backend/settings.py

from pathlib import Path
from pydantic import BaseModel
import os

class Settings(BaseModel):
    ROOT: Path = Path(".")
    DATA_DIR: Path = ROOT / "backend" / "data"
    RAW_DIR: Path = DATA_DIR / "raw"
    PROCESSED_DIR: Path = DATA_DIR / "processed"
    INDEX_DIR: Path = ROOT / "backend" / "index"
    MODEL_DIR: Path = ROOT / "backend" / "models"


    EMBEDDING_MODEL_ID: str = "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2" 


    RERANK_MODEL_ID: str = "BAAI/bge-reranker-v2-m3"


    LLM_REPO_ID: str = "Qwen/Qwen2.5-1.5B-Instruct-GGUF"
    LLM_FILENAME_PATTERN: str = "Qwen2.5-1.5B-Instruct-Q4_K_M.gguf"


    LLM_PATH: Path = MODEL_DIR / "qwen2.5-1.5b-gguf" / "Qwen2.5-1.5B-Instruct-Q4_K_M.gguf"

    N_CHARS_CHUNK: int = 2000
    N_CHARS_OVERLAP: int = 300
    TOP_K: int = 8
    TOP_K_RERANK: int = 3


    LLM_N_CTX: int = 3072 
    LLM_N_GPU_LAYERS: int = 0 


    HF_TOKEN: str | None = (
        os.getenv("HF_TOKEN")
        or os.getenv("HUGGINGFACE_HUB_TOKEN")
        or os.getenv("study")
    )

settings = Settings()