tempor.models.clairvoyance2.interface package

Submodules

Module contents

class tempor.models.clairvoyance2.interface.BaseModel(params: dict[str, Any] | None = None)[source]

Bases: ABC

requirements : Requirements = Requirements(dataset_requirements=DatasetRequirements(requires_static_covariates_present=False, requires_no_missing_data=False, static_covariates_value_type=<DataValueOpts.ANY: 1>, temporal_covariates_value_type=<DataValueOpts.ANY: 1>, temporal_targets_value_type=<DataValueOpts.ANY: 1>, temporal_treatments_value_type=<DataValueOpts.ANY: 1>, event_covariates_value_type=<DataValueOpts.ANY: 1>, event_targets_value_type=<DataValueOpts.ANY: 1>, event_treatments_value_type=<DataValueOpts.ANY: 1>, requires_all_temporal_data_samples_aligned=False, requires_all_temporal_data_regular=False, requires_all_temporal_data_index_numeric=False, requires_all_temporal_containers_shares_index=True), prediction_requirements=None, treatment_effects_requirements=None)
DEFAULT_PARAMS : dict[str, Any] | NamedTuple = {}
check_unknown_params : bool = True
abstract check_model_requirements() → None[source]
check_data_requirements_general(called_at_fit_time: bool, data: Dataset, **kwargs)[source]
fit(data: Dataset, **kwargs) → BaseModel[source]
class tempor.models.clairvoyance2.interface.DatasetRequirements(requires_static_covariates_present: bool = False, requires_no_missing_data: bool = False, static_covariates_value_type: tempor.models.clairvoyance2.interface.requirements.DataValueOpts = <DataValueOpts.ANY: 1>, temporal_covariates_value_type: tempor.models.clairvoyance2.interface.requirements.DataValueOpts = <DataValueOpts.ANY: 1>, temporal_targets_value_type: tempor.models.clairvoyance2.interface.requirements.DataValueOpts = <DataValueOpts.ANY: 1>, temporal_treatments_value_type: tempor.models.clairvoyance2.interface.requirements.DataValueOpts = <DataValueOpts.ANY: 1>, event_covariates_value_type: tempor.models.clairvoyance2.interface.requirements.DataValueOpts = <DataValueOpts.ANY: 1>, event_targets_value_type: tempor.models.clairvoyance2.interface.requirements.DataValueOpts = <DataValueOpts.ANY: 1>, event_treatments_value_type: tempor.models.clairvoyance2.interface.requirements.DataValueOpts = <DataValueOpts.ANY: 1>, requires_all_temporal_data_samples_aligned: bool = False, requires_all_temporal_data_regular: bool = False, requires_all_temporal_data_index_numeric: bool = False, requires_all_temporal_containers_shares_index: bool = True)[source]

Bases: object

requires_static_covariates_present : bool = False
requires_no_missing_data : bool = False
static_covariates_value_type : DataValueOpts = 1
temporal_covariates_value_type : DataValueOpts = 1
temporal_targets_value_type : DataValueOpts = 1
temporal_treatments_value_type : DataValueOpts = 1
event_covariates_value_type : DataValueOpts = 1
event_targets_value_type : DataValueOpts = 1
event_treatments_value_type : DataValueOpts = 1
requires_all_temporal_data_samples_aligned : bool = False
requires_all_temporal_data_regular : bool = False
requires_all_temporal_data_index_numeric : bool = False
requires_all_temporal_containers_shares_index : bool = True
class tempor.models.clairvoyance2.interface.DataStructureOpts(value)[source]

Bases: Enum

An enumeration.

TIME_SERIES = 1
STATIC = 2
EVENT = 3
class tempor.models.clairvoyance2.interface.DataValueOpts(value)[source]

Bases: Enum

An enumeration.

ANY = 1
NUMERIC = 2
NUMERIC_CATEGORICAL = 3
NUMERIC_BINARY = 4
class tempor.models.clairvoyance2.interface.Horizon(horizon_type: HorizonOpts)[source]

Bases: ABC

horizon_type : HorizonOpts
class tempor.models.clairvoyance2.interface.HorizonOpts(value)[source]

Bases: Enum

An enumeration.

N_STEP_AHEAD = 1
TIME_INDEX = 2
class tempor.models.clairvoyance2.interface.NStepAheadHorizon(n_step: int)[source]

Bases: Horizon

n_step : int
class tempor.models.clairvoyance2.interface.PredictionRequirements(target_data_structure: tempor.models.clairvoyance2.interface.requirements.DataStructureOpts = <DataStructureOpts.TIME_SERIES: 1>, horizon_type: tempor.models.clairvoyance2.interface.horizon.HorizonOpts = <HorizonOpts.N_STEP_AHEAD: 1>, min_timesteps_target_when_fit: int = 1, min_timesteps_target_when_predict: int = 1)[source]

Bases: object

target_data_structure : DataStructureOpts = 1
horizon_type : HorizonOpts = 1
min_timesteps_target_when_fit : int = 1
min_timesteps_target_when_predict : int = 1
class tempor.models.clairvoyance2.interface.PredictorModel(params: dict[str, Any] | None = None)[source]

Bases: BaseModel, ABC

check_data_requirements_predict(data: Dataset, horizon: Horizon | None, **kwargs)[source]
predict(data: Dataset, horizon: Horizon | None, **kwargs) → TimeSeriesSamples | StaticSamples[source]
fit(data: Dataset, horizon: Horizon | None = None, **kwargs) → PredictorModel[source]
fit_predict(data: Dataset, horizon: Horizon | None, **kwargs) → TimeSeriesSamples | StaticSamples[source]
check_model_requirements() → None[source]
params : DotMap
inferred_params : DotMap
class tempor.models.clairvoyance2.interface.Requirements(dataset_requirements: tempor.models.clairvoyance2.interface.requirements.DatasetRequirements = DatasetRequirements(requires_static_covariates_present=False, requires_no_missing_data=False, static_covariates_value_type=<DataValueOpts.ANY: 1>, temporal_covariates_value_type=<DataValueOpts.ANY: 1>, temporal_targets_value_type=<DataValueOpts.ANY: 1>, temporal_treatments_value_type=<DataValueOpts.ANY: 1>, event_covariates_value_type=<DataValueOpts.ANY: 1>, event_targets_value_type=<DataValueOpts.ANY: 1>, event_treatments_value_type=<DataValueOpts.ANY: 1>, requires_all_temporal_data_samples_aligned=False, requires_all_temporal_data_regular=False, requires_all_temporal_data_index_numeric=False, requires_all_temporal_containers_shares_index=True), prediction_requirements: Union[tempor.models.clairvoyance2.interface.requirements.PredictionRequirements, NoneType] = None, treatment_effects_requirements: Union[tempor.models.clairvoyance2.interface.requirements.TreatmentEffectsRequirements, NoneType] = None)[source]

Bases: object

dataset_requirements : DatasetRequirements = DatasetRequirements(requires_static_covariates_present=False, requires_no_missing_data=False, static_covariates_value_type=<DataValueOpts.ANY: 1>, temporal_covariates_value_type=<DataValueOpts.ANY: 1>, temporal_targets_value_type=<DataValueOpts.ANY: 1>, temporal_treatments_value_type=<DataValueOpts.ANY: 1>, event_covariates_value_type=<DataValueOpts.ANY: 1>, event_targets_value_type=<DataValueOpts.ANY: 1>, event_treatments_value_type=<DataValueOpts.ANY: 1>, requires_all_temporal_data_samples_aligned=False, requires_all_temporal_data_regular=False, requires_all_temporal_data_index_numeric=False, requires_all_temporal_containers_shares_index=True)
prediction_requirements : PredictionRequirements | None = None
treatment_effects_requirements : TreatmentEffectsRequirements | None = None
class tempor.models.clairvoyance2.interface.SavableModelMixin[source]

Bases: object

params : DotMap
inferred_params : DotMap
save(path: str) → None[source]
classmethod load(path: str)[source]
class tempor.models.clairvoyance2.interface.TimeIndexHorizon(time_index_sequence: collections.abc.Sequence[pandas.core.indexes.range.RangeIndex | pandas.core.indexes.datetimes.DatetimeIndex | pandas.core.indexes.base.Index])[source]

Bases: Horizon

time_index_sequence : Sequence[RangeIndex | DatetimeIndex | Index]
classmethod future_horizon_from_dataset(data: Dataset, forecast_n_future_steps: int, time_delta: int | float | datetime64 = 1) → TimeIndexHorizon[source]
to_numpy_time_series(padding_indicator: float = -999.0, max_len: int | None = None)[source]
to_torch_time_series(padding_indicator: float = -999.0, max_len: int | None = None, **torch_tensor_kwargs)[source]
class tempor.models.clairvoyance2.interface.TransformerModel(params: dict[str, Any] | None = None)[source]

Bases: BaseModel, ABC

check_data_requirements_transform(data: Dataset, **kwargs)[source]
transform(data: Dataset, **kwargs) → Dataset[source]
inverse_transform(data: Dataset, **kwargs) → Dataset[source]
fit_transform(data: Dataset, **kwargs) → Dataset[source]
check_model_requirements() → None[source]
params : DotMap
inferred_params : DotMap
class tempor.models.clairvoyance2.interface.TreatmentEffectsModel(params: dict[str, Any] | None = None)[source]

Bases: PredictorModel, ABC

check_data_requirements_predict_counterfactuals(data: Dataset, sample_index: int, treatment_scenarios: Sequence[TimeSeries | EventSamples], horizon: Horizon | None, **kwargs)[source]
predict_counterfactuals(data: Dataset, sample_index: int, treatment_scenarios: Sequence[DataFrame | ndarray | TimeSeries | EventSamples], horizon: Horizon | None, **kwargs) → Sequence[TimeSeries | EventSamples][source]
fit(data: Dataset, horizon: Horizon | None = None, **kwargs) → TreatmentEffectsModel[source]
check_model_requirements() → None[source]
params : DotMap
inferred_params : DotMap
class tempor.models.clairvoyance2.interface.TreatmentEffectsRequirements(treatment_data_structure: tempor.models.clairvoyance2.interface.requirements.DataStructureOpts = <DataStructureOpts.TIME_SERIES: 1>, min_timesteps_treatment_when_fit: int = 1, min_timesteps_treatment_when_predict: int = 1, min_timesteps_treatment_when_predict_counterfactual: int = 1)[source]

Bases: object

treatment_data_structure : DataStructureOpts = 1
min_timesteps_treatment_when_fit : int = 1
min_timesteps_treatment_when_predict : int = 1
min_timesteps_treatment_when_predict_counterfactual : int = 1