Indicate code snippets whose output need not match the LMFDB

[rvisser7]

Many LMFDB pages have code snippets in various languages (e.g. Sage/Magma/Pari) which recomputes the data displayed on the page. For most of the data, the code snippets give output that matches exactly what the page shows. Though there are some invariants where this is not the case. E.g. for number fields, the Pari code snippet for giving a fundamental system of units is K.fu. While this gives correct output, it often doesn't match exactly the generators given on the LMFDB.

This is of course understandable (and unavoidable), though I've been receiving feedback from a few people, especially those new to the LMFDB, that this can sometimes be a bit confusing. E.g. it's perfectly reasonable to (wrongly) assume that the LMFDB data has been generated directly from the code snippets provided, and then get confused when the output from Pari's K.fu is not exactly the same as what the LMFDB gives.

For such code snippets, I think it might be helpful to include a short comment saying something like "Output not guaranteed to match LMFDB" (or perhaps something even shorter). In fact, this is already done for the character table code snippets on the abstract groups pages.

Here's a (non-exhaustive) list of examples of such code snippets I've found so far:

• Number fields: Integral basis, fundamental units.
• Elliptic curves: Mordell-Weil generators, isogeny class matrix.
• Genus 2 curves: Mordell-Weil generators.
• Abstract groups: Chief series, character table (comment already given).
• Galois groups: Character table, conjugacy class representatives (and maybe group generators?).