Converting Numpy matrix in Qobj

[Simone Azeglio]

Hello! 

I'm trying to implement a Quantum PageRank algorithm with QuTip. The main problem is that I'm using NetworkX to extract the Google Matrix of my network and the return type is a Numpy matrix (ref: https://networkx.github.io/documentation/development/reference/generated/networkx.algorithms.link_analysis.pagerank_alg.google_matrix.html).

Well the problem is that I have to do a tensor product between G-matrix elements and a ket that I have defined with qutip.states.basis, but being the G-matrix a Numpy matrix I cannot do it. 

Thank you in advance for your help ! 

The code is :

...

G = nx.google_matrix(ER) #ER is my network , defined before 

#Projection operator Pi

piOperator = qt.states.projection(N*N, 1, 1)

#Psi0 state 

Psi0 = qt.states.zero_ket(N*N)

for i in xrange(0 , N-1 ):

    auxKet = qt.states.basis(N, i) #Auxiliary vector 

    Psi0 = qt.tensor(auxKet,np.sqrt(G[:,i]))

[jkur...@gmail.com]

Good Morning,

I've also had to work with outside data that are arrays/matrices. Here is what I've found to work (for what it's worth).

Qobj to numpy array / matrix:

data = q.full()

array = numpy.array(data)

matrix = numpy.matrix(data)

Numpy to Qobj

q = Qobj(array)

q = Qobj(matrix)

Numpy operations:

Matrix multiplication in numpy is '*" for matricies and (for 2d) numpy.dot(). (e.g. A.dot(B) for arrays A * B for matrices)

Element wise squaring is numpy.square() (works regardless of the dims)

Tensor on numpy things (not Qobjs):

We've been using numpy.kron(a,b) to tensor a and b (if I remember correctly this works with arrays and matrices)

Don't know if this is helpful not. :)

V/r

JK

[pietro Locatelli]

It was very useful.

Thank you very much