Move away from re2/rea internal mesh zones

[vbaconnet]

This PR gets rid of the internal BCs from the re2/rea formats. The idea is that the "classic" labels w,v,o,on,sym,shl (and its variations) are transformed into labeled zones with indices 1...6.

For example, if the mesh contains boundaries with W, v and on, we will now have to specify "boundary_types": ["w", "v","","","on"].

Note that if one has several boundaries set internally as one of those re2 labels, they will all be marked in the same labeled zone. For example, if I have a channel with a lower and upper wall they will both be marked as labeled zone 1, and one will not be able to modify them separately (say, if I want to top wall to become an open boundary instead).

TODO before undrafting:

• [x] Do the same logic with rea files

NOTE: this has not been tested with meshes that have both internal labeled zones AND internal re2 labels like v,w,o etc.

[vbaconnet]

After some time, this PR is ready for review :) I have only tested with re2 and rea files that only have internal labels like v, w, etc. Not a combination of labeled zones and labels (as mentioned in the PR description).

But I think the logic should work!

[timofeymukha]

Victor, do you still map to zones 1...6 based on type, or is it just appending stuff as we discussed in the meeting?

[vbaconnet]

Victor, do you still map to zones 1...6 based on type, or is it just appending stuff as we discussed in the meeting?

Now it is appending stuff as we discussed!

[timofeymukha]

Appears to look good I think!

Note that if one has several boundaries set internally as one of those re2 labels, they will all be marked in the same labeled zone.

This is not a big problem I think. But generally, I think all the case logic based on the bc type is somewhat redundant now. You anyway do the same thing, besides for doing the addition to the zone number based on the type. It seems one could just throw away the case and always simply add a new zone. But this can be done later if needed.

Do we have a case with mixed internal/labled zones for Nek? To check if it works :-).

[MartinKarp]

I think this looks good! As for the output, when reading a re2/rea file if I understand correctly, at the first occurrence of a hardcoded BC it is printed to stdout which label it is assigned to?

[vbaconnet]

at the first occurrence of a hardcoded BC it is printed to stdout which label it is assigned to?

Yes exactly, it will assign the labeled zone based on order of appearance in the re2/rea file