Skip to content
🚀  Book a demo💡  Explore the Comet website
Doc version: 3.13.98

comet_mpm.api_objects.Model 

Model(client: Client, model_id: str)

A model instance for interacting with Comet MPM model-specific operations.

This class provides high-level methods for querying model predictions, metrics, and feature analysis. It can be configured with panel options for default parameter values.

Parameters:

  • client (Client) –

    The Comet MPM client instance

  • model_id (str) –

    The ID of the model to work with

Initialize a Model instance.

Parameters:

  • client (Client) –

    The Comet MPM client instance for making API calls

  • model_id (str) –

    The ID of the model to work with

  • _config –

    Configuration strategy instance

Attributes

model_id instance-attribute 

model_id = model_id

Functions

get_categorical_features 

get_categorical_features() -> List[str]

Get the list of categorical features available for this model.

Returns:

  • List[str] –

    List[str]: List of categorical feature names

get_custom_metric 

get_custom_metric(
    sql: str,
    start_date: Optional[str] = None,
    end_date: Optional[str] = None,
    interval_type: Optional[str] = None,
    filters: Optional[List[str]] = None,
    model_version: Optional[str] = None,
) -> pd.DataFrame

Execute a custom SQL query to retrieve model metrics.

Parameters:

  • sql (str) –

    SQL query string to execute

  • start_date (Optional[str], default: None ) –

    Start date for filtering results (ISO format)

  • end_date (Optional[str], default: None ) –

    End date for filtering results (ISO format)

  • interval_type (Optional[str], default: None ) –

    Type of interval for aggregation ("DAILY" or "HOURLY")

  • filters (Optional[List[str]], default: None ) –

    List of filters to apply to results

  • model_version (Optional[str], default: None ) –

    Specific model version to query

Returns:

  • DataFrame ( DataFrame ) –

    Results of the SQL query

get_details 

get_details() -> Dict[str, Any]

Get the details of a model.

Returns:

  • Dict[str, Any] –

    Dict[str, Any]: Model details including metadata, configuration, and status

get_feature_category_distribution 

get_feature_category_distribution(
    feature_name: str,
    normalize: bool = False,
    start_date: Optional[str] = None,
    end_date: Optional[str] = None,
    interval_type: Optional[str] = None,
    filters: Optional[List[str]] = None,
    model_version: Optional[str] = None,
) -> pd.DataFrame

Get the distribution of categories for a categorical feature.

Parameters:

  • feature_name (str) –

    Name of the categorical feature

  • normalize (bool, default: False ) –

    If True, returns percentages instead of counts

  • start_date (Optional[str], default: None ) –

    Start date for distribution calculation (ISO format)

  • end_date (Optional[str], default: None ) –

    End date for distribution calculation (ISO format)

  • interval_type (Optional[str], default: None ) –

    Type of interval for aggregation ("DAILY" or "HOURLY")

  • filters (Optional[List[str]], default: None ) –

    List of filters to apply to distribution calculation

  • model_version (Optional[str], default: None ) –

    Specific model version to query

Returns:

  • DataFrame ( DataFrame ) –

    Distribution of feature categories

get_feature_density 

get_feature_density(
    feature_name: str,
    start_date: Optional[str] = None,
    end_date: Optional[str] = None,
    filters: Optional[List[str]] = None,
    model_version: Optional[str] = None,
    interval_type: Optional[str] = None,
) -> pd.DataFrame

Get the probability density function (PDF) of a numeric feature.

Parameters:

  • feature_name (str) –

    Name of the numeric feature

  • start_date (Optional[str], default: None ) –

    Start date for density calculation (ISO format)

  • end_date (Optional[str], default: None ) –

    End date for density calculation (ISO format)

  • filters (Optional[List[str]], default: None ) –

    List of filters to apply to density calculation

  • model_version (Optional[str], default: None ) –

    Specific model version to query

  • interval_type (Optional[str], default: None ) –

    Type of interval for aggregation ("DAILY" or "HOURLY")

Returns:

  • DataFrame ( DataFrame ) –

    Probability density function of the feature values

get_feature_drift 

get_feature_drift(
    feature_name: str,
    algorithm: str = "EMD",
    start_date: Optional[str] = None,
    end_date: Optional[str] = None,
    interval_type: Optional[str] = None,
    filters: Optional[List[str]] = None,
    model_version: Optional[str] = None,
) -> pd.DataFrame

Calculate drift metrics for a specific feature.

Parameters:

  • feature_name (str) –

    Name of the feature to calculate drift for

  • algorithm (str, default: 'EMD' ) –

    Drift calculation algorithm ("EMD", "PSI", or "KL")

  • start_date (Optional[str], default: None ) –

    Start date for drift calculation (ISO format)

  • end_date (Optional[str], default: None ) –

    End date for drift calculation (ISO format)

  • interval_type (Optional[str], default: None ) –

    Type of interval for aggregation ("DAILY" or "HOURLY")

  • filters (Optional[List[str]], default: None ) –

    List of filters to apply to drift calculation

  • model_version (Optional[str], default: None ) –

    Specific model version to query

Returns:

  • DataFrame ( DataFrame ) –

    Drift metrics for the specified feature

get_feature_percentiles 

get_feature_percentiles(
    feature_name: str,
    percentiles: Optional[List[float]] = None,
    start_date: Optional[str] = None,
    end_date: Optional[str] = None,
    filters: Optional[List[str]] = None,
    model_version: Optional[str] = None,
    interval_type: Optional[str] = None,
) -> pd.DataFrame

Get the specified percentiles for a numeric feature.

Parameters:

  • feature_name (str) –

    Name of the numeric feature

  • percentiles (Optional[List[float]], default: None ) –

    List of percentiles to calculate (default: [0, 0.1, 0.25, 0.5, 0.75, 0.9, 1]) Only these specific percentile values are supported

  • start_date (Optional[str], default: None ) –

    Start date for percentile calculation (ISO format)

  • end_date (Optional[str], default: None ) –

    End date for percentile calculation (ISO format)

  • filters (Optional[List[str]], default: None ) –

    List of filters to apply to percentile calculation

  • model_version (Optional[str], default: None ) –

    Specific model version to query

  • interval_type (Optional[str], default: None ) –

    Type of interval for aggregation. Must be one of "DAILY" or "HOURLY". Passing any other value will raise a ValueError.

Returns:

  • DataFrame ( DataFrame ) –

    Percentile values for the specified feature

get_nb_predictions 

get_nb_predictions(
    start_date: Optional[str] = None,
    end_date: Optional[str] = None,
    interval_type: Optional[str] = None,
    filters: Optional[List[str]] = None,
    model_version: Optional[str] = None,
) -> pd.DataFrame

Get the number of predictions for a model within a specified time range.

Parameters:

  • start_date (Optional[str], default: None ) –

    Start date for filtering predictions (ISO format)

  • end_date (Optional[str], default: None ) –

    End date for filtering predictions (ISO format)

  • interval_type (Optional[str], default: None ) –

    Type of interval for aggregation ("DAILY" or "HOURLY")

  • filters (Optional[List[str]], default: None ) –

    List of filters to apply to predictions

  • model_version (Optional[str], default: None ) –

    Specific model version to query

Returns:

  • DataFrame –

    pd.DataFrame: DataFrame containing the number of predictions matching the criteria

get_numerical_features 

get_numerical_features() -> List[str]

Get the list of numerical features available for this model.

Returns:

  • List[str] –

    List[str]: List of numerical feature names

Nov. 5, 2025