gh-144884: Restructure re.sub docs, clarify aspects of repl notation

[cben]

This proposes a "maximalist" re-structuring of re.sub docs for #144884, putting ALL discussion of how-we-match parameters before what-we-do-with-matches.

👉 re.sub in preview

Moving flags with pattern made it easy to say it's only relevant for string patterns.

Consolidated repl notation from two far-away paragraphs to one place.

• Starting from \1 and \g which are the whole purpose of dealing with backslashes!
• Briefly mention \octal wart, 99 limit, and \g<100> avoiding them.
• Draw attention to \\ for getting a literal backslash.
• Clarify that most escapes are supported but \x\u\U\N aren't.
• Move "Unknown escapes of ASCII letters" after listing all the known ones.
• Added a note promoting raw string notation for repl too.

Further, I swapped order to explain callback before backslash notation because it's much shorter — but also to promote it. IMHO, people fear it as a "last-resort escape hatch" while it's actually conceptually simpler than backslashes (#128138 is one example). YMMV 😜

---

The downside is that backslash notation is covered at the end, after rarely important details like adjacent empty matches...

• I tried to mitigate that by a better example early on. Pulled out the "beans and spam" example which was out of place anyway (under "if repl is a function" but this repl is a string!), expanded to demonstrate \1 templating, and flags usage with both string & re.compile.

• A more drastic alternative could be to split sub & subn from "Functions" into a new "Search and replace" section, and explain the notation there at top level, even before the functions — like we do for RE syntax. That could work even better if gh-105636: Add re.Pattern.compile_template() #135992 lands.

• Issue: re.sub() repl notation explanation is disjointed #144884

---

📚 Documentation preview 📚: https://cpython-previews--144891.org.readthedocs.build/

[cben]

I feel we're using too many different verbs for same idea: "are replaced by", "[back]refer[ence]", "uses", "inserts", "converted to", "substitutes in"...
Any advice welcome!

• Technically, describing these as "backreference"s is inexact: backreferences in RE assert equality while matching — here we copy the captured text into replacement. And the syntaxes are somewhat different (see "Ways to reference it" table).
  However, I suspect readers do think of them as flip sides of same idea, and the doc uses "refer" in both senses widely...
  And for \20 inability to express \g<20> vs \g<2>0 distinction 👇 I had trouble phrasing it as well in other ways (perhaps because "reference" doubles as a noun).

• Dropping "substitutes" would be a shame because that's the one place we show what the function name stands for 😐 Ideally we'd use it in opening sentence, but that sounds clumsy in my head.