from __future__ import annotations from functools import lru_cache from pathlib import Path from typing import Literal from pydantic import Field, field_validator, model_validator from pydantic_settings import BaseSettings, SettingsConfigDict _PROJECT_ROOT = Path(__file__).resolve().parents[1] class Settings(BaseSettings): model_config = SettingsConfigDict( env_file=_PROJECT_ROOT / ".env", env_prefix="RAG_", extra="ignore", ) data_dir: Path = Path("data") storage_dir: Path = Path("storage/qdrant") export_dir: Path = Path("exports") qdrant_collection: str = "rag_chunks" max_upload_bytes: int = Field(default=25 * 1024 * 1024, ge=1024) chunk_size: int = Field(default=1500, ge=500) chunk_overlap: int = Field(default=200, ge=0) top_k: int = Field(default=5, ge=1, le=64) retrieval_mode: Literal["dense", "fusion"] = "dense" retrieval_candidate_k: int = Field(default=50, ge=1, le=256) retrieval_dense_weight: float = Field(default=0.25, ge=0.0, le=1.0) retrieval_max_chunks_per_page: int = Field(default=1, ge=1, le=64) retrieval_fallback_to_dense: bool = True retrieval_telemetry_enabled: bool = False retrieval_shadow_sample_rate: float = Field(default=0.0, ge=0.0, le=1.0) retrieval_telemetry_max_bytes: int = Field(default=5 * 1024 * 1024, ge=1024) retrieval_telemetry_retained_events: int = Field(default=5000, ge=100) embedding_warmup_enabled: bool = False llm_provider: Literal["gemini"] = "gemini" llm_model: str = "gemini-flash-lite-latest" llm_temperature: float = Field(default=0.1, ge=0.0, le=2.0) llm_max_new_tokens: int = Field(default=10000, ge=1, le=20000) embedding_provider: Literal["local"] = "local" embedding_model: str = "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2" gemini_api_key: str | None = Field(default=None, validation_alias="GEMINI_API_KEY") api_key: str | None = None api_key_sha256: str | None = Field(default=None, pattern=r"^[0-9a-f]{64}$") gradio_username: str | None = None gradio_password: str | None = None server_name: str = "127.0.0.1" server_port: int = Field(default=7860, ge=1, le=65535) summarize_batch_size: int = Field(default=10, ge=1) summarize_retrieval_k: int = Field(default=12, ge=1, le=128) generation_retrieval_k: int = Field(default=16, ge=1, le=128) quiz_default_count: int = Field(default=8, ge=1, le=50) flashcards_default_count: int = Field(default=15, ge=1, le=100) @field_validator( "gemini_api_key", "api_key", "api_key_sha256", "gradio_username", "gradio_password", mode="before", ) @classmethod def normalize_blank_secrets(cls, value: object) -> object: """Treat blank optional credentials as disabled configuration.""" if isinstance(value, str) and not value.strip(): return None return value @model_validator(mode="after") def validate_config(self) -> Settings: if self.chunk_overlap >= self.chunk_size: raise ValueError("chunk_overlap must be smaller than chunk_size.") if bool(self.gradio_username) != bool(self.gradio_password): raise ValueError( "RAG_GRADIO_USERNAME and RAG_GRADIO_PASSWORD must be configured together." ) if self.retrieval_shadow_sample_rate > 0 and not self.retrieval_telemetry_enabled: raise ValueError( "RAG_RETRIEVAL_TELEMETRY_ENABLED must be true when shadow sampling is enabled." ) for field_name in ("data_dir", "storage_dir", "export_dir"): path = getattr(self, field_name) if not path.is_absolute(): object.__setattr__(self, field_name, _PROJECT_ROOT / path) return self @lru_cache(maxsize=1) def get_settings() -> Settings: return Settings() settings = get_settings()