions dashscope/audio/tts_v2/speech_synthesizer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -151,7 +151,7 @@ def __init__( # pylint: disable=redefined-builtin
instruction=None,
language_hints: list = None,
):
self.task_id = self.genUid()
self.task_id = self.gen_uid()
self.apikey = apikey
self.voice = voice
self.model = model
Expand All @@ -166,11 +166,11 @@ def __init__( # pylint: disable=redefined-builtin
self.instruction = instruction
self.language_hints = language_hints

def genUid(self):
def gen_uid(self):
Comment thread
songguocola marked this conversation as resolved.
# 生成随机UUID
return uuid.uuid4().hex

def getWebsocketHeaders(self, headers, workspace):
def get_websocket_headers(self, headers, workspace):
ua = get_user_agent()
self.headers = {
"user-agent": ua,
Expand All @@ -185,7 +185,7 @@ def getWebsocketHeaders(self, headers, workspace):
}
return self.headers

def getStartRequest(self, additional_params=None):
def get_start_request(self, additional_params=None):
cmd = {
HEADER: {
ACTION_KEY: ActionType.START,
Expand Down Expand Up @@ -223,7 +223,7 @@ def getStartRequest(self, additional_params=None):
] = self.language_hints
return json.dumps(cmd)

def getContinueRequest(self, text):
def get_continue_request(self, text):
cmd = {
HEADER: {
ACTION_KEY: ActionType.CONTINUE,
Expand All @@ -242,7 +242,7 @@ def getContinueRequest(self, text):
}
return json.dumps(cmd)

def getFinishRequest(self):
def get_finish_request(self):
cmd = {
HEADER: {
ACTION_KEY: ActionType.FINISHED,
Expand All @@ -255,6 +255,25 @@ def getFinishRequest(self):
}
return json.dumps(cmd)

def get_flush_request(self):
Comment thread
songguocola marked this conversation as resolved.
cmd = {
HEADER: {
ACTION_KEY: ActionType.CONTINUE, # CONTINUE task
TASK_ID: self.task_id,
"streaming": WebsocketStreamingMode.DUPLEX,
},
"payload": {
"model": self.model,
"task_group": "audio",
"task": "tts",
"function": "SpeechSynthesizer",
"input": {
"flush": True,
},
},
}
return json.dumps(cmd)


class SpeechSynthesizer:
def __init__( # pylint: disable=redefined-builtin
Expand Down Expand Up @@ -373,7 +392,7 @@ def __connect(self, timeout_seconds=5) -> None:
"""
self.ws = websocket.WebSocketApp(
self.url,
header=self.request.getWebsocketHeaders(
header=self.request.get_websocket_headers(
headers=self.headers,
workspace=self.workspace,
),
Expand Down Expand Up @@ -525,7 +544,7 @@ def __start_stream(self):
if self.ws is None:
self.__connect(5)
# 发送run-task指令
request = self.request.getStartRequest(self.additional_params)
request = self.request.get_start_request(self.additional_params)
self.__send_str(request)
if not self.start_event.wait(10):
raise TimeoutError("start speech synthesizer failed within 5s.")
Expand All @@ -539,13 +558,14 @@ def __submit_text(self, text):

if self._stopped.is_set():
raise InvalidTask("speech synthesizer task has stopped.")
request = self.request.getContinueRequest(text)
request = self.request.get_continue_request(text)
self.__send_str(request)

# pylint: disable=useless-return
def streaming_call(self, text: str):
"""
Streaming input mode: You can call the stream_call function multiple times to send text. # noqa: E501 # pylint: disable=line-too-long
Streaming input mode:
You can call the streaming_call function multiple times to send text.
A session will be created on the first call.
The session ends after calling streaming_complete.
Parameters:
Expand All @@ -559,6 +579,19 @@ def streaming_call(self, text: str):
self.__submit_text(text)
return None

def streaming_flush(self):
"""
send tts flush request.
"""
Comment thread
songguocola marked this conversation as resolved.
if not self._is_started:
raise InvalidTask("speech synthesizer has not been started.")

if self._stopped.is_set():
raise InvalidTask("speech synthesizer task has stopped.")
request = self.request.get_flush_request()
self.__send_str(request)
return None

def streaming_complete(self, complete_timeout_millis=600000):
"""
Synchronously stop the streaming input speech synthesis task.
Expand All @@ -575,7 +608,7 @@ def streaming_complete(self, complete_timeout_millis=600000):
raise InvalidTask("speech synthesizer has not been started.")
if self._stopped.is_set():
raise InvalidTask("speech synthesizer task has stopped.")
request = self.request.getFinishRequest()
request = self.request.get_finish_request()
self.__send_str(request)
if complete_timeout_millis is not None and complete_timeout_millis > 0:
if not self.complete_event.wait(
Expand Down Expand Up @@ -623,7 +656,7 @@ def async_streaming_complete(self, complete_timeout_millis=600000):
raise InvalidTask("speech synthesizer has not been started.")
if self._stopped.is_set():
raise InvalidTask("speech synthesizer task has stopped.")
request = self.request.getFinishRequest()
request = self.request.get_finish_request()
self.__send_str(request)
thread = threading.Thread(
target=self.__waiting_for_complete,
Expand All @@ -641,7 +674,7 @@ def streaming_cancel(self):
raise InvalidTask("speech synthesizer has not been started.")
if self._stopped.is_set():
return
request = self.request.getFinishRequest()
request = self.request.get_finish_request()
self.__send_str(request)
self.ws.close()
self.start_event.set()
Expand Down
Loading