since import algopy currently fails, UTPM == None.... and tests fails
==== Though to run successfully that test, another correction is needed in NUMPY
numpy\lib\function_base.py line 4276 has to be commented out.
if a.dtype.kind == "c":
raise TypeError("a must be an array of real numbers")
========RESULT========
nd.test('--doctest-modules', '--disable-warnings')
=============== 164 passed, 8 skipped, 1369 warnings in 42.86s ================
I am still pondering how to present this issue to Numpy.... it appears complicated
This is my dev environment
Python 3.11 Windows 10 numpy 1.26.4
algopy 0.5.7
In init.py for algopy you can comment out
testing
from numpy.testing import Tester
test = Tester().test
Because numpy (1.26.4) no longer has Tester. This code thus fails
====
Impact when fails
import doctest
import numdifftools as nd
nd.test('--doctest-modules', '--disable-warnings')
This nd.test fails ===== Because numdifftools, nd_algopy.py has this code
try:
import algopy
from algopy import UTPM
except ImportError:
UTPM = algopy = None
since import algopy currently fails, UTPM == None.... and tests fails
==== Though to run successfully that test, another correction is needed in NUMPY
numpy\lib\function_base.py line 4276 has to be commented out.
if a.dtype.kind == "c":
raise TypeError("a must be an array of real numbers")
========RESULT========
nd.test('--doctest-modules', '--disable-warnings')
=============== 164 passed, 8 skipped, 1369 warnings in 42.86s ================
I am still pondering how to present this issue to Numpy.... it appears complicated