build/pyinstaller/hooks/hook-voice_dialogue.py

"""
Voice Dialogue PyInstaller Hook
用于配置 Voice Dialogue 应用程序的 PyInstaller 打包规则
"""

import pathlib

from PyInstaller.utils.hooks import (
    collect_submodules,
    collect_data_files,
    collect_system_data_files,
    collect_dynamic_libs
)

# ============================================================================
# 路径配置
# ============================================================================
PROJECT_ROOT = pathlib.Path(__file__).parent.parent.parent.parent
ASSETS_ROOT = PROJECT_ROOT / "assets"

# ============================================================================
# 核心模块收集
# ============================================================================
# 收集主模块的所有子模块
hiddenimports = collect_submodules('voice_dialogue')
datas = collect_data_files('moyoyo_tts', include_py_files=True)
# 收集系统资源文件
datas += collect_system_data_files(ASSETS_ROOT.as_posix(), "assets")

# ============================================================================
# 第三方依赖配置
# ============================================================================

# AI/ML 相关依赖
ML_DEPENDENCIES = [
    "llama_cpp",
    "torch",
    "transformers",
    "pytorch_lightning",
    "huggingface_hub",
    "einops",
]

# 语音处理相关依赖
AUDIO_DEPENDENCIES = [
    "librosa",
    "soundfile",
    "funasr_onnx",
    "pywhispercpp",
    "kokoro_onnx",
    "playsound",
    "pyaudio",
    "silero_vad"
]

# 文本处理相关依赖
TEXT_PROCESSING_DEPENDENCIES = [
    "jieba",
    "jieba_fast",
    "cn2an",
    "g2p_en",
    "pypinyin",
    "misaki",
    "opencc",
    "wordsegment",
    "LangSegment",
    "en_core_web_sm",
    "spacy",
    "spacy_curated_transformers",
]

# Web/网络相关依赖
WEB_DEPENDENCIES = [
    "fastapi",
    "uvicorn",
    "websockets",
    "langchain",
    "langchain_community",
]

# 系统相关依赖
SYSTEM_DEPENDENCIES = [
    "AppKit",  # macOS
    "Foundation",  # macOS
    "ffmpeg",
]

# 其他依赖
OTHER_DEPENDENCIES = [
    'loguru'
]

# 合并所有隐藏导入
hiddenimports.extend(
    ML_DEPENDENCIES +
    AUDIO_DEPENDENCIES +
    TEXT_PROCESSING_DEPENDENCIES +
    WEB_DEPENDENCIES +
    SYSTEM_DEPENDENCIES +
    OTHER_DEPENDENCIES
)

# ============================================================================
# 数据文件收集
# ============================================================================

# 需要收集数据文件的包配置
DATA_PACKAGES = [
    ("jieba_fast", {}),
    ("kokoro_onnx", {}),
    ("language_tags", {}),
    ("py3langid", {}),
    ("inflect", {"include_py_files": True}),
    ("g2p_en", {}),
    ("wordsegment", {}),
    ("espeakng_loader", {}),
    ("spacy", {"include_py_files": True}),
    ("misaki", {}),
    ("silero_vad", {}),
]

# 收集数据文件
for package_name, kwargs in DATA_PACKAGES:
    try:
        datas += collect_data_files(package_name, **kwargs)
    except Exception as e:
        print(f"警告：无法收集包 {package_name} 的数据文件: {e}")

# ============================================================================
# 动态库收集
# ============================================================================

# 需要收集动态库的包
BINARY_PACKAGES = [
    "llama_cpp",
    "pyaudio"
]

binaries = []
for package_name in BINARY_PACKAGES:
    try:
        binaries += collect_dynamic_libs(package_name)
    except Exception as e:
        print(f"警告：无法收集包 {package_name} 的动态库: {e}")