1. 统一为使用基于pydantic的.env环境变量管理配置
2. 全项目基于loguru进行日志管理
This commit is contained in:
Doiiars
@@ -12,15 +12,15 @@ from .text_processing import (
|
||||
format_search_results_for_prompt
|
||||
)
|
||||
|
||||
from .config import Config, load_config
|
||||
from .config import Settings, settings
|
||||
|
||||
__all__ = [
|
||||
"clean_json_tags",
|
||||
"clean_markdown_tags",
|
||||
"remove_reasoning_from_output",
|
||||
"remove_reasoning_from_output",
|
||||
"extract_clean_response",
|
||||
"update_state_with_search_results",
|
||||
"format_search_results_for_prompt",
|
||||
"Config",
|
||||
"load_config"
|
||||
"Settings",
|
||||
"settings"
|
||||
]
|
||||
|
||||
+75
-149
@@ -1,157 +1,83 @@
|
||||
"""
|
||||
Configuration management module for the Media Engine.
|
||||
Configuration management module for the Media Engine (pydantic_settings style).
|
||||
"""
|
||||
|
||||
import os
|
||||
from dataclasses import dataclass
|
||||
from pathlib import Path
|
||||
from pydantic_settings import BaseSettings
|
||||
from pydantic import Field
|
||||
from typing import Optional
|
||||
|
||||
|
||||
def _get_value(source, key: str, default=None, *fallback_keys: str):
|
||||
candidates = (key,) + fallback_keys
|
||||
value = None
|
||||
for candidate in candidates:
|
||||
if isinstance(source, dict):
|
||||
value = source.get(candidate)
|
||||
else:
|
||||
value = getattr(source, candidate, None)
|
||||
if value not in (None, ""):
|
||||
break
|
||||
if value in (None, ""):
|
||||
for candidate in candidates:
|
||||
env_val = os.getenv(candidate)
|
||||
if env_val not in (None, ""):
|
||||
value = env_val
|
||||
break
|
||||
return value if value not in (None, "") else default
|
||||
# 计算 .env 优先级：优先当前工作目录，其次项目根目录
|
||||
PROJECT_ROOT: Path = Path(__file__).resolve().parents[2]
|
||||
CWD_ENV: Path = Path.cwd() / ".env"
|
||||
ENV_FILE: str = str(CWD_ENV if CWD_ENV.exists() else (PROJECT_ROOT / ".env"))
|
||||
|
||||
class Settings(BaseSettings):
|
||||
"""
|
||||
全局配置；支持 .env 和环境变量自动加载。
|
||||
变量名与原 config.py 大写一致，便于平滑过渡。
|
||||
"""
|
||||
# ====================== 数据库配置 ======================
|
||||
DB_HOST: str = Field("your_db_host", description="数据库主机，例如localhost 或 127.0.0.1。我们也提供云数据库资源便捷配置，日均10w+数据，可免费申请，联系我们：670939375@qq.com NOTE：为进行数据合规性审查与服务升级，云数据库自2025年10月1日起暂停接收新的使用申请")
|
||||
DB_PORT: int = Field(3306, description="数据库端口号，默认为3306")
|
||||
DB_USER: str = Field("your_db_user", description="数据库用户名")
|
||||
DB_PASSWORD: str = Field("your_db_password", description="数据库密码")
|
||||
DB_NAME: str = Field("your_db_name", description="数据库名称")
|
||||
DB_CHARSET: str = Field("utf8mb4", description="数据库字符集，推荐utf8mb4，兼容emoji")
|
||||
DB_DIALECT: str = Field("mysql", description="数据库类型，例如 'mysql' 或 'postgresql'。用于支持多种数据库后端（如 SQLAlchemy，请与连接信息共同配置）")
|
||||
|
||||
# ======================= LLM 相关 =======================
|
||||
INSIGHT_ENGINE_API_KEY: str = Field(None, description="Insight Agent（推荐Kimi，https://platform.moonshot.cn/）API密钥，用于主LLM。您可以更改每个部分LLM使用的API，🚩只要兼容OpenAI请求格式都可以，定义好KEY、BASE_URL与MODEL_NAME即可正常使用。重要提醒：我们强烈推荐您先使用推荐的配置申请API，先跑通再进行您的更改！")
|
||||
INSIGHT_ENGINE_BASE_URL: Optional[str] = Field("https://api.moonshot.cn/v1", description="Insight Agent LLM接口BaseUrl，可自定义厂商API")
|
||||
INSIGHT_ENGINE_MODEL_NAME: str = Field("kimi-k2-0711-preview", description="Insight Agent LLM模型名称，如kimi-k2-0711-preview")
|
||||
|
||||
MEDIA_ENGINE_API_KEY: str = Field(None, description="Media Agent（推荐Gemini，这里我用了一个中转厂商，你也可以换成你自己的，申请地址：https://www.chataiapi.com/）API密钥")
|
||||
MEDIA_ENGINE_BASE_URL: Optional[str] = Field("https://www.chataiapi.com/v1", description="Media Agent LLM接口BaseUrl")
|
||||
MEDIA_ENGINE_MODEL_NAME: str = Field("gemini-2.5-pro", description="Media Agent LLM模型名称，如gemini-2.5-pro")
|
||||
|
||||
BOCHA_WEB_SEARCH_API_KEY: Optional[str] = Field(None, description="Bocha Web Search API Key")
|
||||
BOCHA_API_KEY: Optional[str] = Field(None, description="Bocha 兼容键（别名）")
|
||||
|
||||
SEARCH_TIMEOUT: int = Field(240, description="搜索超时（秒）")
|
||||
SEARCH_CONTENT_MAX_LENGTH: int = Field(20000, description="用于提示的最长内容长度")
|
||||
MAX_REFLECTIONS: int = Field(2, description="最大反思轮数")
|
||||
MAX_PARAGRAPHS: int = Field(5, description="最大段落数")
|
||||
|
||||
MINDSPIDER_API_KEY: Optional[str] = Field(None, description="MindSpider API密钥")
|
||||
MINDSPIDER_BASE_URL: Optional[str] = Field("https://api.deepseek.com", description="MindSpider LLM接口BaseUrl")
|
||||
MINDSPIDER_MODEL_NAME: str = Field("deepseek-reasoner", description="MindSpider LLM模型名称，如deepseek-reasoner")
|
||||
|
||||
OUTPUT_DIR: str = Field("reports", description="输出目录")
|
||||
SAVE_INTERMEDIATE_STATES: bool = Field(True, description="是否保存中间状态")
|
||||
|
||||
|
||||
QUERY_ENGINE_API_KEY: str = Field(None, description="Query Agent（推荐DeepSeek，https://www.deepseek.com/）API密钥")
|
||||
QUERY_ENGINE_BASE_URL: Optional[str] = Field("https://api.deepseek.com", description="Query Agent LLM接口BaseUrl")
|
||||
QUERY_ENGINE_MODEL_NAME: str = Field("deepseek-reasoner", description="Query Agent LLM模型，如deepseek-reasoner")
|
||||
|
||||
REPORT_ENGINE_API_KEY: str = Field(None, description="Report Agent（推荐Gemini，这里我用了一个中转厂商，你也可以换成你自己的，申请地址：https://www.chataiapi.com/）API密钥")
|
||||
REPORT_ENGINE_BASE_URL: Optional[str] = Field("https://www.chataiapi.com/v1", description="Report Agent LLM接口BaseUrl")
|
||||
REPORT_ENGINE_MODEL_NAME: str = Field("gemini-2.5-pro", description="Report Agent LLM模型，如gemini-2.5-pro")
|
||||
|
||||
FORUM_HOST_API_KEY: str = Field(None, description="Forum Host（Qwen3最新模型，这里我使用了硅基流动这个平台，申请地址：https://cloud.siliconflow.cn/）API密钥")
|
||||
FORUM_HOST_BASE_URL: Optional[str] = Field("https://api.siliconflow.cn/v1", description="Forum Host LLM BaseUrl")
|
||||
FORUM_HOST_MODEL_NAME: str = Field("Qwen/Qwen3-235B-A22B-Instruct-2507", description="Forum Host LLM模型名，如Qwen/Qwen3-235B-A22B-Instruct-2507")
|
||||
|
||||
KEYWORD_OPTIMIZER_API_KEY: str = Field(None, description="SQL keyword Optimizer（小参数Qwen3模型，这里我使用了硅基流动这个平台，申请地址：https://cloud.siliconflow.cn/）API密钥")
|
||||
KEYWORD_OPTIMIZER_BASE_URL: Optional[str] = Field("https://api.siliconflow.cn/v1", description="Keyword Optimizer BaseUrl")
|
||||
KEYWORD_OPTIMIZER_MODEL_NAME: str = Field("Qwen/Qwen3-30B-A3B-Instruct-2507", description="Keyword Optimizer LLM模型名称，如Qwen/Qwen3-30B-A3B-Instruct-2507")
|
||||
|
||||
# ================== 网络工具配置 ====================
|
||||
TAVILY_API_KEY: str = Field(None, description="Tavily API（申请地址：https://www.tavily.com/）API密钥，用于Tavily网络搜索")
|
||||
BOCHA_BASE_URL: Optional[str] = Field("https://api.bochaai.com/v1/ai-search", description="Bocha AI 搜索BaseUrl或博查网页搜索BaseUrl")
|
||||
BOCHA_WEB_SEARCH_API_KEY: str = Field(None, description="Bocha API（申请地址：https://open.bochaai.com/）API密钥，用于Bocha搜索")
|
||||
|
||||
class Config:
|
||||
env_file = ENV_FILE
|
||||
env_prefix = ""
|
||||
case_sensitive = False
|
||||
extra = "allow"
|
||||
|
||||
|
||||
@dataclass
|
||||
class Config:
|
||||
"""Media Engine configuration."""
|
||||
|
||||
llm_api_key: Optional[str] = None
|
||||
llm_base_url: Optional[str] = None
|
||||
llm_model_name: Optional[str] = None
|
||||
llm_provider: Optional[str] = None # compatibility
|
||||
|
||||
bocha_api_key: Optional[str] = None
|
||||
|
||||
search_timeout: int = 240
|
||||
max_content_length: int = 20000
|
||||
max_reflections: int = 2
|
||||
max_paragraphs: int = 5
|
||||
|
||||
output_dir: str = "reports"
|
||||
save_intermediate_states: bool = True
|
||||
|
||||
def __post_init__(self):
|
||||
if not self.llm_provider and self.llm_model_name:
|
||||
self.llm_provider = self.llm_model_name
|
||||
|
||||
def validate(self) -> bool:
|
||||
if not self.llm_api_key:
|
||||
print("错误: Media Engine LLM API Key 未设置 (MEDIA_ENGINE_API_KEY)。")
|
||||
return False
|
||||
if not self.llm_model_name:
|
||||
print("错误: Media Engine 模型名称未设置 (MEDIA_ENGINE_MODEL_NAME)。")
|
||||
return False
|
||||
if not self.bocha_api_key:
|
||||
print("错误: Bocha API Key 未设置 (BOCHA_WEB_SEARCH_API_KEY)。")
|
||||
return False
|
||||
return True
|
||||
|
||||
@classmethod
|
||||
def from_file(cls, config_file: str) -> "Config":
|
||||
if config_file.endswith(".py"):
|
||||
import importlib.util
|
||||
|
||||
spec = importlib.util.spec_from_file_location("config", config_file)
|
||||
config_module = importlib.util.module_from_spec(spec)
|
||||
spec.loader.exec_module(config_module)
|
||||
|
||||
return cls(
|
||||
llm_api_key=_get_value(config_module, "MEDIA_ENGINE_API_KEY"),
|
||||
llm_base_url=_get_value(config_module, "MEDIA_ENGINE_BASE_URL"),
|
||||
llm_model_name=_get_value(config_module, "MEDIA_ENGINE_MODEL_NAME"),
|
||||
bocha_api_key=_get_value(
|
||||
config_module,
|
||||
"BOCHA_WEB_SEARCH_API_KEY",
|
||||
None,
|
||||
"BOCHA_API_KEY",
|
||||
),
|
||||
search_timeout=int(_get_value(config_module, "SEARCH_TIMEOUT", 240)),
|
||||
max_content_length=int(_get_value(config_module, "SEARCH_CONTENT_MAX_LENGTH", 20000)),
|
||||
max_reflections=int(_get_value(config_module, "MAX_REFLECTIONS", 2)),
|
||||
max_paragraphs=int(_get_value(config_module, "MAX_PARAGRAPHS", 5)),
|
||||
output_dir=_get_value(config_module, "OUTPUT_DIR", "reports"),
|
||||
save_intermediate_states=str(
|
||||
_get_value(config_module, "SAVE_INTERMEDIATE_STATES", "true")
|
||||
).lower()
|
||||
in ("true", "1", "yes"),
|
||||
)
|
||||
|
||||
config_dict = {}
|
||||
if os.path.exists(config_file):
|
||||
with open(config_file, "r", encoding="utf-8") as f:
|
||||
for line in f:
|
||||
line = line.strip()
|
||||
if line and not line.startswith("#") and "=" in line:
|
||||
key, value = line.split("=", 1)
|
||||
config_dict[key.strip()] = value.strip()
|
||||
|
||||
return cls(
|
||||
llm_api_key=_get_value(config_dict, "MEDIA_ENGINE_API_KEY"),
|
||||
llm_base_url=_get_value(config_dict, "MEDIA_ENGINE_BASE_URL"),
|
||||
llm_model_name=_get_value(config_dict, "MEDIA_ENGINE_MODEL_NAME"),
|
||||
bocha_api_key=_get_value(
|
||||
config_dict,
|
||||
"BOCHA_WEB_SEARCH_API_KEY",
|
||||
None,
|
||||
"BOCHA_API_KEY",
|
||||
),
|
||||
search_timeout=int(_get_value(config_dict, "SEARCH_TIMEOUT", 240)),
|
||||
max_content_length=int(_get_value(config_dict, "SEARCH_CONTENT_MAX_LENGTH", 20000)),
|
||||
max_reflections=int(_get_value(config_dict, "MAX_REFLECTIONS", 2)),
|
||||
max_paragraphs=int(_get_value(config_dict, "MAX_PARAGRAPHS", 5)),
|
||||
output_dir=_get_value(config_dict, "OUTPUT_DIR", "reports"),
|
||||
save_intermediate_states=str(
|
||||
_get_value(config_dict, "SAVE_INTERMEDIATE_STATES", "true")
|
||||
).lower()
|
||||
in ("true", "1", "yes"),
|
||||
)
|
||||
|
||||
|
||||
def load_config(config_file: Optional[str] = None) -> Config:
|
||||
if config_file:
|
||||
if not os.path.exists(config_file):
|
||||
raise FileNotFoundError(f"配置文件不存在: {config_file}")
|
||||
file_to_load = config_file
|
||||
else:
|
||||
for candidate in ("config.py", "config.env", ".env"):
|
||||
if os.path.exists(candidate):
|
||||
file_to_load = candidate
|
||||
print(f"已找到配置文件: {candidate}")
|
||||
break
|
||||
else:
|
||||
raise FileNotFoundError("未找到配置文件，请创建 config.py。")
|
||||
|
||||
config = Config.from_file(file_to_load)
|
||||
if not config.validate():
|
||||
raise ValueError("配置校验失败，请检查 config.py 中的相关配置。")
|
||||
return config
|
||||
|
||||
|
||||
def print_config(config: Config):
|
||||
print("\n=== Media Engine 配置 ===")
|
||||
print(f"LLM 模型: {config.llm_model_name}")
|
||||
print(f"LLM Base URL: {config.llm_base_url or '(默认)'}")
|
||||
print(f"Bocha API Key: {'已配置' if config.bocha_api_key else '未配置'}")
|
||||
print(f"搜索超时: {config.search_timeout} 秒")
|
||||
print(f"最长内容长度: {config.max_content_length}")
|
||||
print(f"最大反思次数: {config.max_reflections}")
|
||||
print(f"最大段落数: {config.max_paragraphs}")
|
||||
print(f"输出目录: {config.output_dir}")
|
||||
print(f"保存中间状态: {config.save_intermediate_states}")
|
||||
print(f"LLM API Key: {'已配置' if config.llm_api_key else '未配置'}")
|
||||
print("========================\n")
|
||||
settings = Settings()
|
||||
|
||||
Reference in New Issue
Block a user