config.py

# config.py
# Backward compatible config - imports from src.config for consistency
# This maintains compatibility with existing imports like "from config import settings"

# Import from src.config to ensure consistency
try:
    from src.config import settings, Settings, CacheDirectoryManager
except ImportError:
    # Fallback if src.config not available
    import os
    from pydantic_settings import BaseSettings
    
    class Settings(BaseSettings):
        hf_token: str = os.getenv("HF_TOKEN", "")
        hf_cache_dir: str = os.getenv("HF_HOME", "/tmp/huggingface")
        default_model: str = "mistralai/Mistral-7B-Instruct-v0.2"
        embedding_model: str = "sentence-transformers/all-MiniLM-L6-v2"
        classification_model: str = "cardiffnlp/twitter-roberta-base-emotion"
        max_workers: int = int(os.getenv("MAX_WORKERS", "4"))
        cache_ttl: int = int(os.getenv("CACHE_TTL", "3600"))
        _default_db_path = "/tmp/sessions.db" if os.path.exists("/.dockerenv") or os.path.exists("/tmp") else "sessions.db"
        db_path: str = os.getenv("DB_PATH", _default_db_path)
        _default_faiss_path = "/tmp/embeddings.faiss" if os.path.exists("/.dockerenv") or os.path.exists("/tmp") else "embeddings.faiss"
        faiss_index_path: str = os.getenv("FAISS_INDEX_PATH", _default_faiss_path)
        session_timeout: int = int(os.getenv("SESSION_TIMEOUT", "3600"))
        max_session_size_mb: int = int(os.getenv("MAX_SESSION_SIZE_MB", "10"))
        mobile_max_tokens: int = int(os.getenv("MOBILE_MAX_TOKENS", "800"))
        mobile_timeout: int = int(os.getenv("MOBILE_TIMEOUT", "15000"))
        gradio_port: int = int(os.getenv("GRADIO_PORT", "7860"))
        gradio_host: str = os.getenv("GRADIO_HOST", "0.0.0.0")
        log_level: str = os.getenv("LOG_LEVEL", "INFO")
        log_format: str = os.getenv("LOG_FORMAT", "json")
        
        class Config:
            env_file = ".env"
    
    settings = Settings()

# Context configuration
CONTEXT_CONFIG = {
    'max_context_tokens': int(os.getenv("MAX_CONTEXT_TOKENS", "4000")),
    'cache_ttl_seconds': int(os.getenv("CACHE_TTL_SECONDS", "300")),
    'max_cache_size': int(os.getenv("MAX_CACHE_SIZE", "100")),
    'parallel_processing': os.getenv("PARALLEL_PROCESSING", "True").lower() == "true",
    'context_decay_factor': float(os.getenv("CONTEXT_DECAY_FACTOR", "0.8")),
    'max_interactions_to_keep': int(os.getenv("MAX_INTERACTIONS_TO_KEEP", "10")),
    'enable_metrics': os.getenv("ENABLE_METRICS", "True").lower() == "true",
    'compression_enabled': os.getenv("COMPRESSION_ENABLED", "True").lower() == "true",
    'summarization_threshold': int(os.getenv("SUMMARIZATION_THRESHOLD", "2000"))  # tokens
}

# Model selection for context operations
CONTEXT_MODELS = {
    'summarization': os.getenv("CONTEXT_SUMMARIZATION_MODEL", "Qwen/Qwen2.5-7B-Instruct"),
    'intent': os.getenv("CONTEXT_INTENT_MODEL", "Qwen/Qwen2.5-7B-Instruct"),
    'synthesis': os.getenv("CONTEXT_SYNTHESIS_MODEL", "Qwen/Qwen2.5-72B-Instruct")
}