Thuong Nguyen
Initial deployment: FastAPI backend with data and Dockerfile
f3cb94f
"""
Configuration management for Plant Medicine RAG Backend
"""
from pydantic_settings import BaseSettings
from functools import lru_cache
class Settings(BaseSettings):
"""Application settings loaded from environment variables"""
# Supabase
supabase_url: str
supabase_anon_key: str
supabase_db_uri: str = "" # Optional: for direct DB connection in scripts
# MegLLM
megllm_api_key: str = ""
# Embedding Model
embedding_model_name: str = "AITeamVN/Vietnamese_Embedding"
embedding_dimension: int = 1024 # Actual dimension from model
# CV API
cv_api_url: str = "https://thuonguyenvan-plantsclassify.hf.space"
cv_api_timeout: int = 60
# Data paths
data_dir: str = "data"
photos_dir: str = "inat_representative_photos"
# RAG settings
chunk_max_tokens: int = 250
retrieval_top_k: int = 20
class Config:
env_file = ".env"
case_sensitive = False
@lru_cache()
def get_settings() -> Settings:
"""Get cached settings instance"""
return Settings()