Description

sklearn.datasets.make_circles always generates an even number of samples, even when specifically asked for an uneven number.

Steps/Code to Reproduce

Example:

from sklearn.datasets import make_circles

n_samples = 101
xs, ls = make_circles(n_samples=n_samples)
assert(len(xs)==len(ls)==n_samples)

Expected Results

No Assertion Error should be thrown (in the example) OR the amount of data points generated should match the value supplied to the function OR a warning should be given.

Possible fix: Calculate the number of data points per cluster as in make_moons

Actual Results

Assertion Error, No warning, lower number of points returned than asked for.

Versions

Windows-10-10.0.15063
('Python', '2.7.11 (v2.7.11:6d1b6a68f775, Dec 5 2015, 20:32:19) [MSC v.1500 32 bit (Intel)]')
('NumPy', '1.13.0')
('SciPy', '0.16.1')
('Scikit-Learn', '0.19.1')