"""Configuration management for the refinement engine.
This module provides centralized configuration management for model names,
domain-specific system prompts, and other refinement engine settings.
"""
import os
from domains import get_domain_system_prompt
class ConfigurationManager:
"""Manages configuration settings for the refinement engine."""
@staticmethod
def get_model_name() -> str:
"""Get the current model name for performance hints.
Returns:
The configured Bedrock model ID
"""
return os.getenv("BEDROCK_MODEL_ID", "anthropic.claude-3-sonnet-20240229-v1:0")
@staticmethod
def get_domain_system_prompt(domain: str) -> str:
"""Get domain-specific system prompt.
Args:
domain: The domain type (technical, marketing, legal, etc.)
Returns:
Domain-specific system prompt
"""
return get_domain_system_prompt(domain)
@staticmethod
def get_default_max_iterations() -> int:
"""Get default maximum iterations for refinement.
Returns:
Default maximum number of refinement iterations
"""
return int(os.getenv("MAX_ITERATIONS", "10"))
@staticmethod
def get_default_convergence_threshold() -> float:
"""Get default convergence threshold.
Returns:
Default convergence threshold for refinement completion
"""
return float(os.getenv("CONVERGENCE_THRESHOLD", "0.98"))
@staticmethod
def get_parallel_critiques_count() -> int:
"""Get number of parallel critiques to generate.
Returns:
Number of parallel critique perspectives
"""
return int(os.getenv("PARALLEL_CRITIQUES", "3"))
@staticmethod
def is_cot_enhancement_enabled() -> bool:
"""Check if Chain of Thought enhancement is enabled.
Returns:
True if CoT enhancement should be used
"""
return os.getenv("COT_ENHANCEMENT", "true").lower() == "true"
