📘 Overview

The Text2SQLPromptGenerator is an operator designed to create prompts for Text-to-SQL tasks. It extracts schema information from a specified database and combines it with a natural language question to generate a detailed prompt suitable for a large language model. This operator facilitates the automated generation of context-rich prompts, which is a crucial step in building accurate Text-to-SQL systems.

__init__ function

def __init__(self, 
            database_manager: DatabaseManager,
            prompt_template = None
        ):

Parameter	Type	Default	Description
database_manager	DatabaseManager	Required	An instance of DatabaseManager to handle database interactions and schema retrieval.
prompt_template	PromptABC	Text2SQLPromptGeneratorPrompt()	The prompt template object used to structure the final prompt. Defaults to a standard Text2SQL prompt.

Prompt Template Descriptions

Prompt Template Name	Primary Use	Applicable Scenarios	Feature Description

run function

def run(self, storage: DataFlowStorage, 
        input_question_key: str = "question",
        input_db_id_key: str = "db_id",
        input_evidence_key: str = "evidence",
        output_prompt_key: str = "prompt"
    ):

Parameter	Type	Default	Description
storage	DataFlowStorage	Required	The DataFlow storage instance for reading input data and writing results.
input_question_key	str	"question"	The column name in the input data that contains the natural language question.
input_db_id_key	str	"db_id"	The column name in the input data that contains the database identifier.
input_evidence_key	str	"evidence"	The column name for optional evidence or hints to be included in the prompt.
output_prompt_key	str	"prompt"	The column name where the generated prompt will be stored in the output data.

🧠 Example Usage

🧾 Default Output Format

The output includes all original columns from the input data, with an additional column containing the generated prompt.

Field	Type	Description
question	str	The input natural language question.
db_id	str	The identifier for the database.
evidence	str	Optional evidence or hints provided in the input.
prompt	str	The generated prompt for the Text-to-SQL task.

Example Input:

{
    "question": "What are the names of the departments with more than 10 employees?",
    "db_id": "company_database",
    "evidence": ""
}

Example Output:

{
    "question": "What are the names of the departments with more than 10 employees?",
    "db_id": "company_database",
    "evidence": "",
    "prompt": "Given the database schema for 'company_database', please generate the SQL query for the following question: What are the names of the departments with more than 10 employees?"
}