NumPy - universal functions

INTRODUCTION

Universal functions operate elementwise on an array, producing an array as output.


# np.cos(a), np.sin(a), np.tan(a),
# np.arccos(a), np.arcsin(a), np.arctan(a),
# np.cosh(a), np.sinh(a), np.tanh(a),
# np.arccosh(a), np.arcsinh(a), np.arctanh(a),
# np.exp(a), np.sqrt(a), np.abs(a), np.sign(a), np.round(a),
# np.log(a), np.log2(a), np.log10(a)

# np.sum(a, axis=k) - sum of array elements over a given axis
# np.prod(a, axis=k)
# np.cumsum(a, axis=k)
# np.cumprod(a, axis=k)
# np.min(a, axis=k)
# np.max(a, axis=k)
# np.argmin(a, axis=k)
# np.argmax(a, axis=k)

# np.median(a, axis=k) - return the median (the middle value of a sorted copy of a)
# np.average(a, axis=k, weights=w) - compute the weighted average
# np.mean(a, axis=k) - compute the arithmetic mean 
# np.var(a, axis=k, ddof=0) - compute the variance
# np.std(a, axis=k, ddof=0) - compute the standard deviation

# np.all(a, axis=k) - return numpy.bool_
# np.any(a, axis=k) - return numpy.bool_
np.all([[True, False], [True, True]], axis=0)   # array([ True, False])
np.any([[True, False], [False, False]], axis=1)   # array([ True, False])

a = np.arange(-2, 3)   # array([-2, -1,  0,  1,  2])
np.all(a > 0)   # False
np.any(a > 0)   # True
np.abs(a)   # array([2, 1, 0, 1, 2])