citrine.informatics.reports module

Tools for working with reports.

class citrine.informatics.reports.FeatureImportanceReport

Bases: Serializable[FeatureImportanceReport]

Feature importances for a specific model response.

FeatureImportanceReport objects are constructed from saved models and should not be user-instantiated.

classmethod build(data: dict) → Self: Build an instance of this object from given data.

dump() → dict: Dump this instance.

importances = None

map from feature name to its importance

Type:: dict[str, float]

output_key = None

output descriptor key for which these feature importances are applicable

Type:: str

class citrine.informatics.reports.ModelEvaluationResult

Bases: Serializable[ModelEvaluationResult]

Settings and evaluation metrics for a single algorithm from AutoML model selection.

ModelEvaluationResult objects are included in a ModelSelectionReport and should not be user-instantiated.

classmethod build(data: dict) → Self: Build an instance of this object from given data.

dump() → dict: Dump this instance.

model_settings = None

property responses: Set[str]: Responses the model was evaluated on.

class citrine.informatics.reports.ModelSelectionReport

Bases: Serializable[ModelSelectionReport]

Summary of selection settings and model performance from AutoML model selection.

ModelSelectionReport objects are constructed from saved models and should not be user-instantiated.

classmethod build(data: dict) → Self: Build an instance of this object from given data.

dump() → dict: Dump this instance.

evaluation_results = None

n_folds = None

class citrine.informatics.reports.ModelSummary

Bases: Serializable[ModelSummary]

Summary of information about a single model in a predictor.

ModelSummary objects are constructed from saved models and should not be user-instantiated.

classmethod build(data: dict) → Self: Build an instance of this object from given data.

dump() → dict: Dump this instance.

feature_importances = None

feature importance reports for each output

Type:: List[FeatureImportanceReport]

inputs = None

list of input descriptors

Type:: List[Descriptor]

model_settings = None

model settings, as a dictionary (keys depend on the model type)

Type:: dict

name = None

the name of the model

Type:: str

outputs = None

list of output descriptors

Type:: List[Descriptor]

predictor_name = ''

the name of the predictor that created this model

Type:: str

predictor_uid = None

the unique Citrine id of the predictor that created this model

Type:: Optional[UUID]

selection_summary = None

optional results of AutoML model selection

Type:: Optional[ModelSelectionReport]

training_data_count = None

Number of rows in the training data for the model, if applicable.

Type:: int

type_ = None

the type of the model (e.g., “ML Model”, “Featurizer”, etc.)

Type:: str

class citrine.informatics.reports.PredictorReport

Bases: Serializable[PredictorReport], Report

The performance metrics corresponding to a predictor.

PredictorReport objects are constructed from saved models and should not be user-instantiated.

classmethod build(data: dict) → Self: Build an instance of this object from given data.

dump() → dict: Dump this instance.

failed() → bool: Whether the backend process has completed unsuccessfully.

classmethod get_type(data) → Type[Serializable]: Return the only subtype.

in_progress() → bool: Whether the backend process is in progress.

post_build(): Modify a PredictorReport object in-place after deserialization.

succeeded() → bool: Whether the backend process has completed successfully.

descriptors = []

All descriptors that appear in the predictor

Type:: List[Descriptor]

model_summaries = []

Summaries of all models in the predictor

Type:: List[ModelSummary]

status = None

The status of the report. Possible statuses are PENDING, ERROR, and OK.

Type:: str

uid = None

Unique Citrine id of the predictor report

Type:: UUID

class citrine.informatics.reports.Report

Bases: PolymorphicSerializable[Report], AsynchronousObject

A Citrine Report contains information related to a module.

Abstract type that returns the proper type given a serialized dict.

classmethod build(data: dict) → SelfType: Build the underlying type.

failed() → bool: Whether the backend process has completed unsuccessfully.

classmethod get_type(data) → Type[Serializable]: Return the only subtype.

in_progress() → bool: Whether the backend process is in progress.

abstract post_build(): Executes after a .build() is called in [[Report]].

succeeded() → bool: Whether the backend process has completed successfully.