- Python main entry with CLI (--room-id, --voice, --speed) - blivedm integration for real-time danmaku receiving - Chinese-to-Japanese katakana conversion (pypinyin + mapping table) - Node.js persistent bridge for aquestalk.js TTS synthesis - Audio playback via sounddevice - Message queue for sequential TTS playback
195 lines
4.9 KiB
Python
195 lines
4.9 KiB
Python
"""bililive-touhou-tts — Bilibili Live Danmaku → Yukkuri TTS Reader.
|
|
|
|
Usage:
|
|
python main.py --room-id 12345
|
|
python main.py --room-id 12345 --voice f2 --speed 120
|
|
"""
|
|
|
|
import argparse
|
|
import asyncio
|
|
import logging
|
|
import signal
|
|
import sys
|
|
from pathlib import Path
|
|
|
|
from danmaku_handler import DanmakuClient
|
|
from chinese2kana import chinese_to_kana
|
|
from tts import TTSBridge
|
|
from audio_player import play_wav_async, get_output_devices
|
|
|
|
logger = logging.getLogger("bililive-tts")
|
|
|
|
DEFAULT_VOICE = "f1"
|
|
DEFAULT_SPEED = 100
|
|
|
|
|
|
def parse_args() -> argparse.Namespace:
|
|
parser = argparse.ArgumentParser(
|
|
description="Bilibili Live Danmaku → Yukkuri TTS Reader"
|
|
)
|
|
parser.add_argument(
|
|
"--room-id", "-r",
|
|
type=int,
|
|
required=True,
|
|
help="Bilibili live room ID (from the URL)",
|
|
)
|
|
parser.add_argument(
|
|
"--voice", "-v",
|
|
type=str,
|
|
default=DEFAULT_VOICE,
|
|
choices=["f1", "f2", "m1", "m2", "dvd", "imd1", "jgr", "r1"],
|
|
help=f"Yukkuri voice type (default: {DEFAULT_VOICE})",
|
|
)
|
|
parser.add_argument(
|
|
"--speed", "-s",
|
|
type=int,
|
|
default=DEFAULT_SPEED,
|
|
help=f"Speech speed, 50-300 (default: {DEFAULT_SPEED})",
|
|
)
|
|
parser.add_argument(
|
|
"--sessdata",
|
|
type=str,
|
|
default="",
|
|
help="Bilibili SESSDATA cookie for authenticated access (optional)",
|
|
)
|
|
parser.add_argument(
|
|
"--format", "-f",
|
|
type=str,
|
|
default="{uname}: {msg}",
|
|
help="Danmaku message format. Variables: {uname}, {msg}, {uid}",
|
|
)
|
|
parser.add_argument(
|
|
"--no-numbers",
|
|
action="store_true",
|
|
help="Do not convert Arabic numerals to Chinese words",
|
|
)
|
|
parser.add_argument(
|
|
"--list-devices",
|
|
action="store_true",
|
|
help="List available audio output devices and exit",
|
|
)
|
|
parser.add_argument(
|
|
"--debug", "-d",
|
|
action="store_true",
|
|
help="Enable debug logging",
|
|
)
|
|
return parser.parse_args()
|
|
|
|
|
|
async def tts_worker(
|
|
queue: asyncio.Queue,
|
|
bridge: TTSBridge,
|
|
convert_numbers: bool,
|
|
shutdown_event: asyncio.Event,
|
|
) -> None:
|
|
"""Consume danmaku messages from the queue, convert to kana, and play TTS."""
|
|
logger.info("TTS worker started.")
|
|
|
|
while not shutdown_event.is_set():
|
|
try:
|
|
text = await asyncio.wait_for(queue.get(), timeout=1.0)
|
|
except asyncio.TimeoutError:
|
|
continue
|
|
|
|
try:
|
|
kana = chinese_to_kana(text, convert_numbers=convert_numbers)
|
|
logger.info("Speaking: %s → %s", text, kana)
|
|
|
|
wav_data = await bridge.synthesize(kana)
|
|
await play_wav_async(wav_data)
|
|
|
|
except Exception:
|
|
logger.exception("Error processing danmaku: %s", text)
|
|
|
|
finally:
|
|
queue.task_done()
|
|
|
|
logger.info("TTS worker stopped.")
|
|
|
|
|
|
async def main_async(args: argparse.Namespace) -> int:
|
|
if args.list_devices:
|
|
devices = get_output_devices()
|
|
print("Available audio output devices:")
|
|
for dev in devices:
|
|
print(f" [{dev['index']}] {dev['name']} ({dev['channels']}ch)")
|
|
return 0
|
|
|
|
if not (50 <= args.speed <= 300):
|
|
logger.error("Speed must be between 50 and 300.")
|
|
return 1
|
|
|
|
queue: asyncio.Queue = asyncio.Queue(maxsize=256)
|
|
shutdown_event = asyncio.Event()
|
|
|
|
bridge = TTSBridge(voice=args.voice, speed=args.speed)
|
|
|
|
danmaku_client = DanmakuClient(
|
|
room_id=args.room_id,
|
|
queue=queue,
|
|
sessdata=args.sessdata,
|
|
message_format=args.format,
|
|
)
|
|
|
|
async def handle_signal() -> None:
|
|
"""Wait for shutdown signal."""
|
|
loop = asyncio.get_running_loop()
|
|
sig_event = asyncio.Event()
|
|
|
|
def _handler():
|
|
sig_event.set()
|
|
|
|
for sig in (signal.SIGINT, signal.SIGTERM):
|
|
try:
|
|
loop.add_signal_handler(sig, _handler)
|
|
except NotImplementedError:
|
|
pass
|
|
|
|
await sig_event.wait()
|
|
logger.info("Shutdown signal received.")
|
|
shutdown_event.set()
|
|
|
|
try:
|
|
await bridge.start()
|
|
await danmaku_client.start()
|
|
|
|
worker_task = asyncio.create_task(
|
|
tts_worker(queue, bridge, not args.no_numbers, shutdown_event)
|
|
)
|
|
|
|
await handle_signal()
|
|
|
|
worker_task.cancel()
|
|
try:
|
|
await worker_task
|
|
except asyncio.CancelledError:
|
|
pass
|
|
|
|
except Exception:
|
|
logger.exception("Fatal error")
|
|
return 1
|
|
finally:
|
|
await danmaku_client.stop()
|
|
await bridge.stop()
|
|
|
|
return 0
|
|
|
|
|
|
def main() -> int:
|
|
args = parse_args()
|
|
|
|
logging.basicConfig(
|
|
level=logging.DEBUG if args.debug else logging.INFO,
|
|
format="%(asctime)s [%(levelname)s] %(name)s: %(message)s",
|
|
datefmt="%H:%M:%S",
|
|
)
|
|
|
|
try:
|
|
return asyncio.run(main_async(args))
|
|
except KeyboardInterrupt:
|
|
return 0
|
|
|
|
|
|
if __name__ == "__main__":
|
|
sys.exit(main())
|