model.matrices.c is being coherrsed to int

[dannyopts]

When accessing model.matrices.c the value is coherrsed to int (see failing test below)

This is because when c is created, it has a default value of 0, which is used to create the initial vector.

create_vector(vars, ds.coeffs, fill_value=0, shape=shape)

switching this default to 0.0 fixes the issue.

I will raise a PR to fix this now

def test_matrices_float_c():
    m = Model()

    x = m.add_variables(pd.Series([0, 0]), 1, name="x")
    m.add_objective(x * 1.5)

    c = m.matrices.c
    assert np.all(c == np.array([1.5, 1.5])) 

Fails with

  assert np.all(c == np.array([1.5, 1.5]))

E assert False E + where False = <function all at 0x1044e5b30>(array([1, 1]) == array([1.5, 1.5])

[dannyopts]

This causes a bug when using the direct api that cost coeffs are rounded to ints

>>> m = Model()
>>> x = m.add_variables(name="x")
>>> m.add_objective(0.1 * x)
>>> m.solve("highs", "direct")
>>> m.objective.value
0.0
>>> m.solve("highs", "lp")
>>> m.objective.value
nan