scikit-plot should decorate object instead of add new methods

[leiserfg]

No description provided.

[reiinakano]

Hi @leiserfg ! This is an interesting suggestion. Could you expound with an example of what you want to see?

[leiserfg]

that's what i expected

from sklearn.datasets import load_digits as load_data
from sklearn.naive_bayes import GaussianNB

import matplotlib.pyplot as plt
from scikitplot import classifier_factory

X, y = load_data(return_X_y=True)
nb = GaussianNB()

#canvas can be a subplot instead a plot (imagine that you wanna compare classifiers)
plotter = classifier_factory(nb, canvas=plot)      
plotter.plot_roc_curve(X, y, random_state=1)
plt.show()

[reiinakano]

Hi @leiserfg , sorry for the long time it took to answer this.

Anyway, all scikit-plot functions have an ax parameter that provides this functionality.

In the case of having a particular subplot to use as a canvas, you would do this.

from sklearn.datasets import load_digits as load_data
from sklearn.naive_bayes import GaussianNB

import matplotlib.pyplot as plt
from scikitplot import classifier_factory

X, y = load_data(return_X_y=True)
nb = GaussianNB()

#canvas can be a subplot instead a plot (imagine that you wanna compare classifiers)
plotter = classifier_factory(nb)      
plotter.plot_roc_curve(X, y, random_state=1, ax=plot)
plt.show()