Allow passing `weight=` or `solution_scale=` to `LLL()`

[user202729]

Specifying a weight for the different columns in LLL is a very common use case. In practice, this means you performing LLL to find shortest vector(s), but instead of the usual norm $∑ x_i^2$, you use the norm $∑ (w_i x_i)^2$ where ${w_i}$ are the weights. ($w_i$ must be all nonnegative.)

One case where this is used is in the current implementation of small_roots

    if X is None:
        X = (0.5 * N**(beta**2/delta - epsilon)).ceil()
    verbose("X = %s" % X, level=2)

⋮

    B = Matrix(ZZ, len(g), delta*m + max(delta,t) )
    for i in range(B.nrows()):
        for j in range( g[i].degree()+1 ):
            B[i, j] = g[i][j]*X**j

    B = B.LLL(**kwds)

    f = sum([ZZ(B[0, i]//X**i)*x**i for i in range(B.ncols())])
    R = f.roots()

Here the X**j is the weight of column j — notice all elements of column j are multiplied by X**j then they're immediately divided after running the LLL.

Alternatively we can give a solution_scale which is essentially the inverse of weight.

Possible extension: pass that to approximate_closest_vector too.