Update to support SF v0.11 #3
Conversation
Codecov Report
@@ Coverage Diff @@
## master #3 +/- ##
=======================================
Coverage 100% 100%
=======================================
Files 11 5 -6
Lines 1493 299 -1194
=======================================
- Hits 1493 299 -1194
Continue to review full report at Codecov.
|
| numeric_level = getattr(logging, logLevel.upper(), 10) | ||
| # defaults | ||
| TOL = 1e-2 | ||
| HBAR = 2 |
There was a problem hiding this comment.
worthwhile using an unconventional value here (to help detect bugs)?
There was a problem hiding this comment.
or will you leave that up to travis?
There was a problem hiding this comment.
I actually changed this to HBAR = 1.7 as in SF, and discovered that a huge majority of the tests have hbar = 2 assumptions built into them 🤦♂ I didn't want to go out of scope, but I'll make a new issue to fix that.
| assert np.allclose(state.fock_prob([1, 1]), exp[1], rtol=tol) | ||
| assert np.allclose(state.fock_prob([0, 2]), exp[2], rtol=tol) | ||
|
|
||
| def test_1x2_quadrature_operators(self, eng, tol): |
There was a problem hiding this comment.
Yep, noticed it wasn't being tested previously, and GitHub/Codecov wouldn't pass the coverage checks without it
| self.H = bose_hubbard(1, 2, self.J, self.U) | ||
| self.Hquad = get_quad_operator(self.H, hbar=self.hbar) | ||
|
|
||
| def test_hbar_outside_context(self): |
There was a problem hiding this comment.
no longer need these hbar tests?
There was a problem hiding this comment.
No, this is because of the change in how hbar is handled in SF v0.11; hbar is no longer part of the engine/program context, but a separate global variable.
| """Test S2gate produces correct cov and means""" | ||
| self.logTestName() | ||
| # NOTE: There is currently a bug in strawberry fields, |
No description provided.