fillet/chamfer float comparison with zero via isclose abs_tol=1e-14

[jdegenstein]

Fixes failing test on apple silicon. Closes issue #478

This fixes the following previously failing test:

self = <test_build_part.TestMakeBrakeFormed testMethod=test_make_brake_formed>
    def test_make_brake_formed(self):
        # TODO: Fix so this test doesn't raise a DeprecationWarning from NumPy
        with BuildPart() as bp:
            with BuildLine() as bl:
                m1 = Polyline((0, 0), (5, 6), (10, 1))
>               fillet(m1.vertices(), 1)

As suspected, the behavior of floats on apple silicon is just a bit different. Comparison of a float with 0.0 is a possible source of architecture specific issues. Current recommendation adopted here is to use math.isclose with an appropriate abs_tol e.g. 1e-14.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (3389545) 96.17% compared to head (f79b455) 96.17%.

Additional details and impacted files

@@           Coverage Diff           @@
##              dev     #515   +/-   ##
=======================================
  Coverage   96.17%   96.17%           
=======================================
  Files          24       24           
  Lines        7481     7481           
=======================================
  Hits         7195     7195           
  Misses        286      286           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.