Fix dtype of CZ matrix

[mlazzarin]

I may have found an unwanted behavior. The dtype of CZ matrix in src/qibo/backends/matrices.py is int64 but I think it should be equal to the dtype property of qibo.backends.matrices.Matrices class.

[stavros11]

Thanks for finding this. Looking at the implementation of self.backend.cast for the numpy and other backends, I believe that the final array, here self._CZ will have the backend complex type regardless of the type of the original array m, right? If this is true the self.dtype used in all matrix definitions here is essentially useless.

I am not sure why the macos tests fail though.

[codecov]

Codecov Report

Merging #503 (0e062f0) into rmbackends (f5f0e9f) will not change coverage.
The diff coverage is 100.00%.

@@             Coverage Diff             @@
##           rmbackends     #503   +/-   ##
===========================================
  Coverage       99.99%   99.99%           
===========================================
  Files              84       84           
  Lines           11853    11854    +1     
===========================================
+ Hits            11852    11853    +1     
  Misses              1        1           

Flag	Coverage Δ
unittests	99.99% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/qibo/backends/matrices.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f5f0e9f...0e062f0. Read the comment docs.

[mlazzarin]

Given that we merged #500, I changed the base branch to multiqubit. Shall we merge this?

[scarrazza]

@mlazzarin, yes, sure. I was planning to merge directly to master, but we can instead merge this in multiqubit.