mirror of
https://github.com/xinnan-tech/xiaozhi-esp32-server.git
synced 2026-07-28 01:53:53 +08:00
update:优化提示词
This commit is contained in:
@@ -2,7 +2,10 @@ from config.logger import setup_logging
|
||||
import json
|
||||
import asyncio
|
||||
import time
|
||||
from core.utils.util import remove_punctuation_and_length, get_string_no_punctuation_or_emoji
|
||||
from core.utils.util import (
|
||||
remove_punctuation_and_length,
|
||||
get_string_no_punctuation_or_emoji,
|
||||
)
|
||||
|
||||
TAG = __name__
|
||||
logger = setup_logging()
|
||||
@@ -21,10 +24,11 @@ async def sendAudioMessage(conn, audios, text, text_index=0):
|
||||
|
||||
# 发送结束消息(如果是最后一个文本)
|
||||
if conn.llm_finish_task and text_index == conn.tts_last_text_index:
|
||||
await send_tts_message(conn, 'stop', None)
|
||||
await send_tts_message(conn, "stop", None)
|
||||
if conn.close_after_chat:
|
||||
await conn.close()
|
||||
|
||||
|
||||
# 播放音频
|
||||
async def sendAudio(conn, audios):
|
||||
# 流控参数优化
|
||||
@@ -36,13 +40,12 @@ async def sendAudio(conn, audios):
|
||||
pre_buffer = min(3, len(audios))
|
||||
for i in range(pre_buffer):
|
||||
await conn.websocket.send(audios[i])
|
||||
# conn.logger.bind(tag=TAG).debug(f"预缓冲帧 {i}, 时间: {(time.perf_counter() - start_time) * 1000:.2f}ms")
|
||||
|
||||
# 正常播放剩余帧
|
||||
for opus_packet in audios[pre_buffer:]:
|
||||
if conn.client_abort:
|
||||
return
|
||||
|
||||
|
||||
# 计算预期发送时间
|
||||
expected_time = start_time + (play_position / 1000)
|
||||
current_time = time.perf_counter()
|
||||
@@ -51,19 +54,13 @@ async def sendAudio(conn, audios):
|
||||
await asyncio.sleep(delay)
|
||||
|
||||
await conn.websocket.send(opus_packet)
|
||||
# conn.logger.bind(tag=TAG).debug(f"发送帧,位置: {play_position}ms, 实际间隔: {(time.perf_counter() - current_time) * 1000:.2f}ms")
|
||||
|
||||
play_position += frame_duration
|
||||
|
||||
|
||||
|
||||
async def send_tts_message(conn, state, text=None):
|
||||
"""发送 TTS 状态消息"""
|
||||
message = {
|
||||
"type": "tts",
|
||||
"state": state,
|
||||
"session_id": conn.session_id
|
||||
}
|
||||
message = {"type": "tts", "state": state, "session_id": conn.session_id}
|
||||
if text is not None:
|
||||
message["text"] = text
|
||||
|
||||
@@ -72,7 +69,9 @@ async def send_tts_message(conn, state, text=None):
|
||||
# 播放提示音
|
||||
tts_notify = conn.config.get("enable_stop_tts_notify", False)
|
||||
if tts_notify:
|
||||
stop_tts_notify_voice = conn.config.get("stop_tts_notify_voice", "config/assets/tts_notify.mp3")
|
||||
stop_tts_notify_voice = conn.config.get(
|
||||
"stop_tts_notify_voice", "config/assets/tts_notify.mp3"
|
||||
)
|
||||
audios, duration = conn.tts.audio_to_opus_data(stop_tts_notify_voice)
|
||||
await sendAudio(conn, audios)
|
||||
# 清除服务端讲话状态
|
||||
@@ -85,16 +84,17 @@ async def send_tts_message(conn, state, text=None):
|
||||
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}
|
||||
))
|
||||
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")
|
||||
|
||||
Reference in New Issue
Block a user