Fix error in torsion_bounds function for elliptic curves

[DaveWitteMorris]

Fixes #39580. As pointed out in #39580, the torsion_bounds method of elliptic curves has a bug in its handling of the situation where the monic defining polynomial of the number field has nontrivial denominators in some of its coefficients.

Diagnosis. Assume E is an elliptic curve over a number field K, and let f be a monic defining polynomial of K, so a root a_ of f generates K. The code for torsion_bounds replaces f with a different polynomial (still called f, but obtained by rescaling the variable) that is monic and integral. The code erroneously assumes that a_ is a root of this new polynomial. To fix this, we can rescale a root of the new polynomial, to obtain a root of the original one.

📝 Checklist

• The title is concise and informative.
• The description explains in detail what this PR is about.
• I have linked a relevant issue or discussion.
• I have created tests covering the changes.
• I have updated the documentation and checked the documentation preview.

⌛ Dependencies

[github-actions]

Documentation preview for this PR (built with commit 67cab64; changes) is ready! 🎉
This preview will update shortly after each push to this PR.