Add (and use) a default for IterEv

[scarlehoff]

I wanted to deal with this comment by @felixhekhorn #2028 (comment) and it ended up growing so much that I thought it would be better to separate it in its own PR so it can be reviewed before merging it to #2028.

This does very little:

1. Adds some conditional default (and tests for them)
2. Ensures that evolven3fit will make use of that default and, in addition, will loudly fail if someone tries to mix stuff.

[github-actions]

Greetings from your nice fit 🤖 !
I have good news for you, I just finished my tasks:

• Fit Name: NNBOT-962f2b4e7-2024-04-24
• Fit Report wrt master: https://vp.nnpdf.science/HwKLUMOPR-CBMUX-V_87og==
• Fit Report wrt latest stable reference: https://vp.nnpdf.science/3kh1ZGQdQeaYXClzPAJZ5Q==
• Fit Data: https://data.nnpdf.science/fits/NNBOT-962f2b4e7-2024-04-24.tar.gz

Check the report carefully, and please buy me a ☕ , or better, a GPU 😉!

[scarlehoff]

@niclaurenti @giacomomagni please, have a look at my solution to avoid the need for find_nf in evolven3fit.
For the photon we are taking Q0 as the value the photon needs to be evolved to and so it makes sense (I believe) to take nf0 as well as the target nf (while qedref is an input argument).

I don't think compute_eko_photon is tested anywhere though so please look closely.

I've put find_nf as a method although the theory card is mostly accessed as a dictionary at the moment but if (eventually) the Theory class is passed around this can be used to find nf directly.