Use SB with SymPy polynomial expression

[bqth29]

It is now possible to use the Simulated Bifurcation algorithm by defining the polynomial instances directly from SymPy polynomial expression. This makes the call to the functions more user-readable and more mathematical.

Otherwise, it is of course still possible to define the polynomial instances using coefficient tensors. However, the keywords matrix, vectorand constant are no longer required. The tensors can simply be passed in an arbitrary order for more liberty in the call to the functions.

For example, if you want to solve $\underset{\lbrace0, 1 \rbrace^2}{\hbox{maximize }} x_1^2 - 3x_1x_2 -5x_2^2 + 4x_2 + 7$ you have to strategies:

1. Use a polynomial expression

import simulated_bifurcation as sb
from sympy import poly, symbols

x1, x2 = symbols("x1 x2")
expression = poly(x1 ** 2 - 3 * x1 * x2 - 5 * x2 ** 2 + 4 * x2 + 7)

sb.minimize(expression, domain="binary")

2. Use coefficient tensors

import simulated_bifurcation as sb
import torch

matrix = torch.tensor([[1, -3], [0, -5]])
vector = torch.tensor([0, 4])
constant = 7

sb.minimize(matrix, vector, constant, domain="binary")
# equivalent to: sb.minimize(matrix, constant, vector, domain="binary")
# equivalent to: sb.minimize(constant, vector, matrix, domain="binary")
# etc...

⚠️ Some breaking changes occured when merging #37 so we recommend reading the PR note