[ruff]: Make ruff analyze graph work with jupyter notebooks

[gauthsvenkat]

Summary

ruff analyze graph skips jupyter notebooks. This MR addresses that issue so that jupyter notebooks are also considered in the output.

Closes: #21099

Detailed Summary

ruff analyze graph has the following lines to skip if the source is a jupyter notebook.

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

Lines 130 to 133 in d9cab4d

	if matches!(source_type, PySourceType::Ipynb) {
	debug!("Ignoring Jupyter notebook: {}", path.display());
	continue;
	}

Removing these lines alone is not enough (see this comment on the issue). The reason is because in detect() the file under consideration is read and parsed to be a .py file.

ruff/crates/ruff_graph/src/lib.rs

Lines 25 to 33 in d9cab4d

	pub fn detect(
	db: &ModuleDb,
	path: &SystemPath,
	package: Option<&SystemPath>,
	string_imports: StringImports,
	) -> Result<Self> {
	// Read and parse the source code.
	let source = std::fs::read_to_string(path)?;
	let parsed = parse(&source, ParseOptions::from(Mode::Module))?;

For notebooks, you'd expect this parsing to fail as a .ipynb and .py are different. However that is not the case as under the hood, contents of a jupyter notebook are JSON. What instead is happening is that the jupyter notebooks get parsed as
a giant nested python dictionary.

Fix

• detect() now takes two more arguments; source: the raw content and source_type to correctly parse both .py and .ipynb content.
• in analyze_graph.rs we extract the raw content correctly for both .py and .ipynb.

Test Plan

All existing unit tests pass and added two new tests to make sure the jupyter notebook flow also works.

[github-actions]

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

[MichaReiser]

This is great. Thank you