Loading a case should automatically call extract_**_geometry

[mikekryjak]

The methods extract_1d_tokamak_geometry and extract_2d_tokamak_geometry contain some crucial bits, such as the unnormalisation of geometry variables and calculation of useful fields such as poloidal or radial cell lengths. They are not called automatically at the moment, which can cause confusion. Since Hermes-3 is pretty general purpose, this would need us to have a way to automatically identify if it's a 1D or 2D tokamak problem.

[dschwoerer]

Or a 3D FCI geometry :-)