mirror of
https://github.com/zhouxiaoka/autoclip.git
synced 2026-09-03 06:24:14 +08:00
593cc62bd5
Squashed merge of fix/problem-fixes-from-main. - Desktop client (Tauri v2) wiring + packaging - CI workflow updates (Python 3.11, Rust toolchain, Tauri CLI, WebKit deps) - Backend test fix (test_missing_api_key respects CI-injected env var) - build_backend.py: support CI without venv + Windows-safe ASCII output Desktop build workflows (Linux/Windows/macOS) still failing — tracked in #65.
231 lines
9.9 KiB
Python
231 lines
9.9 KiB
Python
"""
|
||
语音转写配置验证服务
|
||
负责验证配置的有效性和完整性
|
||
"""
|
||
import logging
|
||
from typing import Dict, List, Optional, Tuple
|
||
from pathlib import Path
|
||
from backend.core.desktop_config import SpeechRecognitionSettings, WhisperConfig, ApiConfig
|
||
from backend.services.whisper_model_manager import get_model_manager, ModelStatus
|
||
|
||
logger = logging.getLogger(__name__)
|
||
|
||
|
||
class SpeechConfigValidator:
|
||
"""语音转写配置验证器"""
|
||
|
||
def __init__(self):
|
||
self.model_manager = get_model_manager()
|
||
|
||
def validate_config(self, config: SpeechRecognitionSettings) -> Dict[str, any]:
|
||
"""
|
||
验证语音转写配置
|
||
|
||
Args:
|
||
config: 语音转写配置
|
||
|
||
Returns:
|
||
验证结果字典
|
||
"""
|
||
result = {
|
||
"valid": True,
|
||
"errors": [],
|
||
"warnings": [],
|
||
"recommendations": []
|
||
}
|
||
|
||
# 验证主方法
|
||
method_validation = self._validate_method(config.method)
|
||
if not method_validation["valid"]:
|
||
result["valid"] = False
|
||
result["errors"].extend(method_validation["errors"])
|
||
|
||
# 验证具体配置
|
||
if config.method == "whisper_local":
|
||
whisper_validation = self._validate_whisper_config(config.whisper_config)
|
||
if not whisper_validation["valid"]:
|
||
result["valid"] = False
|
||
result["errors"].extend(whisper_validation["errors"])
|
||
result["warnings"].extend(whisper_validation["warnings"])
|
||
result["recommendations"].extend(whisper_validation["recommendations"])
|
||
|
||
elif config.method in ["openai_api", "azure_speech", "google_speech", "aliyun_speech", "custom_api"]:
|
||
api_validation = self._validate_api_config(config, config.method)
|
||
if not api_validation["valid"]:
|
||
result["valid"] = False
|
||
result["errors"].extend(api_validation["errors"])
|
||
result["warnings"].extend(api_validation["warnings"])
|
||
result["recommendations"].extend(api_validation["recommendations"])
|
||
|
||
# 验证回退配置
|
||
if config.enable_fallback:
|
||
fallback_validation = self._validate_fallback_config(config)
|
||
if not fallback_validation["valid"]:
|
||
result["warnings"].extend(fallback_validation["warnings"])
|
||
|
||
return result
|
||
|
||
def _validate_method(self, method: str) -> Dict[str, any]:
|
||
"""验证方法选择"""
|
||
valid_methods = [
|
||
"whisper_local", "openai_api", "azure_speech",
|
||
"google_speech", "aliyun_speech", "custom_api"
|
||
]
|
||
|
||
if method not in valid_methods:
|
||
return {
|
||
"valid": False,
|
||
"errors": [f"不支持的语音识别方法: {method}"]
|
||
}
|
||
|
||
return {"valid": True, "errors": []}
|
||
|
||
def _validate_whisper_config(self, config: WhisperConfig) -> Dict[str, any]:
|
||
"""验证Whisper配置"""
|
||
result = {"valid": True, "errors": [], "warnings": [], "recommendations": []}
|
||
|
||
# 验证模型名称
|
||
valid_models = ["tiny", "base", "small", "medium", "large"]
|
||
if config.model_name not in valid_models:
|
||
result["valid"] = False
|
||
result["errors"].append(f"不支持的Whisper模型: {config.model_name}")
|
||
|
||
# 检查模型是否已下载
|
||
model_info = self.model_manager.get_model_info(config.model_name)
|
||
if model_info and model_info.status != ModelStatus.DOWNLOADED:
|
||
if model_info.status == ModelStatus.AVAILABLE:
|
||
result["warnings"].append(f"模型 {config.model_name} 未下载，首次使用时会自动下载")
|
||
elif model_info.status == ModelStatus.DOWNLOADING:
|
||
result["warnings"].append(f"模型 {config.model_name} 正在下载中")
|
||
elif model_info.status == ModelStatus.ERROR:
|
||
result["errors"].append(f"模型 {config.model_name} 下载失败")
|
||
|
||
# 验证超时时间
|
||
if config.timeout < 60:
|
||
result["warnings"].append("超时时间过短，建议至少60秒")
|
||
elif config.timeout > 7200:
|
||
result["warnings"].append("超时时间过长，建议不超过2小时")
|
||
|
||
# 验证自定义模型目录
|
||
if config.custom_models_dir:
|
||
custom_dir = Path(config.custom_models_dir)
|
||
if not custom_dir.exists():
|
||
result["errors"].append(f"自定义模型目录不存在: {config.custom_models_dir}")
|
||
elif not custom_dir.is_dir():
|
||
result["errors"].append(f"自定义模型目录不是有效目录: {config.custom_models_dir}")
|
||
|
||
# 添加推荐
|
||
if config.model_name == "tiny":
|
||
result["recommendations"].append("tiny模型速度最快但准确度较低，建议用于实时处理")
|
||
elif config.model_name == "base":
|
||
result["recommendations"].append("base模型是平衡选择，推荐日常使用")
|
||
elif config.model_name in ["small", "medium", "large"]:
|
||
result["recommendations"].append(f"{config.model_name}模型准确度高但速度较慢，适合重要内容")
|
||
|
||
return result
|
||
|
||
def _validate_api_config(self, config: SpeechRecognitionSettings, method: str) -> Dict[str, any]:
|
||
"""验证API配置"""
|
||
result = {"valid": True, "errors": [], "warnings": [], "recommendations": []}
|
||
|
||
# 获取对应的API配置
|
||
if method == "openai_api":
|
||
api_config = config.openai_config
|
||
elif method == "azure_speech":
|
||
api_config = config.azure_config
|
||
elif method == "google_speech":
|
||
api_config = config.google_config
|
||
elif method == "aliyun_speech":
|
||
api_config = config.aliyun_config
|
||
elif method == "custom_api":
|
||
api_config = config.custom_api_config
|
||
else:
|
||
return {"valid": False, "errors": [f"不支持的API方法: {method}"]}
|
||
|
||
# 验证API密钥
|
||
if not api_config.api_key:
|
||
result["valid"] = False
|
||
result["errors"].append(f"{method} API密钥不能为空")
|
||
elif len(api_config.api_key) < 10:
|
||
result["warnings"].append("API密钥长度过短，请检查是否正确")
|
||
|
||
# 验证Azure区域
|
||
if method == "azure_speech" and not api_config.region:
|
||
result["valid"] = False
|
||
result["errors"].append("Azure Speech服务需要指定区域")
|
||
|
||
# 验证自定义API端点
|
||
if method == "custom_api":
|
||
if not api_config.endpoint:
|
||
result["valid"] = False
|
||
result["errors"].append("自定义API需要指定端点URL")
|
||
elif not api_config.endpoint.startswith(("http://", "https://")):
|
||
result["errors"].append("API端点必须是有效的HTTP/HTTPS URL")
|
||
|
||
# 添加推荐
|
||
if method == "openai_api":
|
||
result["recommendations"].append("OpenAI API准确度最高，但需要付费")
|
||
elif method == "azure_speech":
|
||
result["recommendations"].append("Azure Speech适合企业级应用，支持多种语言")
|
||
elif method == "google_speech":
|
||
result["recommendations"].append("Google Speech功能丰富，支持实时识别")
|
||
elif method == "aliyun_speech":
|
||
result["recommendations"].append("阿里云语音识别对中文优化较好")
|
||
|
||
return result
|
||
|
||
def _validate_fallback_config(self, config: SpeechRecognitionSettings) -> Dict[str, any]:
|
||
"""验证回退配置"""
|
||
result = {"valid": True, "warnings": [], "recommendations": []}
|
||
|
||
# 检查回退方法是否与主方法相同
|
||
if config.fallback_method == config.method:
|
||
result["warnings"].append("回退方法与主方法相同，建议选择不同的回退方法")
|
||
|
||
# 检查回退方法是否可用
|
||
if config.fallback_method == "whisper_local":
|
||
model_info = self.model_manager.get_model_info(config.whisper_config.model_name)
|
||
if model_info and model_info.status not in [ModelStatus.DOWNLOADED, ModelStatus.AVAILABLE]:
|
||
result["warnings"].append("回退方法使用的Whisper模型不可用")
|
||
|
||
# 添加推荐
|
||
if config.method != "whisper_local" and config.fallback_method != "whisper_local":
|
||
result["recommendations"].append("建议将Whisper本地模型作为回退方法，确保离线可用")
|
||
|
||
return result
|
||
|
||
def get_config_recommendations(self, config: SpeechRecognitionSettings) -> List[str]:
|
||
"""获取配置建议"""
|
||
recommendations = []
|
||
|
||
# 根据使用场景推荐
|
||
if config.method == "whisper_local":
|
||
if config.whisper_config.model_name == "tiny":
|
||
recommendations.append("tiny模型适合快速处理，但准确度较低")
|
||
elif config.whisper_config.model_name in ["medium", "large"]:
|
||
recommendations.append("大模型准确度高但处理时间长，适合重要内容")
|
||
|
||
# 网络环境建议
|
||
if config.method != "whisper_local":
|
||
recommendations.append("使用API服务需要稳定的网络连接")
|
||
if config.enable_fallback and config.fallback_method == "whisper_local":
|
||
recommendations.append("已配置本地回退，确保离线可用")
|
||
|
||
# 性能建议
|
||
if config.whisper_config.enable_speaker_diarization:
|
||
recommendations.append("说话人分离功能会增加处理时间")
|
||
|
||
return recommendations
|
||
|
||
|
||
# 全局验证器实例
|
||
_validator: Optional[SpeechConfigValidator] = None
|
||
|
||
|
||
def get_config_validator() -> SpeechConfigValidator:
|
||
"""获取配置验证器实例"""
|
||
global _validator
|
||
if _validator is None:
|
||
_validator = SpeechConfigValidator()
|
||
return _validator
|