Negating a Polynomial Returns a Polynomial with an Inconsistent Grid

Negating a polynomial via the unary operator - utilizing the dunder method __neg__() returns a polynomial with different Grid.

Consider the following example, a polynomial with equidistant points as the interpolating grid:

>>> mi = mp.MultiIndexSet.from_degree(3, 5, 1)
>>> coeffs = np.random.rand(len(mi))
>>> grd = mp.Grid.from_value_set(mi, np.linspace(-1, 1, mi.poly_degree+1))[:, np.newaxis])  # equidistant point
>>> lag_poly = mp.LagrangePolynomial(mi, coeffs, grid=grd)
>>> lag_poly.grid.generating_points
array([[ 1. , -1. ,  1. ],
       [ 0.6, -0.6,  0.6],
       [ 0.2, -0.2,  0.2],
       [-0.2,  0.2, -0.2],
       [-0.6,  0.6, -0.6],
       [-1. ,  1. , -1. ]])
>>> neg_lag_poly = -lag_poly
>>> neg_lag_poly.grid.generating_points
array([[ 1.        , -1.        ,  1.        ],
       [-1.        ,  1.        , -1.        ],
       [ 0.30901699, -0.30901699,  0.30901699],
       [-0.30901699,  0.30901699, -0.30901699],
       [ 0.80901699, -0.80901699,  0.80901699],
       [-0.80901699,  0.80901699, -0.80901699]])

The generating points have been reverted to the default (the Chebyshev-Lobatto) nodes. Specifically, the grid before and after negation are not the same which is an unexpected behavior because negating a polynomial should only alter the coefficients (change their signs):

>>> lag_poly.grid == neg_lag_poly.grid
False

Upon inspection, it seems the negation does not include the grid argument when constructing a new instance of negated polynomial.