livekit-plugins-volcenginee 1.3.0__tar.gz → 1.3.2__tar.gz

{livekit_plugins_volcenginee-1.3.0 → livekit_plugins_volcenginee-1.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: livekit-plugins-volcenginee
-Version: 1.3.0
+Version: 1.3.2
 Summary: LiveKit Agent Plugins for Volcengine
 Author-email: linsanzhu <2287225730@qq.com>
 Keywords: audio,livekit,realtime,video,webrtc
@@ -38,6 +38,9 @@ Description-Content-Type: text/markdown
 - `volcengine.TTS` 默认音色为 `zh_female_xiaohe_uranus_bigtts`
 - `volcengine.STT` 现已直接使用豆包大模型流式语音识别接口
 - `volcengine.STT` 已补齐新版握手头 `X-Api-Connect-Id`
+- `volcengine.STT` 切换到新版控制台鉴权，仅使用 `X-Api-Key` 头
+- `volcengine.TTS` 同步切换到新版控制台鉴权，仅使用 `X-Api-Key` 头
+- `volcengine.RealtimeModel` 同步切换到新版控制台鉴权，仅使用 `X-Api-Key` 头
 
 ## ✨ 特性
 
@@ -95,20 +98,20 @@ pip install -e .
 
 | 环境变量 | 描述 | 获取方式 |
 |----------|------|----------|
-| `VOLCENGINE_TTS_ACCESS_TOKEN` | TTS 服务的访问令牌 | [语音合成控制台](https://console.volcengine.com/speech/service/16) |
-| `VOLCENGINE_STT_ACCESS_TOKEN` | STT 服务的访问令牌 | [语音识别控制台](https://console.volcengine.com/speech/service/16) |
+| `VOLCENGINE_TTS_API_KEY` | TTS 服务的 API Key（新版控制台） | [语音合成控制台](https://console.volcengine.com/speech/service/16) |
+| `VOLCENGINE_STT_API_KEY` | STT 服务的 API Key（新版控制台） | [语音识别控制台](https://console.volcengine.com/speech/service/16) |
 | `VOLCENGINE_LLM_API_KEY` | LLM 服务的 API 密钥 | [大模型控制台](https://console.volcengine.com/ark/) |
-| `VOLCENGINE_REALTIME_ACCESS_TOKEN` | 实时服务的访问令牌 | [实时语音控制台](https://console.volcengine.com/speech/service/10011) |
+| `VOLCENGINE_REALTIME_API_KEY` | 实时服务的 API Key（新版控制台） | [实时语音控制台](https://console.volcengine.com/speech/service/10011) |
 | `VOLCENGINE_TTS_RESOURCE_ID` | TTS 资源 ID，默认 `seed-tts-2.0` | [豆包语音合成接口文档](https://www.volcengine.com/docs/6561/1257584) |
 
 ### .env 文件示例
 
 ```bash
 # .env
-VOLCENGINE_TTS_ACCESS_TOKEN=your_tts_token_here
-VOLCENGINE_STT_ACCESS_TOKEN=your_stt_token_here
+VOLCENGINE_TTS_API_KEY=your_tts_api_key_here
+VOLCENGINE_STT_API_KEY=your_stt_api_key_here
 VOLCENGINE_LLM_API_KEY=your_llm_api_key_here
-VOLCENGINE_REALTIME_ACCESS_TOKEN=your_realtime_token_here
+VOLCENGINE_REALTIME_API_KEY=your_realtime_api_key_here
 ```
 
 ## 📖 使用指南
@@ -125,8 +128,7 @@ async def entry_point(ctx: JobContext):
 
     # 使用实时语音模型
     llm = volcengine.RealtimeModel(
-        app_id="your_app_id",
-        access_token="your_access_token",
+        api_key="your_api_key",
         bot_name="智能助手",
         model="O"  # 或 "SC"
     )
@@ -152,11 +154,11 @@ async def entry_point(ctx: JobContext):
     agent = Agent(instructions="You are a helpful assistant.")
 
     # 使用豆包大模型流式语音识别
-    stt = volcengine.STT(app_id="your_app_id")
+    stt = volcengine.STT(api_key="your_api_key")
 
     session = AgentSession(
         stt=stt,
-        tts=volcengine.TTS(app_id="your_tts_app_id"),
+        tts=volcengine.TTS(api_key="your_tts_api_key"),
         llm=volcengine.LLM(model="doubao-1-5-pro-32k-250115")
     )
 
@@ -184,8 +186,7 @@ async def entry_point(ctx: JobContext):
 
     # 基础实时语音模型配置
     realtime_llm = volcengine.RealtimeModel(
-        app_id="your_app_id",
-        access_token="your_access_token",
+        api_key="your_api_key",
         bot_name="豆包",  # 机器人名称
         model="O"  # 默认使用O版本，功能全面
     )
@@ -218,8 +219,7 @@ async def entry_point(ctx: JobContext):
 
     # 高级配置 - 包含网络搜索和RAG功能
     realtime_llm = volcengine.RealtimeModel(
-        app_id="your_app_id",
-        access_token="your_access_token",
+        api_key="your_api_key",
         bot_name="智能助手",
         model="O",  # O版本支持联网搜索和RAG，SC版本专注角色扮演
         enable_volc_websearch=True,  # 启用联网搜索
@@ -327,8 +327,7 @@ async def entry_point(ctx: JobContext):
 
     # SC版本支持角色配置和克隆音色
     realtime_llm = volcengine.RealtimeModel(
-        app_id="your_app_id",
-        access_token="your_access_token",
+        api_key="your_api_key",
         bot_name="小智",
         model="SC",  # SC版本支持character_manifest
         character_manifest='{  # JSON字符串格式的角色配置
@@ -366,12 +365,12 @@ async def entry_point(ctx: JobContext):
     session = AgentSession(
         # 语音识别
         stt=volcengine.STT(
-            app_id="your_stt_app_id",
+            api_key="your_stt_api_key",
             resource_id="volc.seedasr.sauc.duration"
         ),
         # 语音合成
         tts=volcengine.TTS(
-            app_id="your_tts_app_id",
+            api_key="your_tts_api_key",
             resource_id="seed-tts-2.0",
             voice="zh_female_xiaohe_uranus_bigtts"
         ),
@@ -395,13 +394,14 @@ if __name__ == "__main__":
 
 ```python
 volcengine.TTS(
-    app_id: str,           # 应用ID
-    access_token: str | None = None,    # 访问令牌，可用环境变量
-    resource_id: str | None = None,     # 默认 seed-tts-2.0
+    api_key: str,                    # API Key（新版控制台）
+    resource_id: str | None = None,  # 默认 seed-tts-2.0
     voice: str = "zh_female_xiaohe_uranus_bigtts",
-    speed: float = 1.0,
-    volume: float = 1.0,
-    pitch: float = 1.0,
+    speed: float = 1.0,              # 已废弃，保留仅用于兼容
+    volume: float = 1.0,             # 已废弃，保留仅用于兼容
+    pitch: float = 1.0,              # 已废弃，保留仅用于兼容
+    speech_rate: int = 0,            # 语速，[-50, 100]，100=2.0x，-50=0.5x
+    loudness_rate: int = 0,          # 音量，[-50, 100]，100=2.0x，-50=0.5x（mix 音色暂不支持）
     sample_rate: Literal[24000, 16000, 8000] = 16000,
 )
 ```
@@ -421,8 +421,7 @@ volcengine.TTS(
 
 ```python
 volcengine.STT(
-    app_id: str | None = None,
-    access_token: str | None = None,
+    api_key: str | None = None,        # API Key（新版控制台），可用环境变量 VOLCENGINE_STT_API_KEY
     resource_id: str | None = None,  # 例如 volc.bigasr.sauc.duration / volc.seedasr.sauc.duration
     model_name: str = "bigmodel",
     enable_itn: bool = False,
@@ -450,8 +449,7 @@ volcengine.LLM(
 ```python
 volcengine.RealtimeModel(
     # 必需参数
-    app_id: str = None,                    # 应用ID，从环境变量 VOLCENGINE_REALTIME_APP_ID 获取
-    access_token: str = None,              # 访问令牌，从环境变量 VOLCENGINE_REALTIME_ACCESS_TOKEN 获取
+    api_key: str = None,                   # API Key（新版控制台），从环境变量 VOLCENGINE_REALTIME_API_KEY 获取
 
     # 基本配置
     bot_name: str = "豆包",                # 机器人名称
@@ -588,6 +586,9 @@ A: 根据您的应用需求选择合适的版本：
 - 修复无音频返回时静默失败的问题
 - 删除旧版普通 STT，实现统一为豆包大模型流式语音识别
 - `volcengine.STT` 补齐 `X-Api-Connect-Id` 握手头
+- `volcengine.STT` 切换到新版控制台鉴权，移除 `app_id` / `access_token`，改用单一 `api_key` 并发送 `X-Api-Key` 头
+- `volcengine.TTS` 同步切换到新版控制台鉴权，移除 `app_id` / `access_token`，改用单一 `api_key` 并发送 `X-Api-Key` 头
+- `volcengine.RealtimeModel` 同步切换到新版控制台鉴权，移除 `app_id` / `access_token`，改用单一 `api_key` 并发送 `X-Api-Key` 头
 
 ## 🤝 贡献
 

{livekit_plugins_volcenginee-1.3.0 → livekit_plugins_volcenginee-1.3.2}/README.md

@@ -16,6 +16,9 @@
 - `volcengine.TTS` 默认音色为 `zh_female_xiaohe_uranus_bigtts`
 - `volcengine.STT` 现已直接使用豆包大模型流式语音识别接口
 - `volcengine.STT` 已补齐新版握手头 `X-Api-Connect-Id`
+- `volcengine.STT` 切换到新版控制台鉴权，仅使用 `X-Api-Key` 头
+- `volcengine.TTS` 同步切换到新版控制台鉴权，仅使用 `X-Api-Key` 头
+- `volcengine.RealtimeModel` 同步切换到新版控制台鉴权，仅使用 `X-Api-Key` 头
 
 ## ✨ 特性
 
@@ -73,20 +76,20 @@ pip install -e .
 
 | 环境变量 | 描述 | 获取方式 |
 |----------|------|----------|
-| `VOLCENGINE_TTS_ACCESS_TOKEN` | TTS 服务的访问令牌 | [语音合成控制台](https://console.volcengine.com/speech/service/16) |
-| `VOLCENGINE_STT_ACCESS_TOKEN` | STT 服务的访问令牌 | [语音识别控制台](https://console.volcengine.com/speech/service/16) |
+| `VOLCENGINE_TTS_API_KEY` | TTS 服务的 API Key（新版控制台） | [语音合成控制台](https://console.volcengine.com/speech/service/16) |
+| `VOLCENGINE_STT_API_KEY` | STT 服务的 API Key（新版控制台） | [语音识别控制台](https://console.volcengine.com/speech/service/16) |
 | `VOLCENGINE_LLM_API_KEY` | LLM 服务的 API 密钥 | [大模型控制台](https://console.volcengine.com/ark/) |
-| `VOLCENGINE_REALTIME_ACCESS_TOKEN` | 实时服务的访问令牌 | [实时语音控制台](https://console.volcengine.com/speech/service/10011) |
+| `VOLCENGINE_REALTIME_API_KEY` | 实时服务的 API Key（新版控制台） | [实时语音控制台](https://console.volcengine.com/speech/service/10011) |
 | `VOLCENGINE_TTS_RESOURCE_ID` | TTS 资源 ID，默认 `seed-tts-2.0` | [豆包语音合成接口文档](https://www.volcengine.com/docs/6561/1257584) |
 
 ### .env 文件示例
 
 ```bash
 # .env
-VOLCENGINE_TTS_ACCESS_TOKEN=your_tts_token_here
-VOLCENGINE_STT_ACCESS_TOKEN=your_stt_token_here
+VOLCENGINE_TTS_API_KEY=your_tts_api_key_here
+VOLCENGINE_STT_API_KEY=your_stt_api_key_here
 VOLCENGINE_LLM_API_KEY=your_llm_api_key_here
-VOLCENGINE_REALTIME_ACCESS_TOKEN=your_realtime_token_here
+VOLCENGINE_REALTIME_API_KEY=your_realtime_api_key_here
 ```
 
 ## 📖 使用指南
@@ -103,8 +106,7 @@ async def entry_point(ctx: JobContext):
 
     # 使用实时语音模型
     llm = volcengine.RealtimeModel(
-        app_id="your_app_id",
-        access_token="your_access_token",
+        api_key="your_api_key",
         bot_name="智能助手",
         model="O"  # 或 "SC"
     )
@@ -130,11 +132,11 @@ async def entry_point(ctx: JobContext):
     agent = Agent(instructions="You are a helpful assistant.")
 
     # 使用豆包大模型流式语音识别
-    stt = volcengine.STT(app_id="your_app_id")
+    stt = volcengine.STT(api_key="your_api_key")
 
     session = AgentSession(
         stt=stt,
-        tts=volcengine.TTS(app_id="your_tts_app_id"),
+        tts=volcengine.TTS(api_key="your_tts_api_key"),
         llm=volcengine.LLM(model="doubao-1-5-pro-32k-250115")
     )
 
@@ -162,8 +164,7 @@ async def entry_point(ctx: JobContext):
 
     # 基础实时语音模型配置
     realtime_llm = volcengine.RealtimeModel(
-        app_id="your_app_id",
-        access_token="your_access_token",
+        api_key="your_api_key",
         bot_name="豆包",  # 机器人名称
         model="O"  # 默认使用O版本，功能全面
     )
@@ -196,8 +197,7 @@ async def entry_point(ctx: JobContext):
 
     # 高级配置 - 包含网络搜索和RAG功能
     realtime_llm = volcengine.RealtimeModel(
-        app_id="your_app_id",
-        access_token="your_access_token",
+        api_key="your_api_key",
         bot_name="智能助手",
         model="O",  # O版本支持联网搜索和RAG，SC版本专注角色扮演
         enable_volc_websearch=True,  # 启用联网搜索
@@ -305,8 +305,7 @@ async def entry_point(ctx: JobContext):
 
     # SC版本支持角色配置和克隆音色
     realtime_llm = volcengine.RealtimeModel(
-        app_id="your_app_id",
-        access_token="your_access_token",
+        api_key="your_api_key",
         bot_name="小智",
         model="SC",  # SC版本支持character_manifest
         character_manifest='{  # JSON字符串格式的角色配置
@@ -344,12 +343,12 @@ async def entry_point(ctx: JobContext):
     session = AgentSession(
         # 语音识别
         stt=volcengine.STT(
-            app_id="your_stt_app_id",
+            api_key="your_stt_api_key",
             resource_id="volc.seedasr.sauc.duration"
         ),
         # 语音合成
         tts=volcengine.TTS(
-            app_id="your_tts_app_id",
+            api_key="your_tts_api_key",
             resource_id="seed-tts-2.0",
             voice="zh_female_xiaohe_uranus_bigtts"
         ),
@@ -373,13 +372,14 @@ if __name__ == "__main__":
 
 ```python
 volcengine.TTS(
-    app_id: str,           # 应用ID
-    access_token: str | None = None,    # 访问令牌，可用环境变量
-    resource_id: str | None = None,     # 默认 seed-tts-2.0
+    api_key: str,                    # API Key（新版控制台）
+    resource_id: str | None = None,  # 默认 seed-tts-2.0
     voice: str = "zh_female_xiaohe_uranus_bigtts",
-    speed: float = 1.0,
-    volume: float = 1.0,
-    pitch: float = 1.0,
+    speed: float = 1.0,              # 已废弃，保留仅用于兼容
+    volume: float = 1.0,             # 已废弃，保留仅用于兼容
+    pitch: float = 1.0,              # 已废弃，保留仅用于兼容
+    speech_rate: int = 0,            # 语速，[-50, 100]，100=2.0x，-50=0.5x
+    loudness_rate: int = 0,          # 音量，[-50, 100]，100=2.0x，-50=0.5x（mix 音色暂不支持）
     sample_rate: Literal[24000, 16000, 8000] = 16000,
 )
 ```
@@ -399,8 +399,7 @@ volcengine.TTS(
 
 ```python
 volcengine.STT(
-    app_id: str | None = None,
-    access_token: str | None = None,
+    api_key: str | None = None,        # API Key（新版控制台），可用环境变量 VOLCENGINE_STT_API_KEY
     resource_id: str | None = None,  # 例如 volc.bigasr.sauc.duration / volc.seedasr.sauc.duration
     model_name: str = "bigmodel",
     enable_itn: bool = False,
@@ -428,8 +427,7 @@ volcengine.LLM(
 ```python
 volcengine.RealtimeModel(
     # 必需参数
-    app_id: str = None,                    # 应用ID，从环境变量 VOLCENGINE_REALTIME_APP_ID 获取
-    access_token: str = None,              # 访问令牌，从环境变量 VOLCENGINE_REALTIME_ACCESS_TOKEN 获取
+    api_key: str = None,                   # API Key（新版控制台），从环境变量 VOLCENGINE_REALTIME_API_KEY 获取
 
     # 基本配置
     bot_name: str = "豆包",                # 机器人名称
@@ -566,6 +564,9 @@ A: 根据您的应用需求选择合适的版本：
 - 修复无音频返回时静默失败的问题
 - 删除旧版普通 STT，实现统一为豆包大模型流式语音识别
 - `volcengine.STT` 补齐 `X-Api-Connect-Id` 握手头
+- `volcengine.STT` 切换到新版控制台鉴权，移除 `app_id` / `access_token`，改用单一 `api_key` 并发送 `X-Api-Key` 头
+- `volcengine.TTS` 同步切换到新版控制台鉴权，移除 `app_id` / `access_token`，改用单一 `api_key` 并发送 `X-Api-Key` 头
+- `volcengine.RealtimeModel` 同步切换到新版控制台鉴权，移除 `app_id` / `access_token`，改用单一 `api_key` 并发送 `X-Api-Key` 头
 
 ## 🤝 贡献
 

{livekit_plugins_volcenginee-1.3.0 → livekit_plugins_volcenginee-1.3.2}/livekit/plugins/volcengine/llm.py

@@ -2,7 +2,7 @@ from __future__ import annotations
 
 import os
 from dataclasses import dataclass
-from typing import Any
+from typing import Any, Literal
 
 import httpx
 import openai
@@ -31,8 +31,8 @@ class _LLMOptions:
     temperature: NotGivenOr[float]
     parallel_tool_calls: NotGivenOr[bool]
     tool_choice: NotGivenOr[ToolChoice]
-    store: NotGivenOr[bool]
     metadata: NotGivenOr[dict[str, str]]
+    thinking: NotGivenOr[Literal["enabled", "disabled", "auto"] | dict[str, Any]]
 
 
 class LLM(llm.LLM):
@@ -47,8 +47,8 @@ class LLM(llm.LLM):
         temperature: NotGivenOr[float] = NOT_GIVEN,
         parallel_tool_calls: NotGivenOr[bool] = NOT_GIVEN,
         tool_choice: NotGivenOr[ToolChoice] = NOT_GIVEN,
-        store: NotGivenOr[bool] = NOT_GIVEN,
         metadata: NotGivenOr[dict[str, str]] = NOT_GIVEN,
+        thinking: NotGivenOr[Literal["enabled", "disabled", "auto"] | dict[str, Any]] = NOT_GIVEN,
         timeout: httpx.Timeout | None = None,
     ) -> None:
         """
@@ -58,6 +58,11 @@ class LLM(llm.LLM):
             model: The model to use for the LLM, end
             api_key: The API key to use for the LLM.
             base_url: The base URL to use for the LLM.
+            thinking: Controls deep thinking mode. Pass ``"disabled"`` to turn
+                it off, ``"enabled"`` to force it on, ``"auto"`` to let the
+                model decide, or a raw ``{"type": "..."}`` dict for extended
+                parameters. See the Volcengine Chat Completions API reference
+                for details. Omit to use the API default (``enabled``).
         """
         super().__init__()
         self._opts = _LLMOptions(
@@ -66,8 +71,8 @@ class LLM(llm.LLM):
             temperature=temperature,
             parallel_tool_calls=parallel_tool_calls,
             tool_choice=tool_choice,
-            store=store,
             metadata=metadata,
+            thinking=thinking,
         )
         api_key = api_key if is_given(api_key) else os.getenv("VOLCENGINE_LLM_API_KEY")
         if api_key is None:
@@ -110,6 +115,16 @@ class LLM(llm.LLM):
         if is_given(self._opts.user):
             extra["user"] = self._opts.user
 
+        if is_given(self._opts.temperature):
+            extra["temperature"] = self._opts.temperature
+
+        if is_given(self._opts.thinking):
+            extra["thinking"] = (
+                self._opts.thinking
+                if isinstance(self._opts.thinking, dict)
+                else {"type": self._opts.thinking}
+            )
+
         parallel_tool_calls = (
             parallel_tool_calls
             if is_given(parallel_tool_calls)

{livekit_plugins_volcenginee-1.3.0 → livekit_plugins_volcenginee-1.3.2}/livekit/plugins/volcengine/realtime.py

@@ -163,8 +163,7 @@ def parse_response(res):
 
 @dataclass
 class _RealtimeOptions:
-    app_id: str
-    access_token: str
+    api_key: str
     bot_name: str
     system_role: str
     max_session_duration: float | None
@@ -190,10 +189,8 @@ class _RealtimeOptions:
 
     def get_ws_headers(self) -> dict:
         headers = {
-            "X-Api-App-ID": self.app_id,
-            "X-Api-Access-Key": self.access_token,
+            "X-Api-Key": self.api_key,
             "X-Api-Resource-Id": "volc.speech.dialog",  # 固定值
-            "X-Api-App-Key": "PlgvMymc7f3tQnJ6",  # 固定值
             "X-Api-Connect-Id": str(uuid.uuid4()),
         }
         return headers
@@ -262,8 +259,7 @@ class RealtimeModel(llm.RealtimeModel):
         speaking_style: str = "你的说话风格简洁明了，语速适中，语调自然。",
         speaker: str = "zh_female_vv_jupiter_bigtts",
         opening: str | None = None,
-        app_id: str | None = None,
-        access_token: str | None = None,
+        api_key: str | None = None,
         system_role: str | None = None,
         character_manifest: str = None,
         model: Literal["O", "SC"] = "O",
@@ -299,17 +295,11 @@ class RealtimeModel(llm.RealtimeModel):
         logger.info(
             f"Volc Websearch No Result Message: {volc_websearch_no_result_message}"
         )
-        app_id = app_id or os.environ.get("VOLCENGINE_REALTIME_APP_ID")
-        if app_id is None:
-            raise ValueError("VOLCENGINE_REALTIME_APP_ID is required")
-        access_token = access_token or os.environ.get(
-            "VOLCENGINE_REALTIME_ACCESS_TOKEN"
-        )
-        if access_token is None:
-            raise ValueError("VOLCENGINE_REALTIME_ACCESS_TOKEN is required")
+        api_key = api_key or os.environ.get("VOLCENGINE_REALTIME_API_KEY")
+        if api_key is None:
+            raise ValueError("VOLCENGINE_REALTIME_API_KEY is required")
         self._opts = _RealtimeOptions(
-            app_id=app_id,
-            access_token=access_token,
+            api_key=api_key,
             bot_name=bot_name,
             system_role=system_role,
             speaker=speaker,

{livekit_plugins_volcenginee-1.3.0 → livekit_plugins_volcenginee-1.3.2}/livekit/plugins/volcengine/stt.py

@@ -76,8 +76,7 @@ def generate_before_payload(sequence: int):
 
 @dataclass
 class STTOptions:
-    app_id: str | None = None
-    access_token: str | None = None
+    api_key: str | None = None
     source_type: Literal["duration", "concurrent"] = "duration"
     resource_id: str | None = None
 
@@ -167,16 +166,11 @@ class STTOptions:
             header["X-Api-Resource-Id"] = "volc.bigasr.sauc.duration"
         else:
             header["X-Api-Resource-Id"] = "volc.bigasr.sauc.concurrent"
-        if self.app_id is None:
-            self.app_id = os.environ.get("VOLCENGINE_STT_APP_ID", None)
-            if self.app_id is None:
-                raise ValueError("VOLCENGINE_STT_APP_ID is not set")
-        if self.access_token is None:
-            self.access_token = os.environ.get("VOLCENGINE_STT_ACCESS_TOKEN", None)
-            if self.access_token is None:
-                raise ValueError("VOLCENGINE_STT_ACCESS_TOKEN is not set")
-        header["X-Api-Access-Key"] = self.access_token
-        header["X-Api-App-Key"] = self.app_id
+        if self.api_key is None:
+            self.api_key = os.environ.get("VOLCENGINE_STT_API_KEY", None)
+            if self.api_key is None:
+                raise ValueError("VOLCENGINE_STT_API_KEY is not set")
+        header["X-Api-Key"] = self.api_key
         header["X-Api-Connect-Id"] = reqid
         header["X-Api-Request-Id"] = reqid
         return header
@@ -186,9 +180,8 @@ class STT(stt.STT):
     def __init__(
         self,
         *,
-        app_id: str | None = None,
+        api_key: str | None = None,
         base_url: str = "wss://openspeech.bytedance.com/api/v3/sauc/bigmodel",
-        access_token: str | None = None,
         resource_id: str | None = None,
         model_name: str = "bigmodel",
         enable_itn: bool = False,
@@ -208,8 +201,7 @@ class STT(stt.STT):
 
         self._opts = STTOptions(
             base_url=base_url,
-            access_token=access_token,
-            app_id=app_id,
+            api_key=api_key,
             resource_id=resource_id,
             model_name=model_name,
             enable_itn=enable_itn,

{livekit_plugins_volcenginee-1.3.0 → livekit_plugins_volcenginee-1.3.2}/livekit/plugins/volcengine/tts.py

@@ -25,8 +25,7 @@ from .log import logger
 
 
 class _TTSOptions(BaseModel):
-    app_id: str
-    access_token: str | None = None
+    api_key: str
     resource_id: str | None = None
     voice: str = "zh_female_xiaohe_uranus_bigtts"
     base_url: str = "https://openspeech.bytedance.com"
@@ -35,6 +34,8 @@ class _TTSOptions(BaseModel):
     speed: float = Field(1.0, ge=0.2, le=3.0)
     volume: float = Field(1.0, gt=0.1, le=3.0)
     pitch: float = Field(1.0, ge=0.1, le=3.0)
+    speech_rate: int = Field(0, ge=-50, le=100)
+    loudness_rate: int = Field(0, ge=-50, le=100)
 
     def get_http_url(self) -> str:
         return f"{self.base_url}/api/v3/tts/unidirectional"
@@ -54,24 +55,20 @@ class _TTSOptions(BaseModel):
                 "audio_params": {
                     "format": self.encoding,
                     "sample_rate": self.sample_rate,
+                    "loudness_rate": self.loudness_rate,
+                    "speech_rate": self.speech_rate,
                 },
                 "request_id": reqid,
             },
         }
 
     def get_http_header(self, reqid: str | None = None) -> dict[str, str]:
-        if self.access_token is None:
-            self.access_token = os.getenv("VOLCENGINE_TTS_ACCESS_TOKEN")
-            if self.access_token is None:
-                raise ValueError("VOLCENGINE_TTS_ACCESS_TOKEN is not set")
         resource_id = self.resource_id or os.getenv(
             "VOLCENGINE_TTS_RESOURCE_ID", "seed-tts-2.0"
         )
         headers = {
             "Content-Type": "application/json",
-            "X-Api-App-Id": self.app_id,
-            "X-Api-App-Key": self.app_id,
-            "X-Api-Access-Key": self.access_token,
+            "X-Api-Key": self.api_key,
             "X-Api-Resource-Id": resource_id,
         }
         if reqid is not None:
@@ -82,23 +79,28 @@ class _TTSOptions(BaseModel):
 class TTS(tts.TTS):
     def __init__(
         self,
-        app_id: str,
-        access_token: str | None = None,
+        api_key: str,
         resource_id: str | None = None,
         voice: str = "zh_female_xiaohe_uranus_bigtts",
         speed: float = 1.0,
         volume: float = 1.0,
         pitch: float = 1.0,
+        speech_rate: int = 0,
+        loudness_rate: int = 0,
         sample_rate: Literal[24000, 16000, 8000] = 16000,
         http_session: aiohttp.ClientSession | None = None,
     ):
         """VolcEngine TTS
 
         Args:
-            app_id (str): the app id of the tts, you can get it from the console.
-            access_token (str | None, optional): the access token of the tts, if not provided, the value of the environment variable VOLCENGINE_TTS_ACCESS_TOKEN will be used. Defaults to None.
+            api_key (str): the API key of the tts, you can get it from the new console.
             resource_id (str | None, optional): VolcEngine TTS resource id. Use `seed-tts-2.0` for Doubao TTS 2.0 voices, `seed-tts-1.0` / `seed-tts-1.0-concurr` for Doubao TTS 1.0 voices, `seed-icl-2.0` for voice cloning 2.0, and `seed-icl-1.0` / `seed-icl-1.0-concurr` for voice cloning 1.0. Defaults to None.
             voice (str, optional): the voice id used by the tts request. Defaults to `zh_female_xiaohe_uranus_bigtts`.
+            speed (float, optional): kept for backwards compatibility, no longer sent to the API. Defaults to 1.0.
+            volume (float, optional): kept for backwards compatibility, no longer sent to the API. Defaults to 1.0.
+            pitch (float, optional): kept for backwards compatibility, no longer sent to the API. Defaults to 1.0.
+            speech_rate (int, optional): speech rate, range [-50, 100]. 100 = 2.0x, -50 = 0.5x. Defaults to 0 (normal).
+            loudness_rate (int, optional): volume gain, range [-50, 100]. 100 = 2.0x, -50 = 0.5x. Not supported for mix voices. Defaults to 0 (normal).
             sample_rate (Literal[24000, 16000, 8000], optional): the sample rate of the tts. Defaults to 24000.
             streaming (bool, optional): whether to use the streaming api. Defaults to True.
             http_session (aiohttp.ClientSession | None, optional): the http session to use. Defaults to None.
@@ -109,14 +111,15 @@ class TTS(tts.TTS):
             num_channels=1,
         )
         self._opts = _TTSOptions(
-            app_id=app_id,
-            access_token=access_token,
+            api_key=api_key,
             resource_id=resource_id,
             voice=voice,
             sample_rate=sample_rate,
             speed=speed,
             volume=volume,
             pitch=pitch,
+            speech_rate=speech_rate,
+            loudness_rate=loudness_rate,
         )
         self._session = http_session
         self._streams = weakref.WeakSet[SynthesizeStream]()

livekit_plugins_volcenginee-1.3.2/livekit/plugins/volcengine/version.py

@@ -0,0 +1 @@
+__version__ = "1.3.2"

livekit_plugins_volcenginee-1.3.0/livekit/plugins/volcengine/version.py

@@ -1 +0,0 @@
-__version__ = "1.3.0"