Reflection Processor

​

Classes

​

ReflectionProcessor

Processor for reflection-based self-evaluation and improvement.

​

Parameters

Parameter	Type	Default	Description
config	ReflectionConfig	-	Configuration for reflection process

​

Functions

process_with_reflection

Process a task response through reflection and improvement cycles. Parameters:

• agent (Agent): The main agent that generated the response
• task (Task): The task being processed
• initial_response (Any): The initial response to evaluate and improve

Returns:

• Any: The final improved response

_create_evaluator_agent

Create an evaluator agent for reflection. Parameters:

• main_agent (Agent): The main agent to create evaluator from

Returns:

• Agent: The evaluator agent

_evaluate_response

Evaluate a response using the evaluator agent. Parameters:

• evaluator (Agent): The evaluator agent
• task (Task): The task being evaluated
• response (str): The response to evaluate
• state (ReflectionState): The current reflection state

Returns:

• EvaluationResult: The evaluation result

_generate_improved_response

Generate an improved response based on evaluation feedback. Parameters:

• agent (Agent): The main agent
• task (Task): The task being processed
• previous_response (str): The previous response
• evaluation (EvaluationResult): The evaluation result
• state (ReflectionState): The current reflection state

Returns:

• Any: The improved response

_extract_response_text

Extract text representation from response. Parameters:

• response (Any): The response to extract text from

Returns:

• str: The extracted text

_convert_to_response_format

Convert improved response back to original format. Parameters:

• response_text (str): The improved response text
• original_response (Any): The original response
• task (Task): The task being processed

Returns:

• Any: The converted response

_build_evaluation_context

Build context for evaluation. Parameters:

• task (Task): The task being evaluated
• state (ReflectionState): The current reflection state

Returns:

• str: The evaluation context

_build_improvement_context

Build context for improvement. Parameters:

• task (Task): The task being processed
• state (ReflectionState): The current reflection state

Returns:

• str: The improvement context

_create_fallback_evaluation

Create a fallback evaluation when the evaluator fails. Parameters:

• response (str): The response to evaluate
• error (str): The error message

Returns:

• EvaluationResult: The fallback evaluation result