Closed jolyonb closed 5 years ago
Decrease in coverage is due to the python3 orthogonal/unitary matrix samplers.
Welp, all the code is done. I need to write a bunch of tests now, but that's a job for tomorrow.
Sorting out the code to generate unit determinant matrices and zero determinant matrices was rather messy. I welcome any suggestions to clean things up.
Pushed tests to here so I can test python3 for orthogonal and unitary sampling. Almost ready for review!
Sorry that I'm using Travis for testing. I haven't set up a python3 test environment. Perhaps I should...
In other news, this is one of our biggest PRs!
Weird... Looks like codecov only got one of the two runs. Anyway, I have one more test to writenow...
Yeah, looks like only got python 3 tests....weird
@ChristopherChudzicki This is ready for review. I also updated codecov options a little.
Cool, I will review ~tomorrow morning~ right after lunch.
I'm just about to push a new commit once I get the tests passing in py2... Note that I've added a whole lot more tests for the SquareMatrices
case (apparently I forgot to test all the different features).
Well, that took longer than I had hoped. Was good that I put in tests for all of the properties; I caught some bugs! Ready for re-review.
Merging #222 into master will not change coverage. The diff coverage is
100%
.
@@ Coverage Diff @@
## master #222 +/- ##
======================================
Coverage 100% 100%
======================================
Files 27 28 +1
Lines 2211 2580 +369
======================================
+ Hits 2211 2580 +369
This PR brings the long-awaited matrix sampling structure together. It largely follows the structure laid out in #188; biggest difference is that I went for a
dimension
property for square matrices instead ofshape
. We can now sample matrices of all types!Resolves #188 Resolves #148 Resolves #83 (about time!)