Support multivariate polynomials

[mjcrock]

The galois.Poly class currently supports univariate polynomials, but not multivariate polynomials, over $GF(p^m)$. I would like to propose supporting this functionality in galois.

The following code snippet demonstrates how to generate a random univariate polynomial which we may evaluate on scalars, arrays, or square matrices:

from galois import GF, Poly

# specify desired polynomial parameters
field, degree = GF(3**5), 4

# randomly generate a univariate polynomial, e.g.
# Poly(41x^4 + 16x^3 + 147x^2 + 232x + 202, GF(3^5))
f_uni = Poly.Random(degree, field)

By contrast, the following pseudo-code demonstrates how one might generate a random multivariate polynomial which we may evaluate on arrays of shape (..., n_variables) only:

from galois import GF, Poly

# specify desired polynomial parameters
field, n_variables, degree = GF(3**5), 2, 2

# randomly generate a multivariate polynomial, e.g.
# Poly(241x_1^2 + 207x_0^2 + 36x_1x_0 + 5x_1 + 19x_1 + 112, GF(3^5))
f_multi = Poly.Random(degree, n_variables, field)

I look forward to hearing any thoughts and suggestions!