src/core/config.py

import os
from pydantic_settings import BaseSettings, SettingsConfigDict

class Settings(BaseSettings):
    model_config = SettingsConfigDict(env_file=".env", env_file_encoding='utf-8', extra='ignore')
    
    PROJECT_NAME: str = "RAG API Service"
    API_V1_STR: str = "/api/v1"
    
    QDRANT_HOST: str = os.getenv("QDRANT_HOST", "localhost")
    QDRANT_PORT: int = int(os.getenv("QDRANT_PORT", "6333"))
    QDRANT_URL: str = os.getenv("QDRANT_URL", "")          # Cloud URL (overrides host/port)
    QDRANT_API_KEY: str = os.getenv("QDRANT_API_KEY", "")  # Cloud API Key
    QDRANT_COLLECTION: str = os.getenv("QDRANT_COLLECTION", "news_articles")
    
    CLICKHOUSE_HOST: str = os.getenv("CLICKHOUSE_HOST", "localhost")
    CLICKHOUSE_PORT: int = int(os.getenv("CLICKHOUSE_PORT", "8123"))
    CLICKHOUSE_USER: str = os.getenv("CLICKHOUSE_USER", "default")
    CLICKHOUSE_PASSWORD: str = os.getenv("CLICKHOUSE_PASSWORD", "")
    CLICKHOUSE_DB: str = os.getenv("CLICKHOUSE_DB", "default")
    CLICKHOUSE_SECURE: bool = os.getenv("CLICKHOUSE_SECURE", "false").lower() == "true"

    
    # Embedding Model Config
    EMBEDDING_MODEL: str = os.getenv("EMBEDDING_MODEL", "BAAI/bge-m3")
    VECTOR_SIZE: int = int(os.getenv("VECTOR_SIZE", "1024"))
    RERANKER_MODEL: str = os.getenv("RERANKER_MODEL", "BAAI/bge-reranker-v2-m3")
    
    # PostgreSQL / Neon Config
    DATABASE_URL: str = os.getenv("DATABASE_URL", "")  # Full Neon URL (overrides individual fields)
    POSTGRES_USER: str = os.getenv("POSTGRES_USER", "postgres")
    POSTGRES_PASSWORD: str = os.getenv("POSTGRES_PASSWORD", "postgres")
    POSTGRES_SERVER: str = os.getenv("POSTGRES_SERVER", "localhost")
    POSTGRES_PORT: str = os.getenv("POSTGRES_PORT", "5432")
    POSTGRES_DB: str = os.getenv("POSTGRES_DB", "rag_interactions")

    @property
    def SQLALCHEMY_DATABASE_URI(self) -> str:
        if self.DATABASE_URL:
            return self.DATABASE_URL
        return f"postgresql://{self.POSTGRES_USER}:{self.POSTGRES_PASSWORD}@{self.POSTGRES_SERVER}:{self.POSTGRES_PORT}/{self.POSTGRES_DB}"

    # LLM Settings
    # Supported providers: "groq", "gemini", "together", "openai", "ollama"
    LLM_PROVIDER: str = os.getenv("LLM_PROVIDER", "groq")

    # Groq — free, 200+ tok/s, llama-3.3-70b-versatile | https://console.groq.com
    OPENAI_API_KEY: str = os.getenv("OPENAI_API_KEY", "")
    GROQ_API_KEY: str = os.getenv("GROQ_API_KEY", "")
    GROQ_MODEL: str = os.getenv("GROQ_MODEL", "llama-3.3-70b-versatile")

    # Google Gemini — free tier (15 RPM / 1M TPM) | https://aistudio.google.com/apikey
    GEMINI_API_KEY: str = os.getenv("GEMINI_API_KEY", "")
    GEMINI_MODEL: str = os.getenv("GEMINI_MODEL", "gemini-1.5-flash")

    # Together AI — free $25 credit | https://api.together.ai
    TOGETHER_API_KEY: str = os.getenv("TOGETHER_API_KEY", "")
    TOGETHER_MODEL: str = os.getenv("TOGETHER_MODEL", "meta-llama/Llama-3.3-70B-Instruct-Turbo")

    # HuggingFace Inference API — free with HF token | https://huggingface.co/settings/tokens
    HF_TOKEN: str = os.getenv("HF_TOKEN", "")
    HF_MODEL: str = os.getenv("HF_MODEL", "meta-llama/Llama-3.1-8B-Instruct")

    # OpenRouter — free model pool | https://openrouter.ai/keys
    OPENROUTER_API_KEY: str = os.getenv("OPENROUTER_API_KEY", "")

    # Ollama — local inference
    OLLAMA_HOST: str = os.getenv("OLLAMA_HOST", "http://localhost:11434")
    OLLAMA_MODEL: str = os.getenv("OLLAMA_MODEL", "llama3.2")
    
    # Redis Settings
    REDIS_URL: str = os.getenv("REDIS_URL", "")          # Full URL (Upstash) - overrides host/port
    REDIS_HOST: str = os.getenv("REDIS_HOST", "localhost")
    REDIS_PORT: int = int(os.getenv("REDIS_PORT", "6380"))
    REDIS_DB: int = int(os.getenv("REDIS_DB", "0"))
    REDIS_PASSWORD: str = os.getenv("REDIS_PASSWORD", "")
    
    # Hybrid Search Settings
    ENABLE_HYBRID_SEARCH: bool = os.getenv("ENABLE_HYBRID_SEARCH", "true").lower() == "true"
    LIVE_SEARCH_TIMEOUT: float = float(os.getenv("LIVE_SEARCH_TIMEOUT", "1.5"))
    LIVE_SEARCH_MAX_RESULTS: int = int(os.getenv("LIVE_SEARCH_MAX_RESULTS", "15"))
    LIVE_SEARCH_WEIGHT: float = float(os.getenv("LIVE_SEARCH_WEIGHT", "0.5"))
    DB_SEARCH_WEIGHT: float = float(os.getenv("DB_SEARCH_WEIGHT", "0.5"))
    
    # Jina Reader Settings (Full Article Extraction)
    ENABLE_JINA_READER: bool = os.getenv("ENABLE_JINA_READER", "true").lower() == "true"
    JINA_READER_TIMEOUT: float = float(os.getenv("JINA_READER_TIMEOUT", "8.0"))
    JINA_READER_MAX_CONCURRENT: int = int(os.getenv("JINA_READER_MAX_CONCURRENT", "10"))
    JINA_API_KEY: str = os.getenv("JINA_API_KEY", "")  # Get free key at https://jina.ai

    # Jina Reranker API
    JINA_RERANKER_ENABLED: bool = os.getenv("JINA_RERANKER_ENABLED", "true").lower() == "true"
    JINA_RERANKER_MODEL: str = os.getenv("JINA_RERANKER_MODEL", "jina-reranker-v3")
    JINA_RERANKER_TIMEOUT: float = float(os.getenv("JINA_RERANKER_TIMEOUT", "5.0"))
    
    # NewsAPI Settings (Real-Time News Search)
    NEWSAPI_KEY: str = os.getenv("NEWSAPI_KEY", "")  # Get free key at https://newsapi.org/register
    NEWSAPI_ENABLED: bool = os.getenv("NEWSAPI_ENABLED", "true").lower() == "true"
    NEWSAPI_TIMEOUT: float = float(os.getenv("NEWSAPI_TIMEOUT", "2.0"))
    NEWSAPI_MAX_RESULTS: int = int(os.getenv("NEWSAPI_MAX_RESULTS", "20"))
    
    # Cache Settings (TTL in seconds)
    CACHE_RESPONSE_TTL: int = int(os.getenv("CACHE_RESPONSE_TTL", "300"))      # 5 minutes
    CACHE_LIVE_TTL: int = int(os.getenv("CACHE_LIVE_TTL", "600"))              # 10 minutes
    CACHE_TRANSLATION_TTL: int = int(os.getenv("CACHE_TRANSLATION_TTL", "3600"))  # 1 hour
    CACHE_INTENT_TTL: int = int(os.getenv("CACHE_INTENT_TTL", "3600"))         # 1 hour
    
    # Security Settings
    SECRET_KEY: str = os.getenv("SECRET_KEY", "a_very_secret_key_change_me_in_production")
    ACCESS_TOKEN_EXPIRE_MINUTES: int = int(os.getenv("ACCESS_TOKEN_EXPIRE_MINUTES", "60"))

    # Kafka Settings (for Top Stories — read-only consumer)
    KAFKA_BOOTSTRAP_SERVERS: str = os.getenv("KAFKA_BOOTSTRAP_SERVERS", "")
    KAFKA_SSL_CA: str = os.getenv("KAFKA_SSL_CA", "")
    KAFKA_SSL_CERT: str = os.getenv("KAFKA_SSL_CERT", "")
    KAFKA_SSL_KEY: str = os.getenv("KAFKA_SSL_KEY", "")
    KAFKA_TOPIC_PROCESSED: str = os.getenv("KAFKA_TOPIC_PROCESSED", "news.processed")
    
settings = Settings()