* Fix compilation error
* Add implementation of pseudo-inverse
The implementation is based on this publication:
Courrieu, P. (2008). Fast Computation of Moore-Penrose Inverse Matrices, 8(2), 25–29. http://arxiv.org/abs/0804.4809
It is a fully templated implementation to guaranty type correctness.
* Add tests for pseudoinverse
* Apply suggestions from code review
Co-Authored-By: Mathieu Bresciani <brescianimathieu@gmail.com>
* Adapt fullRankCholesky tolerance to type
* Add pseudoinverse test with effectiveness matrix
* Fix coverage
* Fix rebase issue
* Fix SquareMatrix test, add null Matrix test
Before the print function was just implicitly called somewhere
and that's why we had 100% line coverage. With this we have actual
testing of the functions.
The original implementation with no wrapping on size_t is more readable
but the compiler errors with:
internal compiler error: in trunc_int_for_mode, at explow.c:55
I read up and it's apparently a loop optimization problem.
Inspired by https://stackoverflow.com/a/27224697/6326048
I used a far less readable implementation that works fine and
wrote a comment to explain it.
I had a look at the implementation of `slice`, and I found it odd that it doesn't have a copy loop. The current implementation does a raw memcpy of the underlying contiguous row-major data. As far as I can tell, this could only slice along rows. Interestingly, I found that the tests only tested for slicing along rows, so this bug would go unnoticed.
I added some tests that checks slicing along columns also. I have a feeling this would break, and we need to fix the implementation of `slice`. However I could be wrong, and hence I'm submitting these tests first to verify.
kritz
Daniel Agar
Matthias Grob
Julien Lecoeur
Julian Kent
Julian Kent
James Goppert
Tanja Baumann
Martina Rivizzigno
Oskar Weigl