This rule raises an issue when a deprecated Numpy alias of a built-in type is used.

Why is this an issue?

In NumPy, some built-in types such as int have aliases in the form of numpy.int. However, these aliases have been deprecated and should not be used anymore.

The following deprecated aliases should be replaced with their built-in alternatives:

Deprecated name Equivalent built-in type

numpy.bool

bool

numpy.int

int

numpy.float

float

numpy.complex

complex

numpy.object

object

numpy.str

str

numpy.long

int

numpy.unicode

str

How to fix it

To fix this issue, make sure to replace deprecated NumPy type aliases with their corresponding built-in types.

Code examples

Noncompliant code example

import numpy as np
def foo():
    x = np.int(42)  # Noncompliant: deprecated type alias

Compliant solution

import numpy as np
def foo():
    x = 42  # Compliant

Resources

Documentation