numpy_integration.md

Integration with numpy

Every logiq's object have the .npm() method that returns a numpy.matrix object.

>>> M = Op.Y.npm()
>>> type(M)
numpy.matrix

>>> stdbasis.npm() == matrix.Id(2).npm()
matrix([[ True,  True],
        [ True,  True]])

This matrix is the real data of the object so do not modify it!
You can easily copy it if you need.

The only object with a different behaviour is Qstate.

To get the state as a numpy.matrix you have to get the internal state as a vector and then call the .npm():

>>> q = Qstate.random(2)
>>> intState = q.state #this contains the internal state as a vector
>>> intState.npm()
matrix([[-0.41113669+0.59933783j],
        [ 0.62821559-0.27767959j]])

The mainly motivations of this are to avoid involuntary modifications of the internal state of a Qstate:

>>> q = Qstate.random(2)
>>> intState = q.state # this is a copy of the real internal state
>>> Op.Y | q
>>> intState.npm() == q.state.npm()
matrix([[False],
        [False]])

---

Creation using numpy's objects

Except for Qstate every other classes generate their data using numpy.matrix(input), so every type of input accepted from numpy.matrix could be a good input.

The classes vector, matrix, Op and Basis initialize their content in a way similar to this:

class obj:
    def __init__(self, input):
        self.content = numpy.matrix(input)

So every input allowed from numpy.matrix are possible good input for the class and this cause a great integration with numpy's classes!