Closed
Description
Reproducing code example:
I've been using this context manager in some code to control array printing:
import contextlib
import numpy as np
import numpy.core.arrayprint as arrayprint
@contextlib.contextmanager
def printoptions(strip_zeros=False, **kwargs):
"""Allows you to set NumPy array print formatting locally.
Taken from:
http://stackoverflow.com/questions/2891790/pretty-printing-of-numpy-array
"""
origcall = arrayprint.FloatFormat.__call__
def __call__(self, x, strip_zeros=strip_zeros):
return origcall.__call__(self, x, strip_zeros)
arrayprint.FloatFormat.__call__ = __call__
original = np.get_printoptions()
np.set_printoptions(**kwargs)
yield
np.set_printoptions(**original)
arrayprint.FloatFormat.__call__ = origcallFloatFormat was removed in NumPy 1.18.0 but prior to that #9941 removed the strip_zeros kwarg from FloatFormat.__call__. This seems to be a backwards incompatible change. It also seems like that kwarg likely did not have a unit test.
Regardless, I'm trying to figure out how to reproduce the behavior of strip_zeros=False prior to PR #9941.
Error message:
Numpy/Python version information:
In [10]: import sys, numpy; print(numpy.__version__, sys.version)
1.18.1 3.6.10 | packaged by conda-forge | (default, Apr 6 2020, 14:52:36)
[GCC 7.3.0]