generate_dq_rule_candidates_with_fabric_ai

Public callable

Append AI-suggested DQ rule candidates to a profiling DataFrame.

Use this after profiling and before rule approval to accelerate candidate generation while keeping deterministic enforcement separate.

Parameters:

Name	Type	Description	Default
profile_df	Any	Fabric PySpark DataFrame with an .ai.generate_response extension. Expected row columns include column_name, data_type, null_count, distinct_count, and row_count for prompt placeholders.	required
business_context	str	Static context embedded in prompt text.	''
dataset_name	str | None	Static dataset name embedded in prompt text.	None
output_col	str	Output column for generated JSON text.	"ai_dq_rule_candidate"
error_col	str	Error column written by AI Functions.	"ai_dq_rule_error"
response_format	str	Fabric AI response format.	"json_object"
concurrency	int	Fabric AI row-level concurrency setting.	20

Returns:

Type	Description
Any	Enriched DataFrame containing AI suggestion output columns.

Notes

This helper executes AI generation only and does not enforce deterministic quality rules.

Source code in src/fabricops_kit/ai.py

def generate_dq_rule_candidates_with_fabric_ai(profile_df, business_context="", dataset_name=None, output_col="ai_dq_rule_candidate", error_col="ai_dq_rule_error", response_format="json_object", concurrency=20, config: FrameworkConfig | None = None):
    """Append AI-suggested DQ rule candidates to a profiling DataFrame.

    Use this after profiling and before rule approval to accelerate candidate
    generation while keeping deterministic enforcement separate.


    Parameters
    ----------
    profile_df : Any
        Fabric PySpark DataFrame with an ``.ai.generate_response`` extension.
        Expected row columns include ``column_name``, ``data_type``, ``null_count``,
        ``distinct_count``, and ``row_count`` for prompt placeholders.
    business_context : str, optional
        Static context embedded in prompt text.
    dataset_name : str | None, optional
        Static dataset name embedded in prompt text.
    output_col : str, default="ai_dq_rule_candidate"
        Output column for generated JSON text.
    error_col : str, default="ai_dq_rule_error"
        Error column written by AI Functions.
    response_format : str, default="json_object"
        Fabric AI response format.
    concurrency : int, default=20
        Fabric AI row-level concurrency setting.

    Returns
    -------
    Any
        Enriched DataFrame containing AI suggestion output columns.

    Notes
    -----
    This helper executes AI generation only and does not enforce deterministic
    quality rules.
    """
    _require_fabric_ai_dataframe(profile_df, "generate_dq_rule_candidates_with_fabric_ai")
    prompt = build_dq_rule_candidate_prompt(business_context=business_context, dataset_name=dataset_name, config=config)
    return profile_df.ai.generate_response(prompt=prompt, is_prompt_template=True, output_col=output_col, error_col=error_col, response_format=response_format, concurrency=concurrency)