Gmsh data for Alugrid and ug needs to be manually loadbalances. Sadly, ug
doesn't implement parameter load balancing for elements at the moment (only vertices).
It might be possbile to have all parameters on all processes.

Use always level 0 entities. This is for user convenience. The data was always
only available for level 0. Now the gridcreator also checks this and if not we take
the level 0 father instead to evaluate the parameter - more robust.

This test should also work in parallel now and a parallel test is added

We now get the parameter using the insertion index because that's the index
used by the factory in the gmsh reader. This also fixes how UG and ALU handle
reading gmsh files differently in parallel again.

Unfortunately the gmsh reader doesn't provide any load balancing features for the
parameters so that they might not always work in parallel

We now get the parameter using the insertion index because that's the index
used by the factory in the gmsh reader. This also fixes how UG and ALU handle
reading gmsh files differently in parallel again.

Unfortunately the gmsh reader doesn't provide any load balancing features for the
parameters so that they might not always work in parallel?

Fix/lowrekepsilon

See merge request !1031

Cleanup/component co2

Closes #489

See merge request !1034

[components][simpleco2] remove simpleco2 as it is not used anywhere and can be replaced by C02.hh when the correct table is used

Freeflow/kepsilon improvements

See merge request !1029

For wall functions, near-wall cells are always matchinPoints and should get a fixed value.

Freeflow/komegancni

See merge request !1028

Freeflow/komega

See merge request !1027

[math] Implement weighted means

See merge request !1024

[io] Use mapper size for determining length of output vectors

See merge request !1025

[poromechanics] pull fluid density out of volume variables

See merge request !1022