wsy182
6dee2a2ff3
- 调整视频处理频率从每帧处理改为每2帧处理 - 添加膝角趋势平滑算法减少单帧抖动误判 - 改进对角伸展检测逻辑支持准备位手臂上举 - 优化状态机确保严格回到准备姿态才计数 - 添加姿态丢失时的候选帧清理机制 - 更新音频文件生成路径至resources目录 - 改进macOS音频生成使用AIFF格式提高质量 - 添加详细的帧处理日志输出间隔配置
130 lines
3.3 KiB
Python
130 lines
3.3 KiB
Python
# app/audio/generate.py
|
|
from __future__ import annotations
|
|
|
|
import platform
|
|
import shutil
|
|
import subprocess
|
|
from pathlib import Path
|
|
|
|
from loguru import logger
|
|
|
|
|
|
def generate_rep_audio_files(
|
|
*,
|
|
max_count: int,
|
|
rate: int,
|
|
output_dir: Path,
|
|
overwrite: bool = False,
|
|
) -> None:
|
|
"""
|
|
确保 0~max_count 的运动次数语音 wav 文件存在。
|
|
|
|
默认生成到：
|
|
|
|
resources/audio/reps/0.aiff # macOS
|
|
resources/audio/reps/0.wav # Windows / Linux
|
|
...
|
|
resources/audio/reps/200.aiff 或 200.wav
|
|
|
|
服务启动时调用一次即可。
|
|
"""
|
|
output_dir.mkdir(parents=True, exist_ok=True)
|
|
system = platform.system().lower()
|
|
suffix = ".aiff" if system == "darwin" else ".wav"
|
|
|
|
missing_counts = [
|
|
count
|
|
for count in range(0, max_count + 1)
|
|
if overwrite or not _audio_path(output_dir, count, suffix=suffix).exists()
|
|
]
|
|
|
|
if not missing_counts:
|
|
logger.info("Rep audio files already prepared: {}", output_dir)
|
|
return
|
|
|
|
logger.info(
|
|
"Preparing rep audio files, system={}, count={}, output_dir={}",
|
|
system,
|
|
len(missing_counts),
|
|
output_dir,
|
|
)
|
|
|
|
if system == "darwin":
|
|
_generate_with_macos_say(
|
|
counts=missing_counts,
|
|
output_dir=output_dir,
|
|
rate=rate,
|
|
)
|
|
else:
|
|
_generate_with_pyttsx3(
|
|
counts=missing_counts,
|
|
output_dir=output_dir,
|
|
rate=rate,
|
|
)
|
|
|
|
logger.info("Rep audio files prepared: {}", output_dir)
|
|
|
|
|
|
def _generate_with_macos_say(
|
|
*,
|
|
counts: list[int],
|
|
output_dir: Path,
|
|
rate: int,
|
|
) -> None:
|
|
"""macOS 使用 say 命令生成 wav。"""
|
|
if platform.system().lower() != "darwin":
|
|
raise RuntimeError("say command is only available on macOS")
|
|
|
|
if shutil.which("say") is None:
|
|
raise RuntimeError("macOS say command not found")
|
|
|
|
for count in counts:
|
|
audio_file = _audio_path(output_dir, count, suffix=".aiff")
|
|
|
|
try:
|
|
subprocess.run(
|
|
[
|
|
"say",
|
|
"-r",
|
|
str(rate),
|
|
"--file-format=AIFF",
|
|
"-o",
|
|
str(audio_file),
|
|
str(count),
|
|
],
|
|
stdout=subprocess.DEVNULL,
|
|
stderr=subprocess.PIPE,
|
|
text=True,
|
|
check=True,
|
|
)
|
|
except subprocess.CalledProcessError as exc:
|
|
message = exc.stderr.strip() or f"exit status {exc.returncode}"
|
|
raise RuntimeError(f"Failed to generate {audio_file}: {message}") from exc
|
|
|
|
|
|
def _generate_with_pyttsx3(
|
|
*,
|
|
counts: list[int],
|
|
output_dir: Path,
|
|
rate: int,
|
|
) -> None:
|
|
"""Windows / Linux 使用 pyttsx3 生成 wav。"""
|
|
try:
|
|
import pyttsx3
|
|
except Exception as exc:
|
|
raise RuntimeError(f"pyttsx3 unavailable: {exc}") from exc
|
|
|
|
engine = pyttsx3.init()
|
|
engine.setProperty("rate", rate)
|
|
engine.setProperty("volume", 1.0)
|
|
|
|
for count in counts:
|
|
audio_file = _audio_path(output_dir, count, suffix=".wav")
|
|
engine.save_to_file(str(count), str(audio_file))
|
|
|
|
engine.runAndWait()
|
|
|
|
|
|
def _audio_path(output_dir: Path, count: int, *, suffix: str) -> Path:
|
|
return output_dir / f"{count}{suffix}"
|