skpro.workflow.manager package

Submodules

skpro.workflow.manager.data module

class skpro.workflow.manager.data.DataManager(X=None, y=None, split=0.2, name=None, random_state=None)

Bases: object

A helper to manage datasets more easily. Test/training split is carried out behind the scenes whenever new data is being assigned

Parameters:

• X (np.array | string) – Features or ‘boston’, ‘diabetes’ to load sklearn datasets, url of file
• y (np.array) – Labels
• split (float, default=0.2) – Train/test split
• name (string, default=None) – Optional name to be used in the object representation
• random_state (int, default=None) – Optional random state to be used during split

X_train

Training features

Type:np.array

X_test

Training labels

Type:np.array

y_train

Test features

Type:np.array

y_test

Test labels

Type:np.array

X

clone()

Clones the data manager

Returns: Return type:A copy of the data manager itself

data(copy=True)

Returns the data

Parameters:copy (boolean, default=True) – If false, reference copy will be used Returns: Return type:X, y

shuffle(random_state=None)

Shuffles the data

Parameters:random_state (int, default=None) – Optional random state Returns: Return type:None

y

skpro.workflow.manager.data.load_file(file, return_X_y=False)

skpro.workflow.manager.data.load_url(url, return_X_y=False)

skpro.workflow.manager.data.retrieve_data(url)

skpro.workflow.manager.models module

class skpro.workflow.manager.models.ModelManager

Bases: object

Model manager

all()

group(name)

info()

register(model, tuning=None, group=None, distinguish_tuning=True)

Module contents