ruff/crates/ruff_python_formatter/src/lib.rs

use anyhow::Result;
use ruff_formatter::{format, Formatted, IndentStyle, SimpleFormatOptions};
use rustpython_parser::lexer::LexResult;

use crate::attachment::attach;
use crate::context::ASTFormatContext;
use crate::core::locator::Locator;
use crate::core::rustpython_helpers;
use crate::cst::Stmt;
use crate::newlines::normalize_newlines;
use crate::parentheses::normalize_parentheses;

mod attachment;
pub mod builders;
pub mod cli;
pub mod context;
mod core;
mod cst;
mod format;
mod newlines;
mod parentheses;
pub mod shared_traits;
pub mod trivia;

pub fn fmt(contents: &str) -> Result<Formatted<ASTFormatContext>> {
    // Tokenize once.
    let tokens: Vec<LexResult> = rustpython_helpers::tokenize(contents);

    // Extract trivia.
    let trivia = trivia::extract_trivia_tokens(&tokens);

    // Parse the AST.
    let python_ast = rustpython_helpers::parse_program_tokens(tokens, "<filename>")?;

    // Convert to a CST.
    let mut python_cst: Vec<Stmt> = python_ast.into_iter().map(Into::into).collect();

    // Attach trivia.
    attach(&mut python_cst, trivia);
    normalize_newlines(&mut python_cst);
    normalize_parentheses(&mut python_cst);

    format!(
        ASTFormatContext::new(
            SimpleFormatOptions {
                indent_style: IndentStyle::Space(4),
                line_width: 88.try_into().unwrap(),
            },
            Locator::new(contents)
        ),
        [format::builders::block(&python_cst)]
    )
    .map_err(Into::into)
}

#[cfg(test)]
mod tests {
    use std::fs;
    use std::path::Path;

    use anyhow::Result;

    use crate::fmt;
    use ruff_testing_macros::fixture;
    use similar::TextDiff;
    use std::fmt::{Formatter, Write};

    #[fixture(
        pattern = "resources/test/fixtures/black/**/*.py",
        // Excluded tests because they reach unreachable when attaching tokens
        exclude = [
            "*comments.py",
            "*comments[3,5,8].py",
            "*comments_non_breaking_space.py",
            "*docstring_preview.py",
            "*docstring.py",
            "*fmtonoff.py",
            "*fmtskip8.py",
        ])
    ]
    #[test]
    fn black_test(input_path: &Path) -> Result<()> {
        let content = fs::read_to_string(input_path)?;

        let formatted = fmt(&content)?;

        let expected_path = input_path.with_extension("py.expect");
        let expected_output = fs::read_to_string(&expected_path)
            .unwrap_or_else(|_| panic!("Expected Black output file '{expected_path:?}' to exist"));

        let printed = formatted.print()?;
        let formatted_code = printed.as_code();

        if formatted_code == expected_output {
            // Black and Ruff formatting matches. Delete any existing snapshot files because the Black output
            // already perfectly captures the expected output.
            // The following code mimics insta's logic generating the snapshot name for a test.
            let workspace_path = std::env::var("CARGO_MANIFEST_DIR").unwrap();
            let snapshot_name = insta::_function_name!()
                .strip_prefix(&format!("{}::", module_path!()))
                .unwrap();
            let module_path = module_path!().replace("::", "__");

            let snapshot_path = Path::new(&workspace_path)
                .join("src/snapshots")
                .join(format!(
                    "{module_path}__{}.snap",
                    snapshot_name.replace(&['/', '\\'][..], "__")
                ));

            if snapshot_path.exists() && snapshot_path.is_file() {
                // SAFETY: This is a convenience feature. That's why we don't want to abort
                // when deleting a no longer needed snapshot fails.
                fs::remove_file(&snapshot_path).ok();
            }

            let new_snapshot_path = snapshot_path.with_extension("snap.new");
            if new_snapshot_path.exists() && new_snapshot_path.is_file() {
                // SAFETY: This is a convenience feature. That's why we don't want to abort
                // when deleting a no longer needed snapshot fails.
                fs::remove_file(&new_snapshot_path).ok();
            }
        } else {
            // Black and Ruff have different formatting. Write out a snapshot that covers the differences
            // today.
            let mut snapshot = String::new();
            write!(snapshot, "{}", Header::new("Input"))?;
            write!(snapshot, "{}", CodeFrame::new("py", &content))?;

            write!(snapshot, "{}", Header::new("Black Differences"))?;

            let diff = TextDiff::from_lines(expected_output.as_str(), formatted_code)
                .unified_diff()
                .header("Black", "Ruff")
                .to_string();

            write!(snapshot, "{}", CodeFrame::new("diff", &diff))?;

            write!(snapshot, "{}", Header::new("Ruff Output"))?;
            write!(snapshot, "{}", CodeFrame::new("py", formatted_code))?;

            write!(snapshot, "{}", Header::new("Black Output"))?;
            write!(snapshot, "{}", CodeFrame::new("py", &expected_output))?;

            insta::with_settings!({ omit_expression => false, input_file => input_path }, {
                insta::assert_snapshot!(snapshot);
            });
        }

        Ok(())
    }

    /// Use this test to debug the formatting of some snipped
    #[ignore]
    #[test]
    fn quick_test() {
        let src = r#"
{
    k: v for k, v in a_very_long_variable_name_that_exceeds_the_line_length_by_far_keep_going
}
"#;
        let formatted = fmt(src).unwrap();

        // Uncomment the `dbg` to print the IR.
        // Use `dbg_write!(f, []) instead of `write!(f, [])` in your formatting code to print some IR
        // inside of a `Format` implementation
        // dbg!(formatted.document());

        let printed = formatted.print().unwrap();

        assert_eq!(
            printed.as_code(),
            r#"{
    k: v
    for k, v in a_very_long_variable_name_that_exceeds_the_line_length_by_far_keep_going
}"#
        );
    }

    struct Header<'a> {
        title: &'a str,
    }

    impl<'a> Header<'a> {
        fn new(title: &'a str) -> Self {
            Self { title }
        }
    }

    impl std::fmt::Display for Header<'_> {
        fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
            writeln!(f, "## {}", self.title)?;
            writeln!(f)
        }
    }

    struct CodeFrame<'a> {
        language: &'a str,
        code: &'a str,
    }

    impl<'a> CodeFrame<'a> {
        fn new(language: &'a str, code: &'a str) -> Self {
            Self { language, code }
        }
    }

    impl std::fmt::Display for CodeFrame<'_> {
        fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
            writeln!(f, "```{}", self.language)?;
            write!(f, "{}", self.code)?;
            writeln!(f, "```")?;
            writeln!(f)
        }
    }
}