Use assert_almost_equal in test_template_comparison.py

[olebole]

On 32-bit x86, test_template_match_multidim_spectrum fails during build due to slightly different floating points:

    def test_template_match_multidim_spectrum():
        """[…]"""
        np.random.seed(42)
    
        # Create test spectra
        spec_axis1 = np.linspace(0, 50, 50) * u.AA
        spec_axis2 = np.linspace(0, 50, 50) * u.AA
    
        spec = Spectrum1D(spectral_axis=spec_axis1,
                          flux=np.random.sample(50) * u.Jy,
                          uncertainty=StdDevUncertainty(np.random.sample(50)))
        multidim_spec = Spectrum1D(spectral_axis=spec_axis2,
                                   flux=np.random.sample((2, 50)) * u.Jy,
                                   uncertainty=StdDevUncertainty(np.random.sample((2, 50))))
    
        # Get result from template_match
        tm_result = template_comparison.template_match(spec, multidim_spec)
    
>       assert tm_result[1] == 250.26870401777543
E       assert 250.26870401777546 == 250.26870401777543

The equal here should be an "almost equal". This patch fixes this in test_template_comparison.py. Although there are direct comparisons in many other test files, I pragmatically left them untouched since they didn't cause a failure, and I am not always sure whether a direct equal was meant there.

[eteq]

Yep, this is definitely better. Thanks @olebole!

Longer term it might be better to replace these tests with something that's more heuristic anyway (i.e. a "is the chi squared better in this case versus that other one" instead of an exact numerical test), but this is definitely an improvement.

[eteq]

cc @javerbukh as an FYI to just for future related development on the redshift-finding machinery.