Add exceptional classes and indices

[emikelsons]

Added the attributes exceptional_classes and exceptional_divisor_indices for FTheory models for more convenient access to this information. I also made slight improvements to the internal _set_all_attributes method.
Before this is merged, I would also like to add changes, that would change the exceptional_classes attribute, whenever a blow up is performed. @HereAround @apturner

[HereAround]

@apturner if the tests are correct, these changes here somehow break what we stated in our book chapter. But I fail to see exactly what. Any idea?

(@emikelsons Did you already rebase to master. Just want to make sure that the Hecke documentation error is gone and not responsible for any of this.)

Generally speaking, this looks good to me.

[emikelsons]

@apturner if the tests are correct, these changes here somehow break what we stated in our book chapter. But I fail to see exactly what. Any idea?

(@emikelsons Did you already rebase to master. Just want to make sure that the Hecke documentation error is gone and not responsible for any of this.)

Generally speaking, this looks good to me.

I did rebase, but if the errors that persist were indeed caused by me - then I'm not sure. I added additional optional arguments to the blow_up method, but I'm not sure, if this could cause the book tests to fail

[benlorenz]

Regarding the book tests, one ambient space prints slightly different now:

  julia> t1 = blow_up(t, [\"x\", \"y\", \"x1\"]; coordinate_name = \"e1\")
  Partially resolved global Tate model over a concrete base
  
  julia> amb1 = ambient_space(t1)
- Normal toric variety
+ Normal, simplicial toric variety

(https://github.com/oscar-system/Oscar.jl/actions/runs/14264819453/job/39984171773?pr=4779#step:7:921)

Regarding the other tests: This looks like the timeout / IO issues that we have seen quite a bit during this week, this hopefully disappears when re-run when the CI runner is less busy.

[benlorenz]

It would also be fine to adjust the book-test output to new printing, we never claimed that this would stay unchanged. (assuming the new output is also correct)
Only the input lines should stay unchanged in 1.x

[emikelsons]

It would also be fine to adjust the book-test output to new printing, we never claimed that this would stay unchanged. (assuming the new output is also correct) Only the input lines should stay unchanged in 1.x

I see, I wasn't familiar with the book test claims and restrictions. This certainly makes things easier as my current fix isn't great

[HereAround]

Thank you @emikelsons for working on this!