[docs] improve and fix entry for analyze.include-dependencies

[purajit]

Summary

Changes two things about the entry:

• make the example valid TOML - inline tables must be a single line, at least till v1.1.0 is released,
  but also while in the future the toml version used by ruff might handle it, it would probably be
  good to stick to a spec that's readable by the vast majority of other tools and versions as well,
  especially if people are using pyproject.toml. The current example leads to ruff failure.
  See Allow newlines and trailing comma in inline tables toml-lang/toml#904
• adds a line about the ability to add non-Python files to the map, which I think is a specific and
  important feature people should know about (in fact, I would assume this could potentially
  become the single biggest use-case for this).

Test Plan

Ran doc creation as described in the contribution guide.

[purajit]

new entry

[github-actions]

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

[purajit]

For the second point, I was trying to see if that was intentional or not; couldn't find anything
in the original PR, and the code

ruff/crates/ruff/src/commands/analyze_graph.rs

Lines 156 to 159 in 8a98d88

	if let Some((root, globs)) = include_dependencies {
	let mut glob_resolver = glob_resolver.lock().unwrap();
	imports.extend(glob_resolver.resolve(root, globs));
	}

doesn't suggest anything strongly either; GlobResolver resolves all files. My main hesitation
is that it's consistently called an "import map", and there are already issues/discussions about
expanding the scope of analyze graph assuming that these are only Python deps.. I hope it
stays, though I understand if it won't.

[charliermarsh]

Thanks! It's an intentional behavior.