search

materialsproject / pymatgen

Python Materials Genomics (pymatgen) is a robust materials analysis code that defines classes for structures and molecules with support for many electronic structure codes. It powers the Materials Project.
https://pymatgen.org
Other
1.52k stars 867 forks source link

tests use equality to compare floating point numbers #4158

Open drew-parsons opened 2 weeks ago

drew-parsons commented 2 weeks ago

Floating point numbers (angles) are compared in transformation tests at

https://github.com/materialsproject/pymatgen/blob/bd9fba9ec62437b5b62fbd0b2c2c723216cc5a2c/tests/transformations/test_advanced_transformations.py#L840

Using equality to compare floating point numbers inevitably fails e.g. https://buildd.debian.org/status/fetch.php?pkg=pymatgen&arch=amd64&ver=2024.10.29%2Bdfsg1-1&stamp=1731061592&raw=0

        # make sure that the orthorhombic supercell is different from the cubic cell
        assert not np.array_equal(
            supercell_generator_cubic.transformation_matrix,
            supercell_generator_orthorhombic.transformation_matrix,
        )
        assert transformed_orthorhombic.lattice.abc != transformed_cubic.lattice.abc
        # only angels are expected to be the same because of force_90_degrees = True
>       assert transformed_cubic.lattice.angles == transformed_orthorhombic.lattice.angles
E       assert (89.999970160...0000000086139) == (89.999970160...0000000086142)
E         
E         At index 2 diff: 90.00000000086139 != 90.00000000086142
E         Use -v to get more diff

/<<PKGBUILDDIR>>/.pybuild/test_python3.12/tests/transformations/test_advanced_transformations.py:840: AssertionError

Floating point comparison needs to be used instead

assert_allclose(transformed_cubic.lattice.angles, lattice.angles)
DanielYang59 commented 1 week ago

Thanks for reporting, would fix this.