Files
xiaozhi-esp32-server/core/handle/sendAudioHandle.py
T

77 lines
2.3 KiB
Python
Raw Normal View History

2025-02-18 00:07:19 +08:00
from config.logger import setup_logging
2025-02-02 23:01:14 +08:00
import json
import asyncio
import time
from core.utils.util import remove_punctuation_and_length, get_string_no_punctuation_or_emoji
2025-02-18 00:07:19 +08:00
TAG = __name__
logger = setup_logging()
2025-02-02 23:01:14 +08:00
2025-03-01 01:54:55 +08:00
async def isLLMWantToFinish(last_text):
2025-02-14 23:09:12 +08:00
_, last_text_without_punctuation = remove_punctuation_and_length(last_text)
if "再见" in last_text_without_punctuation or "拜拜" in last_text_without_punctuation:
return True
return False
2025-03-01 01:54:55 +08:00
async def sendAudioMessage(conn, audios, text):
# 发送 tts.start
2025-02-02 23:01:14 +08:00
if text == conn.tts_first_text:
2025-02-18 00:07:19 +08:00
logger.bind(tag=TAG).info(f"发送第一段语音: {text}")
2025-02-02 23:01:14 +08:00
conn.tts_start_speak_time = time.time()
2025-03-01 01:54:55 +08:00
await send_tts_message(conn, "sentence_start", text)
2025-02-02 23:01:14 +08:00
# 发送音频数据
2025-03-01 01:54:55 +08:00
frame_duration = 60 # 初始帧持续时间(毫秒)
start_time = time.time() # 记录开始时间
for idx, opus_packet in enumerate(audios):
2025-03-01 01:54:55 +08:00
if conn.client_abort:
return
# 计算当前包的预期发送时间
expected_time = start_time + idx * (frame_duration / 1000)
current_time = time.time()
# 如果未到预期时间则等待差值
if current_time < expected_time:
await asyncio.sleep(expected_time - current_time)
# 发送音频包
2025-02-02 23:01:14 +08:00
await conn.websocket.send(opus_packet)
if conn.llm_finish_task and text == conn.tts_last_text:
2025-03-01 01:54:55 +08:00
await send_tts_message(conn, 'stop')
if await isLLMWantToFinish(text):
await conn.close()
2025-02-02 23:01:14 +08:00
async def send_tts_message(conn, state, text=None):
"""发送 TTS 状态消息"""
message = {
2025-02-02 23:01:14 +08:00
"type": "tts",
"state": state,
2025-02-02 23:01:14 +08:00
"session_id": conn.session_id
}
if text is not None:
message["text"] = text
2025-02-02 23:01:14 +08:00
await conn.websocket.send(json.dumps(message))
if state == "stop":
conn.clearSpeakStatus()
2025-02-02 23:01:14 +08:00
2025-02-11 13:51:44 +08:00
async def send_stt_message(conn, text):
"""发送 STT 状态消息"""
stt_text = get_string_no_punctuation_or_emoji(text)
await conn.websocket.send(json.dumps({
"type": "stt",
"text": stt_text,
"session_id": conn.session_id}
))
await conn.websocket.send(
json.dumps({
"type": "llm",
"text": "😊",
"emotion": "happy",
"session_id": conn.session_id}
))
await send_tts_message(conn, "start")