Use a dummy CRS instead of the Dutch CRS in ribasim testmodels

[evetion]

#1339 added setting crs to our models, and made it mandatory to do so. Now all of our testmodels have EPSG:28992 set (the Dutch crs), but points are actually set with local coordinates like point(0,0) (https://github.com/Deltares/Ribasim/blob/main/python/ribasim_testmodels/ribasim_testmodels/allocation.py#L36).

Arguably, we need to set a dummy engineering crs for these models, like:

"""ENGCRS["Dummy CRS for testing",
    EDATUM["Unknown engineering datum"],
    CS[Cartesian,2],
        AXIS["X",unspecified,
            ORDER[1],
            LENGTHUNIT["unknown",0]],
        AXIS["Y",unspecified,
            ORDER[2],
            LENGTHUNIT["unknown",0]]]"""

#nitpick

[visr]

There is no EPSG for this? I feel like having such a long code in the TOML that is often used as example is worse.
They aren't really unknown either, they are known to be some imaginary cartesian coordinate system.

[evetion]

There is no EPSG for this? I feel like having such a long code in the TOML that is often used as example is worse. They aren't really unknown either, they are known to be some imaginary cartesian coordinate system.

That's the engineering part of the datum, cartesian (x, y) in an imaginery space somewhere. There's no EPSG code for it, apart from some deprecated example ones (EPSG:5806 comes to mind)

I agree it's ugly, but I would prefer it over being wrong 😄. And this is a logical extension from the crs PR, or would you want to limit it to EPSG codes only?

[visr]

I don't necessarily want to limit it to EPSG codes only, though we should stimulate people to use them when possible.

I agree it's ugly, but I would prefer it over being wrong

I prefer wrong, since wrong only affects plotting it on the map. We cannot plot it on any base map since the locations don't exist. Perhaps EPSG:5806 is fine to use? Strange that it is deprecated with no alternative.

[visr]

Let's revisit when a suitable dummy EPSG is added ;)