Skip to main content

databricks.labs.dqx.anomaly.scoring_orchestrator

Orchestrates anomaly scoring: route global vs segmented, run global model pipeline.

run_anomaly_scoring​

def run_anomaly_scoring(df_to_score: DataFrame, config: ScoringConfig,
                        registry_table: str, model_name: str) -> DataFrame

Route to segmented or global scoring and return scored DataFrame (caller drops row_id_col).

try_segmented_scoring_fallback​

def try_segmented_scoring_fallback(
        df: DataFrame, config: ScoringConfig,
        registry_client: AnomalyModelRegistry) -> DataFrame | None

Try to score using segmented models as fallback. Returns None if no segments found.