Paperlib MCP

""" 可配置参数模块 支持从环境变量或 config.yaml 读取参数。 优先级: 环境变量 > config.yaml > 默认值 """ import os from pathlib import Path from functools import lru_cache # 默认值 DEFAULTS = { # Claim Grouping "CLAIM_SPLIT_THRESHOLD": 150, # 触发拆分的组大小阈值 "CLAIM_TARGET_SIZE": 120, # 拆分后目标组大小 "CLAIM_MAX_PER_DOC": 100, # 每篇文档最大 claim 数 "CLAIM_TOP_K_PER_GROUP": 10, # 导出时每组 top-k claims # Taxonomy "TAXONOMY_MIN_PRIORITY": 1, # 最小优先级 (越小越优先) "TAXONOMY_GENERAL_THRESHOLD": 0.3, # general 占比警告阈值 # Topic Selection "TOPIC_MIN_DF": 3, # topic 最小文档频率 "TOPIC_SELECTION_STRATEGY": "min_df", # 选择策略: min_df | max_weight # Relation Canonicalization "REL_QUALIFIER_KEYS": "sample_size,time_period,geography,industry", # 保留的 qualifier keys } @lru_cache(maxsize=1) def load_config() -> dict: """加载配置，支持 YAML 文件""" config = DEFAULTS.copy() # 尝试加载 config.yaml config_path = Path(__file__).parent.parent.parent.parent / "config.yaml" if config_path.exists(): try: import yaml with open(config_path, 'r') as f: yaml_config = yaml.safe_load(f) or {} for key, value in yaml_config.items(): key_upper = key.upper() if key_upper in config: config[key_upper] = value except ImportError: pass # yaml not installed except Exception as e: print(f"Warning: Failed to load config.yaml: {e}") # 环境变量覆盖 for key in config: env_value = os.getenv(key) if env_value is not None: # 类型转换 default = DEFAULTS[key] if isinstance(default, int): config[key] = int(env_value) elif isinstance(default, float): config[key] = float(env_value) elif isinstance(default, bool): config[key] = env_value.lower() in ('true', '1', 'yes') else: config[key] = env_value return config def get(key: str, default=None): """获取配置值""" config = load_config() return config.get(key, default) def reload(): """重新加载配置""" load_config.cache_clear() return load_config() # 便捷访问 def claim_split_threshold() -> int: return get("CLAIM_SPLIT_THRESHOLD") def claim_target_size() -> int: return get("CLAIM_TARGET_SIZE") def claim_max_per_doc() -> int: return get("CLAIM_MAX_PER_DOC") def claim_top_k_per_group() -> int: return get("CLAIM_TOP_K_PER_GROUP") def topic_min_df() -> int: return get("TOPIC_MIN_DF") def taxonomy_general_threshold() -> float: return get("TAXONOMY_GENERAL_THRESHOLD") def rel_qualifier_keys() -> list[str]: keys = get("REL_QUALIFIER_KEYS", "") return [k.strip() for k in keys.split(",") if k.strip()]