snowflake.ml.monitoring.entities.model_monitor_config.ModelMonitorSourceConfig¶
- class snowflake.ml.monitoring.entities.model_monitor_config.ModelMonitorSourceConfig(source: str | None = None, timestamp_column: str | None = None, id_columns: list[str] | None = None, prediction_score_columns: list[str] | None = None, prediction_class_columns: list[str] | None = None, actual_score_columns: list[str] | None = None, actual_class_columns: list[str] | None = None, baseline: str | None = None, segment_columns: list[str] | None = None, custom_metric_columns: list[str] | None = None, timestamp_custom_metric_table: str | None = None, ground_truth: str | None = None)¶
Bases:
objectConfiguration for the source of data to be monitored.
Pass
sourceandtimestamp_columnto monitor a user table. Omit both to monitor the model inference table.ground_truthis only valid when monitoring the model inference table, and must be passed withid_columns.Attributes
- actual_class_columns: list[str] | None = None¶
List of columns in the source containing actual classes for classification models. Inferred when omitted for the model inference table.
- actual_score_columns: list[str] | None = None¶
List of columns in the source containing actual scores. Inferred when omitted for the model inference table.
- baseline: str | None = None¶
Name of table containing the baseline data.
- custom_metric_columns: list[str] | None = None¶
List of columns in the source containing custom metrics. Not accepted for the model inference table.
- ground_truth: str | None = None¶
Name of the ground truth table. Model inference table monitors only. Pass this with
id_columns.
- id_columns: list[str] | None = None¶
List of columns containing unique identifiers. Required with
source. Pass withground_truth.
- prediction_class_columns: list[str] | None = None¶
List of columns in the source containing prediction classes for classification models. Inferred when omitted for the model inference table.
- prediction_score_columns: list[str] | None = None¶
List of columns in the source containing prediction scores. Can be regression scores for regression models and probability scores for classification models. Inferred when omitted for the model inference table.
- segment_columns: list[str] | None = None¶
List of columns in the source containing segment information for grouped monitoring.
- source: str | None = None¶
Name of table or view containing monitoring data. Omit this to monitor the model inference table.
- timestamp_column: str | None = None¶
Name of column in the source containing timestamp. Required with
source. Omit this for the model inference table.
- timestamp_custom_metric_table: str | None = None¶
Optional name of a table containing timestamp-based custom metrics. Can be specified unqualified or fully qualified as database.schema.table. Not accepted for the model inference table.