Hass/xiaozhi-esp32-server
1
0
Fork 0
mirror of https://github.com/xinnan-tech/xiaozhi-esp32-server.git synced 2026-07-28 10:03:54 +08:00
Code Issues Packages Projects Releases Wiki Activity

更新paddlespeechtts供应器

Browse Source
This commit is contained in:
FAN-yeB
2025-08-26 11:02:57 +08:00
parent ae228f3b86
commit 15291bfcf2
1 changed files with 34 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files
main/xiaozhi-server/core/providers/tts/paddle_speech.py
+34 -14
View File
@@ -1,4 +1,3 @@
import asyncio import asyncio
import json import json
import base64 import base64
@@ -9,6 +8,7 @@ import wave
import websockets import websockets
from core.providers.tts.base import TTSProviderBase from core.providers.tts.base import TTSProviderBase
from config.logger import setup_logging from config.logger import setup_logging
from datetime import datetime
TAG = __name__ TAG = __name__
logger = setup_logging() logger = setup_logging()
@@ -19,11 +19,12 @@ class TTSProvider(TTSProviderBase):
super().__init__(config, delete_audio_file) super().__init__(config, delete_audio_file)
self.url = config.get("url", "ws://192.168.1.10:8092/paddlespeech/tts/streaming") self.url = config.get("url", "ws://192.168.1.10:8092/paddlespeech/tts/streaming")
self.protocol = config.get("protocol", "websocket") self.protocol = config.get("protocol", "websocket")
if config.get("private_voice"): if config.get("private_voice"):
self.spk_id = int(config.get("private_voice")) self.spk_id = int(config.get("private_voice"))
else: else:
self.spk_id = int(config.get("spk_id", "0")) self.spk_id = int(config.get("spk_id", "0"))
sample_rate = config.get("sample_rate", 24000) sample_rate = config.get("sample_rate", 24000)
self.sample_rate = float(sample_rate) if sample_rate else 24000 self.sample_rate = float(sample_rate) if sample_rate else 24000
@@ -33,7 +34,12 @@ class TTSProvider(TTSProviderBase):
volume = config.get("volume", 1.0) volume = config.get("volume", 1.0)
self.volume = float(volume) if volume else 1.0 self.volume = float(volume) if volume else 1.0
self.save_path = config.get("save_path", "None") self.delete_audio_file = config.get("delete_audio_file", True)
if not self.delete_audio_file:
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
self.save_path = config.get("save_path", f"./streaming_tts_{timestamp}.wav")
else:
self.save_path = None
async def pcm_to_wav(self, pcm_data: bytes, sample_rate: int = 24000, num_channels: int = 1, async def pcm_to_wav(self, pcm_data: bytes, sample_rate: int = 24000, num_channels: int = 1,
bits_per_sample: int = 16) -> bytes: bits_per_sample: int = 16) -> bytes:
@@ -56,22 +62,22 @@ class TTSProvider(TTSProviderBase):
return wav_io.getvalue() return wav_io.getvalue()
async def text_to_speak(self, text, output_file=None): async def text_to_speak(self, text, output_file):
if self.protocol == "websocket": if self.protocol == "websocket":
return await self.text_streaming(text) return await self.text_streaming(text, output_file)
elif self.protocol == "http": elif self.protocol == "http":
return await self.text(text) return await self.text(text, output_file)
else: else:
raise ValueError("Unsupported protocol. Please use 'websocket' or 'http'.") raise ValueError("Unsupported protocol. Please use 'websocket' or 'http'.")
async def text(self, text): async def text(self, text, output_file):
request_json = { request_json = {
"text": text, "text": text,
"spk_id": self.spk_id, "spk_id": self.spk_id,
"speed": self.speed, "speed": self.speed,
"volume": self.volume, "volume": self.volume,
"sample_rate": self.sample_rate, "sample_rate": self.sample_rate,
"save_path": None, "save_path": self.save_path
} }
try: try:
@@ -82,7 +88,11 @@ class TTSProvider(TTSProviderBase):
if resp_json.get("success"): if resp_json.get("success"):
data = resp_json["result"] data = resp_json["result"]
audio_bytes = base64.b64decode(data["audio"]) audio_bytes = base64.b64decode(data["audio"])
return audio_bytes if output_file:
with open(output_file, "wb") as file_to_save:
file_to_save.write(audio_bytes)
else:
return audio_bytes
else: else:
raise Exception( raise Exception(
f"Error: {resp_json.get('message', 'Unknown error')} while processing text: {text}") f"Error: {resp_json.get('message', 'Unknown error')} while processing text: {text}")
@@ -92,7 +102,7 @@ class TTSProvider(TTSProviderBase):
except Exception as e: except Exception as e:
raise Exception(f"Error during TTS HTTP request: {e} while processing text: {text}") raise Exception(f"Error during TTS HTTP request: {e} while processing text: {text}")
async def text_streaming(self, text): async def text_streaming(self, text, output_file):
try: try:
# 使用 websockets 异步连接到 WebSocket 服务器 # 使用 websockets 异步连接到 WebSocket 服务器
async with websockets.connect(self.url) as ws: async with websockets.connect(self.url) as ws:
@@ -148,8 +158,18 @@ class TTSProvider(TTSProviderBase):
# 接收结束响应避免服务抛出异常 # 接收结束响应避免服务抛出异常
await ws.recv() await ws.recv()
# 直接返回音频数据 # 根据配置决定是否保存文件
return wav_data if not self.delete_audio_file and self.save_path:
with open(self.save_path, "wb") as f:
f.write(wav_data)
logger.bind(tag=TAG).info(f"音频文件已保存到: {self.save_path}")
# 返回或保存音频数据
if output_file:
with open(output_file, "wb") as file_to_save:
file_to_save.write(wav_data)
else:
return wav_data
except Exception as e: except Exception as e:
raise Exception(f"Error during TTS WebSocket request: {e} while processing text: {text}") raise Exception(f"Error during TTS WebSocket request: {e} while processing text: {text}")