128 lines
5.1 KiB
Python
128 lines
5.1 KiB
Python
from typing import List, Union, Optional
|
||
from pydantic import AnyHttpUrl, validator
|
||
from pydantic_settings import BaseSettings
|
||
|
||
|
||
class Settings(BaseSettings):
|
||
PROJECT_NAME: str = "DeepAudit"
|
||
API_V1_STR: str = "/api/v1"
|
||
|
||
# SECURITY
|
||
SECRET_KEY: str = "changethis_in_production_to_a_long_random_string"
|
||
ALGORITHM: str = "HS256"
|
||
ACCESS_TOKEN_EXPIRE_MINUTES: int = 60 * 24 * 8 # 8 days
|
||
|
||
# CORS
|
||
BACKEND_CORS_ORIGINS: List[AnyHttpUrl] = []
|
||
|
||
@validator("BACKEND_CORS_ORIGINS", pre=True)
|
||
def assemble_cors_origins(cls, v: Union[str, List[str]]) -> Union[List[str], str]:
|
||
if isinstance(v, str) and not v.startswith("["):
|
||
return [i.strip() for i in v.split(",")]
|
||
elif isinstance(v, (list, str)):
|
||
return v
|
||
raise ValueError(v)
|
||
|
||
# POSTGRES
|
||
POSTGRES_SERVER: str = "db"
|
||
POSTGRES_USER: str = "postgres"
|
||
POSTGRES_PASSWORD: str = "postgres"
|
||
POSTGRES_DB: str = "deepaudit"
|
||
DATABASE_URL: str | None = None
|
||
|
||
@validator("DATABASE_URL", pre=True)
|
||
def assemble_db_connection(cls, v: str | None, values: dict[str, any]) -> str:
|
||
if isinstance(v, str):
|
||
return v
|
||
return str(f"postgresql+asyncpg://{values.get('POSTGRES_USER')}:{values.get('POSTGRES_PASSWORD')}@{values.get('POSTGRES_SERVER')}/{values.get('POSTGRES_DB')}")
|
||
|
||
# LLM配置
|
||
LLM_PROVIDER: str = "openai" # gemini, openai, claude, qwen, deepseek, zhipu, moonshot, baidu, minimax, doubao, ollama
|
||
LLM_API_KEY: Optional[str] = None
|
||
LLM_MODEL: Optional[str] = None # 不指定时使用provider的默认模型
|
||
LLM_BASE_URL: Optional[str] = None # 自定义API端点（如中转站）
|
||
LLM_TIMEOUT: int = 300 # 超时时间（秒）
|
||
LLM_TEMPERATURE: float = 0.1
|
||
LLM_MAX_TOKENS: int = 4096
|
||
|
||
# 各LLM提供商的API Key配置（兼容单独配置）
|
||
OPENAI_API_KEY: Optional[str] = None
|
||
OPENAI_BASE_URL: Optional[str] = None
|
||
GEMINI_API_KEY: Optional[str] = None
|
||
CLAUDE_API_KEY: Optional[str] = None
|
||
QWEN_API_KEY: Optional[str] = None
|
||
DEEPSEEK_API_KEY: Optional[str] = None
|
||
ZHIPU_API_KEY: Optional[str] = None
|
||
MOONSHOT_API_KEY: Optional[str] = None
|
||
BAIDU_API_KEY: Optional[str] = None # 格式: api_key:secret_key
|
||
MINIMAX_API_KEY: Optional[str] = None
|
||
DOUBAO_API_KEY: Optional[str] = None
|
||
OLLAMA_BASE_URL: Optional[str] = "http://localhost:11434/v1"
|
||
|
||
# GitHub配置
|
||
GITHUB_TOKEN: Optional[str] = None
|
||
|
||
# GitLab配置
|
||
GITLAB_TOKEN: Optional[str] = None
|
||
|
||
# Gitea配置
|
||
GITEA_TOKEN: Optional[str] = None
|
||
GITEA_HOST_URL: Optional[str] = "http://localhost:3000"
|
||
GITEA_BOT_TOKEN: Optional[str] = None
|
||
GITEA_WEBHOOK_SECRET: Optional[str] = None
|
||
|
||
# 扫描配置
|
||
MAX_ANALYZE_FILES: int = 0 # 最大分析文件数，0表示无限制
|
||
MAX_FILE_SIZE_BYTES: int = 200 * 1024 # 最大文件大小 200KB
|
||
LLM_CONCURRENCY: int = 10 # LLM并发数
|
||
LLM_GAP_MS: int = 0 # LLM请求间隔（毫秒）
|
||
|
||
# ZIP文件存储配置
|
||
ZIP_STORAGE_PATH: str = "./uploads/zip_files" # ZIP文件存储目录
|
||
|
||
# 输出语言配置 - 支持 zh-CN（中文）和 en-US（英文）
|
||
OUTPUT_LANGUAGE: str = "zh-CN"
|
||
|
||
# ============ Agent 模块配置 ============
|
||
|
||
# 嵌入模型配置（独立于 LLM 配置）
|
||
EMBEDDING_PROVIDER: str = "openai" # openai, azure, ollama, cohere, huggingface, jina, qwen
|
||
EMBEDDING_MODEL: str = "text-embedding-3-small"
|
||
EMBEDDING_API_KEY: Optional[str] = None # 嵌入模型专用 API Key（留空则使用 LLM_API_KEY）
|
||
EMBEDDING_BASE_URL: Optional[str] = None # 嵌入模型专用 Base URL（留空使用提供商默认地址）
|
||
EMBEDDING_DIMENSION: int = 0 # 嵌入模型维度（0 表示自动检测或由代码逻辑根据模型确定）
|
||
|
||
# 向量数据库配置
|
||
VECTOR_DB_PATH: str = "./data/vector_db" # 向量数据库持久化目录
|
||
|
||
# SSH配置
|
||
SSH_CONFIG_PATH: str = "./data/ssh" # SSH配置目录（存储known_hosts等）
|
||
SSH_CLONE_TIMEOUT: int = 300 # SSH克隆超时时间（秒）
|
||
SSH_TEST_TIMEOUT: int = 15 # SSH测试连接超时时间（秒）
|
||
SSH_CONNECT_TIMEOUT: int = 10 # SSH连接超时时间（秒）
|
||
|
||
# Agent 配置
|
||
AGENT_MAX_ITERATIONS: int = 50 # Agent 最大迭代次数
|
||
AGENT_TOKEN_BUDGET: int = 100000 # Agent Token 预算
|
||
AGENT_TIMEOUT_SECONDS: int = 1800 # Agent 超时时间（30分钟）
|
||
|
||
# 沙箱配置（必须）
|
||
SANDBOX_IMAGE: str = "deepaudit/sandbox:latest" # 沙箱 Docker 镜像
|
||
SANDBOX_MEMORY_LIMIT: str = "512m" # 沙箱内存限制
|
||
SANDBOX_CPU_LIMIT: float = 1.0 # 沙箱 CPU 限制
|
||
SANDBOX_TIMEOUT: int = 60 # 沙箱命令超时（秒）
|
||
SANDBOX_NETWORK_MODE: str = "none" # 沙箱网络模式 (none, bridge)
|
||
|
||
# RAG 配置
|
||
RAG_CHUNK_SIZE: int = 1500 # 代码块大小（Token）
|
||
RAG_CHUNK_OVERLAP: int = 50 # 代码块重叠（Token）
|
||
RAG_TOP_K: int = 10 # 检索返回数量
|
||
|
||
class Config:
|
||
case_sensitive = True
|
||
env_file = ".env"
|
||
extra = "ignore" # 忽略额外的环境变量（如 VITE_* 前端变量）
|
||
|
||
|
||
settings = Settings()
|