Files
bililive-touhou-tts/tts.py
T
chun_qiu fd5c732d8a feat: web UI + QR login + portable build
- Add web UI panel (dark theme) at / with REST API
- Add Bilibili QR code scan login with browser header emulation
- Add wanakana for English -> katakana conversion
- Add volume control (0%-200%)
- Add message format with '说、' pause marker
- PyInstaller --onedir build for portable exe
- Rate-limit danmaku logging (1/20)
- Fix aiohttp.access log spam
- Add README.md
2026-08-08 13:22:58 +08:00

130 lines
4.3 KiB
Python

"""Node.js TTS Bridge — Persistent Process Manager.
Starts a long-lived Node.js subprocess running tts_bridge.js. Communication uses
stdin/stdout with temporary files for text and WAV data.
Supports PyInstaller frozen mode: uses local node.exe and tts_bridge.js.
"""
import asyncio
import os
import pathlib
import sys
import tempfile
import logging
logger = logging.getLogger(__name__)
def _get_portable_root() -> pathlib.Path:
return pathlib.Path(sys.executable).parent.parent.parent
def _get_node_exe() -> str:
if getattr(sys, "frozen", False):
p = _get_portable_root() / "node.exe"
if p.exists():
return str(p)
return "node"
def _resolve_path(rel: str) -> str:
if getattr(sys, "frozen", False):
p = _get_portable_root() / rel
if p.exists():
return str(p)
return str(pathlib.Path(__file__).parent / rel)
class TTSBridge:
"""Manages a persistent Node.js subprocess for aquestalk.js TTS synthesis."""
def __init__(self, voice: str = "f1", speed: int = 100, bridge_script: str | None = None):
self._voice = voice
self._speed = speed
self._process: asyncio.subprocess.Process | None = None
self._bridge_script = bridge_script or _resolve_path("tts_bridge.js")
self._node_exe = _get_node_exe()
async def start(self) -> None:
if not pathlib.Path(self._bridge_script).exists():
raise FileNotFoundError(f"Bridge script not found: {self._bridge_script}")
logger.info("Starting Node.js TTS bridge (voice=%s, speed=%d)...", self._voice, self._speed)
self._process = await asyncio.create_subprocess_exec(
self._node_exe,
self._bridge_script,
"--voice", self._voice,
"--speed", str(self._speed),
stdin=asyncio.subprocess.PIPE,
stdout=asyncio.subprocess.PIPE,
stderr=asyncio.subprocess.PIPE,
)
line = await asyncio.wait_for(self._process.stdout.readline(), timeout=60)
line_str = line.decode("utf-8").strip()
if "READY" not in line_str:
stderr_data = await self._process.stderr.read()
raise RuntimeError(
f"Bridge failed to start. Got: {line_str}. Stderr: {stderr_data.decode('utf-8', errors='replace')}"
)
logger.info("Node.js TTS bridge ready.")
async def synthesize(self, kana_text: str) -> bytes:
if self._process is None or self._process.stdin is None:
raise RuntimeError("Bridge not started. Call start() first.")
tmpdir = pathlib.Path(tempfile.mkdtemp(prefix="tts_"))
try:
input_path = tmpdir / "input.txt"
output_path = tmpdir / "output.wav"
input_path.write_text(kana_text, encoding="utf-8")
cmd_line = f"{input_path}|{output_path}\n"
self._process.stdin.write(cmd_line.encode("utf-8"))
await self._process.stdin.drain()
line = await asyncio.wait_for(self._process.stdout.readline(), timeout=120)
line_str = line.decode("utf-8").strip()
if line_str.startswith("ERR:"):
raise RuntimeError(f"TTS synthesis error: {line_str[4:]}")
if not line_str.startswith("OK:") or line_str[3:] != str(output_path):
raise RuntimeError(f"Unexpected bridge response: {line_str}")
wav_data = output_path.read_bytes()
return wav_data
finally:
for f in tmpdir.iterdir():
try:
f.unlink()
except OSError:
pass
try:
tmpdir.rmdir()
except OSError:
pass
async def stop(self) -> None:
if self._process is not None:
logger.info("Stopping TTS bridge...")
try:
if self._process.stdin is not None:
self._process.stdin.close()
except OSError:
pass
try:
self._process.terminate()
await asyncio.wait_for(self._process.wait(), timeout=5)
except asyncio.TimeoutError:
self._process.kill()
await self._process.wait()
logger.info("TTS bridge stopped.")
self._process = None