scxpand.util.classes

Functions

ensure_model_type(model_type)

Convert string to ModelType enum if needed, with validation.

Classes

BaseParams()	Abstract base class for all parameter classes.
DataAugmentParams([mask_rate, noise_std, ...])
DataLoaderParams(batch_size, shuffle, ...)
LRSchedulerParams(lr_scheduler_type, ...)
LRSchedulerType(*values)
ModelType(*values)	Enumeration of supported model types.
OptimizerParams(optimizer_type, adam_betas, ...)
OptimizerType(*values)	Enumeration of supported optimizer types.
SamplerType(*values)	Enumeration of supported sampler types.

class scxpand.util.classes.BaseParams

Abstract base class for all parameter classes.

Provides a common interface for parameter classes with a shared get_model_type method. All parameter classes should inherit from this base class to ensure consistency.

abstractmethod classmethod get_model_type()

Return the model type identifier for this parameter class.

Return type:

ModelType

Returns:

ModelType enum value for the model type

class scxpand.util.classes.DataAugmentParams(mask_rate=0.0, noise_std=0.0, soft_loss_beta=1.0)

__init__(mask_rate=0.0, noise_std=0.0, soft_loss_beta=1.0)

mask_rate: float = 0.0

noise_std: float = 0.0

soft_loss_beta: float = 1.0

class scxpand.util.classes.DataLoaderParams(batch_size, shuffle, sampler_type)

__init__(batch_size, shuffle, sampler_type)

batch_size: int

sampler_type: SamplerType

shuffle: bool

class scxpand.util.classes.LRSchedulerParams(lr_scheduler_type, lr_scheduler_config)

__init__(lr_scheduler_type, lr_scheduler_config)

lr_scheduler_config: dict

lr_scheduler_type: LRSchedulerType

class scxpand.util.classes.LRSchedulerType(*values)

CONSTANT_LR = 'ConstantLR'

COSINE_ANNEALING_LR = 'CosineAnnealingLR'

NO_SCHEDULER = 'NoScheduler'

ONE_CYCLE_LR = 'OneCycleLR'

REDUCE_LR_ON_PLATEAU = 'ReduceLROnPlateau'

STEP_LR = 'StepLR'

class scxpand.util.classes.ModelType(*values)

Enumeration of supported model types.

AUTOENCODER = 'autoencoder'

LIGHTGBM = 'lightgbm'

LOGISTIC = 'logistic'

MLP = 'mlp'

SVM = 'svm'

class scxpand.util.classes.OptimizerParams(optimizer_type, adam_betas, weight_decay, max_grad_norm, init_learning_rate)

__init__(optimizer_type, adam_betas, weight_decay, max_grad_norm, init_learning_rate)

adam_betas: tuple

init_learning_rate: float

max_grad_norm: float

optimizer_type: OptimizerType

weight_decay: float

class scxpand.util.classes.OptimizerType(*values)

Enumeration of supported optimizer types.

ADAM = 'Adam'

ADAMW = 'AdamW'

SGD = 'SGD'

class scxpand.util.classes.SamplerType(*values)

Enumeration of supported sampler types.

BALANCED_LABELS = 'balanced_labels'

BALANCED_TYPES = 'balanced_types'

RANDOM = 'random'

scxpand.util.classes.ensure_model_type(model_type)

Convert string to ModelType enum if needed, with validation.

Return type:

ModelType