Added matrix_division support for QuArray. #34
Conversation
| @@ -109,7 +109,7 @@ Base.scale(qarr::AbstractQuArray, num::Number) = scale(num, qarr) | |||
| *(num::Number, qarr::AbstractQuArray) = scale(num, qarr) | |||
| *(qarr::AbstractQuArray, num::Number) = scale(qarr, num) | |||
| /(qarr::AbstractQuArray, num::Number) = scale(1/num, qarr) | |||
|
|
|||
| \(qarr::AbstractQuArray, v::AbstractQuArray) = QuArray(\(coeffs(qarr), coeffs(v))) | |||
There was a problem hiding this comment.
You should determine the return type using similar_type (see +,-,* for example). It's probably better to restrict the arguments to 2 AbstractQuMatrix and AbstractQuMatrix and AbstractQuVector (so you need two separate functions). Anything else won't work.
| @@ -110,6 +108,17 @@ Base.scale(qarr::AbstractQuArray, num::Number) = scale(num, qarr) | |||
| *(qarr::AbstractQuArray, num::Number) = scale(qarr, num) | |||
| /(qarr::AbstractQuArray, num::Number) = scale(1/num, qarr) | |||
|
|
|||
| function \{B<:OrthonormalBasis}(qm1::AbstractQuMatrix{B}, qm2::AbstractQuMatrix{B}) | |||
| div = \(rawcoeffs(qm1),rawcoeffs(qm2)) | |||
|
Please add brief comments stating what the functions are doing (solving linear systems) and some simple tests. |
|
@acroy the tests seem to pass on julia 0.3 and fail on the other julia 0.4, and also I have hard coded the tests, any better way around would be really helpful. |
| # tests for Matrix Division | ||
| v1 = [0.5+0*im, 0.+0.*im] | ||
| qv1 = normalize!(QuArray(v1)) | ||
| @assert coeffs(\(sigmaz, sigmax)) == sparse([0. 1.;-1. 0.]) |
There was a problem hiding this comment.
Try to use full(sigmaz) and full(sigmax) here and in the next line. As I said via PM it might actually be better for the Pauli matrices to be dense anyways, but this can be done separately.
|
Since #33 made some changes in arraymath you will need to rebase against master before we are able to merge. |
A*X == B where A,X,B are
AbstractQuArray.