Top boundary reference and method

[dbrakenhoff]

Adds a reference to the top boundary element. Mostly useful for transient models to get access to top boundary element and inspect the top boundary condition. Also added to steady model to be consistent.

In steady, ml.aq.topbc or inhom.topbc holds reference to ConstantStar, AreaSinkInhom or XsectionAreaSinkInhom depending on the type of model/inhom.

In transient xsection.topbc holds reference to HstarXsection or AreaSinkXsection.

Adds method .get_bc() to transient element class to obtain time series of any .tsandbc which could be useful for plotting or comparison to other time series etc.

riv = tft.XsectionMaq(..., tsandhstar=[(0, 1), (1, 2), (2, 0)])
riv.topbc  # --> HstarXsection element

t = np.arange(0, 2, 0.1)
riv.topbc.get_bc(t)  # --> returns array of hstar values

[mbakker7]

In the current transient code, aquifer 0 can be bounded on top by a leaky layer. If the topboundary='semi', then flow is semi-confined and in a cross-section model, tsandhstar can be specified to model a varying water table above the semi-confining top. But when topboundary='leak', then the top of the leaky layer is impermeable. This option is useful when a sand layer is covered by a clay layer and during a transient simulation water can leaky into or out of the clay layer. I don't think this option is used much (and probably poorly documented). But in this PR I get the idea that it is removed. Is that the case?

[dbrakenhoff]

But in this PR I get the idea that it is removed. Is that the case?

No, in timflow.steady there were references to the top being "leaky" which I think were never renamed to "semi"? In timflow.steady there are only two options, "conf" and "semi" so I modified the docs/error messages slightly to reflect that.

[mbakker7]

Sorry, I missed that. There is indeed no "leaky" in timflow.steady. So let's merge this.