mirror of
https://github.com/Sosokker/plain-rag.git
synced 2025-12-18 14:34:05 +01:00
64 lines
1.6 KiB
Python
64 lines
1.6 KiB
Python
import os
|
|
from functools import lru_cache
|
|
from pathlib import Path
|
|
|
|
from pydantic import Field, ValidationError
|
|
from pydantic_settings import BaseSettings, SettingsConfigDict
|
|
from structlog import get_logger
|
|
|
|
ROOT = Path(__file__).resolve().parent.parent.parent
|
|
ENV_FILE = ROOT / ".env"
|
|
|
|
logger = get_logger()
|
|
|
|
|
|
class Settings(BaseSettings):
|
|
# Database
|
|
DB_SERVER: str = "localhost"
|
|
DB_USER: str = "postgres"
|
|
DB_PASSWORD: str = ""
|
|
DB_NAME: str = "chat_hub"
|
|
DB_PORT: str = "5432"
|
|
|
|
# Vector Store
|
|
VECTOR_STORE_TYPE: str = "pgvector" # or "chroma", "faiss", etc.
|
|
EMBEDDING_MODEL: str = "sentence-transformers/all-MiniLM-L6-v2"
|
|
RERANKER_MODEL: str = "cross-encoder/ms-marco-MiniLM-L-6-v2"
|
|
|
|
# File uploads
|
|
ALLOWED_DOCUMENT_TYPES: list[str] = Field(default=["pdf", "txt", "md"])
|
|
|
|
# LLM Models
|
|
GEMINI_API_KEY: str = "secret"
|
|
|
|
# Prompt
|
|
# may be use prompt.pys
|
|
SYSTEM_PROMPT: str = "You are a helpful assistant that answers questions based on the provided context."
|
|
|
|
model_config = SettingsConfigDict(
|
|
env_file=ENV_FILE,
|
|
env_file_encoding="utf-8",
|
|
case_sensitive=True,
|
|
extra="ignore",
|
|
)
|
|
|
|
|
|
@lru_cache
|
|
def get_settings() -> Settings:
|
|
"""
|
|
Get cached settings instance.
|
|
This function uses lru_cache to prevent re-reading the environment on each call.
|
|
"""
|
|
try:
|
|
return Settings()
|
|
except ValidationError:
|
|
logger.exception("Error loading settings")
|
|
raise
|
|
|
|
|
|
settings = get_settings()
|
|
|
|
# Set environment variables for third-party libraries
|
|
os.environ["GEMINI_API_KEY"] = settings.GEMINI_API_KEY
|
|
os.environ["TOKENIZERS_PARALLELISM"] = "false"
|