config.py

import os
import logging
import warnings
from dataclasses import dataclass, field

from dotenv import load_dotenv

load_dotenv()

# Suppress noisy third-party logs
os.environ["TOKENIZERS_PARALLELISM"] = "false"
logging.getLogger("sentence_transformers").setLevel(logging.WARNING)
logging.getLogger("transformers").setLevel(logging.WARNING)
logging.getLogger("huggingface_hub").setLevel(logging.WARNING)
warnings.filterwarnings("ignore", message=".*Pydantic V1.*")
warnings.filterwarnings("ignore", message=".*urllib3.*")
warnings.filterwarnings("ignore", message=".*HuggingFaceEmbeddings.*")
warnings.filterwarnings("ignore", category=DeprecationWarning)


@dataclass
class Settings:
    llm_base_url: str = field(
        default_factory=lambda: os.getenv("LLM_BASE_URL", "")
    )
    llm_model: str = field(
        default_factory=lambda: os.getenv("LLM_MODEL", "")
    )
    llm_api_key: str = field(
        default_factory=lambda: os.getenv("LLM_API_KEY", "")
    )

    @property
    def is_llm_configured(self) -> bool:
        return bool(self.llm_base_url and self.llm_model)
    embedding_model: str = field(
        default_factory=lambda: os.getenv(
            "EMBEDDING_MODEL", "sentence-transformers/all-MiniLM-L6-v2"
        )
    )
    faiss_index_path: str = field(
        default_factory=lambda: os.getenv("FAISS_INDEX_PATH", "rag/faiss_index")
    )
    memory_dir: str = field(
        default_factory=lambda: os.getenv("MEMORY_DIR", "memory/data")
    )
    ocr_confidence_threshold: float = 0.6
    asr_confidence_threshold: float = 0.6
    verifier_confidence_threshold: float = 0.7
    rag_top_k: int = 5
    max_solver_retries: int = 2


settings = Settings()