[BUG] SurfaceFlux doesn't account for Soret effect

[RemDelaporteMathurin]

Describe the bug
When running a simulation with the Soret effect, I realised that setting a SurfaceFlux derived quantity would result in the wrong values. This is because the soret flag in compute seems to be ignored.

To Reproduce
Simple 1D steady state simulation, checking the flux balance with soret activated.

import numpy as np
import festim as F

my_model = F.Simulation()
my_model.mesh = F.MeshFromVertices(vertices=np.linspace(0, 1, num=100))
my_model.materials = F.Material(id=1, D_0=1, E_D=0, Q=2)
my_model.T = F.Temperature(value=700 + 1000 * F.x)

my_model.boundary_conditions = [
    F.DirichletBC(surfaces=[1], value=10, field=0),
    F.DirichletBC(surfaces=[2], value=1, field=0),
]
my_model.settings = F.Settings(
    absolute_tolerance=1e-10, relative_tolerance=1e-10, transient=False, soret=True
)

flux_left = F.SurfaceFlux(field=0, surface=1)
flux_right = F.SurfaceFlux(field=0, surface=2)
my_model.exports = [F.DerivedQuantities([flux_left, flux_right])]

my_model.initialise()
my_model.run()


# these two should be equal in steady state
print(flux_left.data)
print(flux_right.data)

Produces:

[370.7625458744958]
[1.1058666646404112]

Expected behavior
The two fluxes should have the same absolute value.

To convince us even further, try and add assert False in the compute method with soret:

    def compute(self, soret=False):
        flux = f.assemble(
            self.prop * f.dot(f.grad(self.function), self.n) * self.ds(self.surface)
        )
        if soret and self.field in [0, "0", "solute"]:
            assert False
            flux += f.assemble(
                self.prop
                * self.function
                * self.Q
                / (k_B * self.T**2)
                * f.dot(f.grad(self.T), self.n)
                * self.ds(self.surface)
            )
        return flux

And you will see everything runs fine.

Moreover, I found this TODO comment:

FESTIM/festim/exports/derived_quantities/derived_quantities.py

Line 157 in 40cdd22

# TODO need to support for soret flag in surface flux