Tests are broken with scipy 1.16.0

[Tommimon]

I tried to run the tests but 2 of them fail with a version of scipy greater or equal to 1.16.0 with the following log:

FAIL: test_82 (numerical.specialfunctions.TestNumericalSpecialFunctions.test_82)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/root/ModularCalculator/tests/testrunner.py", line 66, in do_test
    raise err
  File "/root/ModularCalculator/tests/testrunner.py", line 29, in do_test
    self.assertEqual(actual, expected, msg=expr)
AssertionError: Number('5945711222452133', '4503599627370496') != Number('1486427805088563', '1125899906842624') : sp_nctdtrit(1, 1, 0.5)

======================================================================
FAIL: test_87 (numerical.specialfunctions.TestNumericalSpecialFunctions.test_87)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/root/ModularCalculator/tests/testrunner.py", line 66, in do_test
    raise err
  File "/root/ModularCalculator/tests/testrunner.py", line 29, in do_test
    self.assertEqual(actual, expected, msg=expr)
AssertionError: Number('-1661586938413673', '4503599627370496') != Number('-6646347753654691', '18014398509481984') : sp_log_ndtr(0.5)

----------------------------------------------------------------------

I was running this test in an Ubuntu container but I tried also other distros with the same result.
The problem seams that scipy changed something in how these functions are calculated because I get slightly different results with different versions:

Python 3.12.3 (main, Nov  6 2025, 13:44:16) [GCC 13.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import scipy
>>> scipy.__version__
'1.15.3'
>>> print(scipy.special.nctdtrit(1,1,0.5))
1.3202130989218857
>>> 


Python 3.12.3 (main, Nov  6 2025, 13:44:16) [GCC 13.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import scipy
>>> scipy.__version__
'1.16.0'
>>> print(scipy.special.nctdtrit(1,1,0.5))
1.3202130993877088
>>> 

I think the best approach to solve this is to add a condition in the tests and if a version of scipy grater than 1.16.0 is detected, the new values should be used:

• Number('5945711222452133', '4503599627370496') for sp_nctdtrit(1, 1, 0.5)
• Number('-1661586938413673', '4503599627370496') for sp_log_ndtr(0.5)

Let me know if you can reproduce the same issue and if I can help to get the tests to work also with newer versions of scipy.

[JordanL2]

Yeah some tests always break with a new scipy version. I don't think it's necessary to support multiple scipy versions, I'll just update the Flatpak to use 1.16.0 and fix the tests.