📘 Overview

AgenticRAGDepthQAGenerator is an operator used to generate deeper questions based on existing QA pairs. It iteratively refines and expands upon an initial question to create more complex and in-depth queries over multiple rounds.

__init__ function

def __init__(self,
             llm_serving: LLMServingABC = None,
             n_rounds:int = 2
             ):

init Parameters

Parameter Name	Type	Default Value	Description
llm_serving	LLMServingABC	None	An LLM serving instance for executing inference and generation.
n_rounds	int	2	The number of rounds to generate deeper questions.

Prompt Template Descriptions

Prompt Template Name	Primary Use	Applicable Scenarios	Feature Description
This operator does not use prompt templates; it generates and processes deeper question-answer pairs directly through multi-round reasoning steps.

run function

def run(
        self,
        storage: DataFlowStorage,
        input_key:str = "question",
        output_key:str = "depth_question",
        ):

Executes the main operator logic: reads the input DataFrame from storage, generates deeper questions over multiple rounds of reasoning, and writes the results—including newly generated columns—back to storage.

Parameters

Name	Type	Default Value	Description
storage	DataFlowStorage	Required	The DataFlow storage instance, responsible for reading and writing data.
input_key	str	"question"	The input column name, corresponding to the question field.
output_key	str	"depth_question"	The base name for the output columns that will store the generated deeper questions.

🧠 Example Usage

from dataflow.operators.agentic_rag.generate.agenticrag_depth_qa_generator import AgenticRAGDepthQAGenerator
from dataflow.utils.storage import DataFlowStorage

# Initialize the operator
generator = AgenticRAGDepthQAGenerator(
    llm_serving=your_llm_serving_instance,
    n_rounds=2
)

# Run the operator
storage = DataFlowStorage()
generator.run(
    storage=storage,
    input_key="question",
    output_key="depth_question"
)

🧾 Default Output Format (Output Format)

Field	Type	Description
identifier	str	Content identifier extracted from the original question
new_identifier_	str	New identifiers generated in round i
relation_	str	Relations between identifiers in round i
depth_question_	str	Deeper questions generated in round i

Example Input:

{
  "question": "What is machine learning?"
}

Example Output:

{
  "question": "What is machine learning?",
  "identifier": "machine learning concepts",
  "new_identifier_1": "artificial intelligence algorithms",
  "relation_1": "is a subset of",
  "depth_question_1": "How do artificial intelligence algorithms work in machine learning systems?",
  "new_identifier_2": "statistical learning theory",
  "relation_2": "provides foundation for",
  "depth_question_2": "What role does statistical learning theory play in the development of machine learning algorithms?"
}