ruff_python_parser/

lib.rs

//! This crate can be used to parse Python source code into an Abstract
//! Syntax Tree.
//!
//! ## Overview
//!
//! The process by which source code is parsed into an AST can be broken down
//! into two general stages: [lexical analysis] and [parsing].
//!
//! During lexical analysis, the source code is converted into a stream of lexical
//! tokens that represent the smallest meaningful units of the language. For example,
//! the source code `print("Hello world")` would _roughly_ be converted into the following
//! stream of tokens:
//!
//! ```text
//! Name("print"), LeftParen, String("Hello world"), RightParen
//! ```
//!
//! These tokens are then consumed by the `ruff_python_parser`, which matches them against a set of
//! grammar rules to verify that the source code is syntactically valid and to construct
//! an AST that represents the source code.
//!
//! During parsing, the `ruff_python_parser` consumes the tokens generated by the lexer and constructs
//! a tree representation of the source code. The tree is made up of nodes that represent
//! the different syntactic constructs of the language. If the source code is syntactically
//! invalid, parsing fails and an error is returned. After a successful parse, the AST can
//! be used to perform further analysis on the source code. Continuing with the example
//! above, the AST generated by the `ruff_python_parser` would _roughly_ look something like this:
//!
//! ```text
//! node: Expr {
//!     value: {
//!         node: Call {
//!             func: {
//!                 node: Name {
//!                     id: "print",
//!                     ctx: Load,
//!                 },
//!             },
//!             args: [
//!                 node: Constant {
//!                     value: Str("Hello World"),
//!                     kind: None,
//!                 },
//!             ],
//!             keywords: [],
//!         },
//!     },
//! },
//!```
//!
//! **Note:** The Tokens/ASTs shown above are not the exact tokens/ASTs generated by the `ruff_python_parser`.
//! Refer to the [playground](https://play.ruff.rs) for the correct representation.
//!
//! ## Source code layout
//!
//! The functionality of this crate is split into several modules:
//!
//! - [lexer]: This module contains the lexer and is responsible for generating the tokens.
//! - parser: This module contains an interface to the [Parsed] and is responsible for generating the AST.
//! - mode: This module contains the definition of the different modes that the `ruff_python_parser` can be in.
//!
//! [lexical analysis]: https://en.wikipedia.org/wiki/Lexical_analysis
//! [parsing]: https://en.wikipedia.org/wiki/Parsing
//! [lexer]: crate::lexer

pub use crate::error::{
    InterpolatedStringErrorType, LexicalErrorType, ParseError, ParseErrorType,
    UnsupportedSyntaxError, UnsupportedSyntaxErrorKind,
};
pub use crate::parser::ParseOptions;

use crate::parser::Parser;

use ruff_python_ast::token::{Token, TokenFlags, TokenKind, Tokens};
use ruff_python_ast::{
    AtomicNodeIndex, Expr, Mod, ModExpression, ModModule, PySourceType, StringFlags, StringLiteral,
    Suite,
};
use ruff_text_size::{Ranged, TextRange};

mod error;
pub mod lexer;
mod parser;
pub mod semantic_errors;
mod string;
mod token_set;
mod token_source;
pub mod typing;

/// Parse a full Python module usually consisting of multiple lines.
///
/// This is a convenience function that can be used to parse a full Python program without having to
/// specify the [`Mode`] or the location. It is probably what you want to use most of the time.
///
/// # Example
///
/// For example, parsing a simple function definition and a call to that function:
///
/// ```
/// use ruff_python_parser::parse_module;
///
/// let source = r#"
/// def foo():
///    return 42
///
/// print(foo())
/// "#;
///
/// let module = parse_module(source);
/// assert!(module.is_ok());
/// ```
pub fn parse_module(source: &str) -> Result<Parsed<ModModule>, ParseError> {
    Parser::new(source, ParseOptions::from(Mode::Module))
        .parse()
        .try_into_module()
        .unwrap()
        .into_result()
}

/// Parses a single Python expression.
///
/// This convenience function can be used to parse a single expression without having to
/// specify the Mode or the location.
///
/// # Example
///
/// For example, parsing a single expression denoting the addition of two numbers:
///
/// ```
/// use ruff_python_parser::parse_expression;
///
/// let expr = parse_expression("1 + 2");
/// assert!(expr.is_ok());
/// ```
pub fn parse_expression(source: &str) -> Result<Parsed<ModExpression>, ParseError> {
    Parser::new(source, ParseOptions::from(Mode::Expression))
        .parse()
        .try_into_expression()
        .unwrap()
        .into_result()
}

/// Parses a Python expression for the given range in the source.
///
/// This function allows to specify the range of the expression in the source code, other than
/// that, it behaves exactly like [`parse_expression`].
///
/// # Example
///
/// Parsing one of the numeric literal which is part of an addition expression:
///
/// ```
/// use ruff_python_parser::parse_expression_range;
/// # use ruff_text_size::{TextRange, TextSize};
///
/// let parsed = parse_expression_range("11 + 22 + 33", TextRange::new(TextSize::new(5), TextSize::new(7)));
/// assert!(parsed.is_ok());
/// ```
pub fn parse_expression_range(
    source: &str,
    range: TextRange,
) -> Result<Parsed<ModExpression>, ParseError> {
    let source = &source[..range.end().to_usize()];
    Parser::new_starts_at(source, range.start(), ParseOptions::from(Mode::Expression))
        .parse()
        .try_into_expression()
        .unwrap()
        .into_result()
}

/// Parses a Python expression as if it is parenthesized.
///
/// It behaves similarly to [`parse_expression_range`] but allows what would be valid within parenthesis
///
/// # Example
///
/// Parsing an expression that would be valid within parenthesis:
///
/// ```
/// use ruff_python_parser::parse_parenthesized_expression_range;
/// # use ruff_text_size::{TextRange, TextSize};
///
/// let parsed = parse_parenthesized_expression_range("'''\n int | str'''", TextRange::new(TextSize::new(3), TextSize::new(14)));
/// assert!(parsed.is_ok());
pub fn parse_parenthesized_expression_range(
    source: &str,
    range: TextRange,
) -> Result<Parsed<ModExpression>, ParseError> {
    let source = &source[..range.end().to_usize()];
    let parsed = Parser::new_starts_at(
        source,
        range.start(),
        ParseOptions::from(Mode::ParenthesizedExpression),
    )
    .parse();
    parsed.try_into_expression().unwrap().into_result()
}

/// Parses a Python expression from a string annotation.
///
/// # Example
///
/// Parsing a string annotation:
///
/// ```
/// use ruff_python_parser::parse_string_annotation;
/// use ruff_python_ast::{StringLiteral, StringLiteralFlags, AtomicNodeIndex};
/// use ruff_text_size::{TextRange, TextSize};
///
/// let string = StringLiteral {
///     value: "'''\n int | str'''".to_string().into_boxed_str(),
///     flags: StringLiteralFlags::empty(),
///     range: TextRange::new(TextSize::new(0), TextSize::new(16)),
///     node_index: AtomicNodeIndex::NONE
/// };
/// let parsed = parse_string_annotation("'''\n int | str'''", &string);
/// assert!(!parsed.is_ok());
/// ```
pub fn parse_string_annotation(
    source: &str,
    string: &StringLiteral,
) -> Result<Parsed<ModExpression>, ParseError> {
    let range = string
        .range()
        .add_start(string.flags.opener_len())
        .sub_end(string.flags.closer_len());
    let source = &source[..range.end().to_usize()];
    if string.flags.is_triple_quoted() {
        parse_parenthesized_expression_range(source, range)
    } else {
        parse_expression_range(source, range)
    }
}

/// Parse the given Python source code using the specified [`ParseOptions`].
///
/// This function is the most general function to parse Python code. Based on the [`Mode`] supplied
/// via the [`ParseOptions`], it can be used to parse a single expression, a full Python program,
/// an interactive expression or a Python program containing IPython escape commands.
///
/// # Example
///
/// If we want to parse a simple expression, we can use the [`Mode::Expression`] mode during
/// parsing:
///
/// ```
/// use ruff_python_parser::{parse, Mode, ParseOptions};
///
/// let parsed = parse("1 + 2", ParseOptions::from(Mode::Expression));
/// assert!(parsed.is_ok());
/// ```
///
/// Alternatively, we can parse a full Python program consisting of multiple lines:
///
/// ```
/// use ruff_python_parser::{parse, Mode, ParseOptions};
///
/// let source = r#"
/// class Greeter:
///
///   def greet(self):
///    print("Hello, world!")
/// "#;
/// let parsed = parse(source, ParseOptions::from(Mode::Module));
/// assert!(parsed.is_ok());
/// ```
///
/// Additionally, we can parse a Python program containing IPython escapes:
///
/// ```
/// use ruff_python_parser::{parse, Mode, ParseOptions};
///
/// let source = r#"
/// %timeit 1 + 2
/// ?str.replace
/// !ls
/// "#;
/// let parsed = parse(source, ParseOptions::from(Mode::Ipython));
/// assert!(parsed.is_ok());
/// ```
pub fn parse(source: &str, options: ParseOptions) -> Result<Parsed<Mod>, ParseError> {
    parse_unchecked(source, options).into_result()
}

/// Parse the given Python source code using the specified [`ParseOptions`].
///
/// This is same as the [`parse`] function except that it doesn't check for any [`ParseError`]
/// and returns the [`Parsed`] as is.
pub fn parse_unchecked(source: &str, options: ParseOptions) -> Parsed<Mod> {
    Parser::new(source, options).parse()
}

/// Parse the given Python source code using the specified [`PySourceType`].
pub fn parse_unchecked_source(source: &str, source_type: PySourceType) -> Parsed<ModModule> {
    // SAFETY: Safe because `PySourceType` always parses to a `ModModule`
    Parser::new(source, ParseOptions::from(source_type))
        .parse()
        .try_into_module()
        .unwrap()
}

/// Parses each `range` of `source` as an independent module and concatenates the results into a
/// single [`Parsed<ModModule>`] whose nodes keep their offsets into `source`.
///
/// This validates sources such as Jupyter notebooks, where each cell must be syntactically valid on
/// its own while later cells can still reference earlier definitions.
/// The `ranges` must be ordered and non-overlapping.
///
/// Consecutive ranges must be separated by a single-byte `\n`: each range ends just before the
/// separator and the next range starts just after it, as Ruff's notebook cells are. A syntax error
/// anchored at a cell's trailing offset then lands on that separator, the cell's own last line, so
/// it is attributed to that cell rather than to the following one.
pub fn parse_cells_unchecked(
    source: &str,
    ranges: impl IntoIterator<Item = TextRange>,
    options: &ParseOptions,
) -> Parsed<ModModule> {
    let mut ranges = ranges.into_iter().peekable();
    let mut body = Suite::new();
    let mut tokens = Vec::new();
    let mut errors = Vec::new();
    let mut unsupported_syntax_errors = Vec::new();
    let mut module_range: Option<TextRange> = None;

    while let Some(range) = ranges.next() {
        if let Some(previous) = module_range {
            assert!(previous.end() <= range.start());
        }

        // The cell is lexed from `range.start()`, so the slice must keep the leading text to
        // preserve absolute offsets into the concatenated source.
        let cell_source = &source[TextRange::up_to(range.end())];
        let Parsed {
            syntax,
            tokens: cell_tokens,
            errors: cell_errors,
            unsupported_syntax_errors: cell_unsupported_syntax_errors,
        } = Parser::new_starts_at(cell_source, range.start(), options.clone())
            .parse()
            .try_into_module()
            .expect("module options should parse into a module");

        body.extend(syntax.body);
        tokens.extend(cell_tokens);
        errors.extend(cell_errors);
        unsupported_syntax_errors.extend(cell_unsupported_syntax_errors);

        // Each range excludes its trailing `\n` separator (see the doc comment above), leaving a
        // one-byte gap in the token stream. Cover it with a `NonLogicalNewline` so token-based
        // checks don't treat the separator as another logical line terminator. The final cell's
        // separator is the file-final newline and is deliberately left uncovered.
        if let Some(next) = ranges.peek() {
            let separator = TextRange::new(range.end(), next.start());
            assert_eq!(&source[separator], "\n");
            tokens.push(Token::new(
                TokenKind::NonLogicalNewline,
                separator,
                TokenFlags::empty(),
            ));
        }

        module_range = Some(match module_range {
            Some(previous) => TextRange::new(previous.start(), range.end()),
            None => range,
        });
    }

    body.shrink_to_fit();
    tokens.shrink_to_fit();
    errors.shrink_to_fit();
    unsupported_syntax_errors.shrink_to_fit();

    Parsed {
        syntax: ModModule {
            node_index: AtomicNodeIndex::NONE,
            range: module_range.unwrap_or_default(),
            body,
        },
        tokens: Tokens::new(tokens),
        errors,
        unsupported_syntax_errors,
    }
}

/// Represents the parsed source code.
#[derive(Debug, PartialEq, Clone, get_size2::GetSize)]
pub struct Parsed<T> {
    syntax: T,
    tokens: Tokens,
    errors: Vec<ParseError>,
    unsupported_syntax_errors: Vec<UnsupportedSyntaxError>,
}

impl<T> Parsed<T> {
    /// Returns the syntax node represented by this parsed output.
    pub fn syntax(&self) -> &T {
        &self.syntax
    }

    /// Returns all the tokens for the parsed output.
    pub fn tokens(&self) -> &Tokens {
        &self.tokens
    }

    /// Returns a list of syntax errors found during parsing.
    pub fn errors(&self) -> &[ParseError] {
        &self.errors
    }

    /// Returns a list of version-related syntax errors found during parsing.
    pub fn unsupported_syntax_errors(&self) -> &[UnsupportedSyntaxError] {
        &self.unsupported_syntax_errors
    }

    /// Consumes the [`Parsed`] output and returns the contained syntax node.
    pub fn into_syntax(self) -> T {
        self.syntax
    }

    /// Consumes the [`Parsed`] output and returns a list of syntax errors found during parsing.
    fn into_errors(self) -> Vec<ParseError> {
        self.errors
    }

    /// Returns `true` if the parsed source code is valid i.e., it has no [`ParseError`]s.
    ///
    /// Note that this does not include version-related [`UnsupportedSyntaxError`]s.
    ///
    /// See [`Parsed::has_no_syntax_errors`] for a version that takes these into account.
    pub fn has_valid_syntax(&self) -> bool {
        self.errors.is_empty()
    }

    /// Returns `true` if the parsed source code is invalid i.e., it has [`ParseError`]s.
    ///
    /// Note that this does not include version-related [`UnsupportedSyntaxError`]s.
    ///
    /// See [`Parsed::has_no_syntax_errors`] for a version that takes these into account.
    pub fn has_invalid_syntax(&self) -> bool {
        !self.has_valid_syntax()
    }

    /// Returns `true` if the parsed source code does not contain any [`ParseError`]s *or*
    /// [`UnsupportedSyntaxError`]s.
    ///
    /// See [`Parsed::has_valid_syntax`] for a version specific to [`ParseError`]s.
    pub fn has_no_syntax_errors(&self) -> bool {
        self.has_valid_syntax() && self.unsupported_syntax_errors.is_empty()
    }

    /// Returns `true` if the parsed source code contains any [`ParseError`]s *or*
    /// [`UnsupportedSyntaxError`]s.
    ///
    /// See [`Parsed::has_invalid_syntax`] for a version specific to [`ParseError`]s.
    pub fn has_syntax_errors(&self) -> bool {
        !self.has_no_syntax_errors()
    }

    /// Returns the [`Parsed`] output as a [`Result`], returning [`Ok`] if it has no syntax errors,
    /// or [`Err`] containing the first [`ParseError`] encountered.
    ///
    /// Note that any [`unsupported_syntax_errors`](Parsed::unsupported_syntax_errors) will not
    /// cause [`Err`] to be returned.
    pub fn as_result(&self) -> Result<&Parsed<T>, &[ParseError]> {
        if self.has_valid_syntax() {
            Ok(self)
        } else {
            Err(&self.errors)
        }
    }

    /// Consumes the [`Parsed`] output and returns a [`Result`] which is [`Ok`] if it has no syntax
    /// errors, or [`Err`] containing the first [`ParseError`] encountered.
    ///
    /// Note that any [`unsupported_syntax_errors`](Parsed::unsupported_syntax_errors) will not
    /// cause [`Err`] to be returned.
    fn into_result(self) -> Result<Parsed<T>, ParseError> {
        if self.has_valid_syntax() {
            Ok(self)
        } else {
            Err(self.into_errors().into_iter().next().unwrap())
        }
    }
}

impl Parsed<Mod> {
    /// Attempts to convert the [`Parsed<Mod>`] into a [`Parsed<ModModule>`].
    ///
    /// This method checks if the `syntax` field of the output is a [`Mod::Module`]. If it is, the
    /// method returns [`Some(Parsed<ModModule>)`] with the contained module. Otherwise, it
    /// returns [`None`].
    ///
    /// [`Some(Parsed<ModModule>)`]: Some
    pub fn try_into_module(self) -> Option<Parsed<ModModule>> {
        match self.syntax {
            Mod::Module(module) => Some(Parsed {
                syntax: module,
                tokens: self.tokens,
                errors: self.errors,
                unsupported_syntax_errors: self.unsupported_syntax_errors,
            }),
            Mod::Expression(_) => None,
        }
    }

    /// Attempts to convert the [`Parsed<Mod>`] into a [`Parsed<ModExpression>`].
    ///
    /// This method checks if the `syntax` field of the output is a [`Mod::Expression`]. If it is,
    /// the method returns [`Some(Parsed<ModExpression>)`] with the contained expression.
    /// Otherwise, it returns [`None`].
    ///
    /// [`Some(Parsed<ModExpression>)`]: Some
    fn try_into_expression(self) -> Option<Parsed<ModExpression>> {
        match self.syntax {
            Mod::Module(_) => None,
            Mod::Expression(expression) => Some(Parsed {
                syntax: expression,
                tokens: self.tokens,
                errors: self.errors,
                unsupported_syntax_errors: self.unsupported_syntax_errors,
            }),
        }
    }
}

impl Parsed<ModModule> {
    /// Returns the module body contained in this parsed output as a [`Suite`].
    pub fn suite(&self) -> &Suite {
        &self.syntax.body
    }

    /// Consumes the [`Parsed`] output and returns the module body as a [`Suite`].
    pub fn into_suite(self) -> Suite {
        self.syntax.body
    }
}

impl Parsed<ModExpression> {
    /// Returns the expression contained in this parsed output.
    pub fn expr(&self) -> &Expr {
        &self.syntax.body
    }

    /// Returns a mutable reference to the expression contained in this parsed output.
    fn expr_mut(&mut self) -> &mut Expr {
        &mut self.syntax.body
    }

    /// Consumes the [`Parsed`] output and returns the contained [`Expr`].
    pub fn into_expr(self) -> Expr {
        *self.syntax.body
    }
}

/// Control in the different modes by which a source file can be parsed.
///
/// The mode argument specifies in what way code must be parsed.
#[derive(Clone, Copy, Debug, Hash, PartialEq, Eq)]
pub enum Mode {
    /// The code consists of a sequence of statements.
    Module,

    /// The code consists of a single expression.
    Expression,

    /// The code consists of a single expression and is parsed as if it is parenthesized. The parentheses themselves aren't required.
    /// This allows for having valid multiline expression without the need of parentheses
    /// and is specifically useful for parsing string annotations.
    ParenthesizedExpression,

    /// The code consists of a sequence of statements which can include the
    /// escape commands that are part of IPython syntax.
    ///
    /// ## Supported escape commands:
    ///
    /// - [Magic command system] which is limited to [line magics] and can start
    ///   with `?` or `??`.
    /// - [Dynamic object information] which can start with `?` or `??`.
    /// - [System shell access] which can start with `!` or `!!`.
    /// - [Automatic parentheses and quotes] which can start with `/`, `;`, or `,`.
    ///
    /// [Magic command system]: https://ipython.readthedocs.io/en/stable/interactive/reference.html#magic-command-system
    /// [line magics]: https://ipython.readthedocs.io/en/stable/interactive/magics.html#line-magics
    /// [Dynamic object information]: https://ipython.readthedocs.io/en/stable/interactive/reference.html#dynamic-object-information
    /// [System shell access]: https://ipython.readthedocs.io/en/stable/interactive/reference.html#system-shell-access
    /// [Automatic parentheses and quotes]: https://ipython.readthedocs.io/en/stable/interactive/reference.html#automatic-parentheses-and-quotes
    Ipython,
}

impl std::str::FromStr for Mode {
    type Err = ModeParseError;
    fn from_str(s: &str) -> Result<Self, ModeParseError> {
        match s {
            "exec" | "single" => Ok(Mode::Module),
            "eval" => Ok(Mode::Expression),
            "ipython" => Ok(Mode::Ipython),
            _ => Err(ModeParseError),
        }
    }
}

/// A type that can be represented as [Mode].
pub trait AsMode {
    fn as_mode(&self) -> Mode;
}

impl AsMode for PySourceType {
    fn as_mode(&self) -> Mode {
        match self {
            PySourceType::Python | PySourceType::Stub => Mode::Module,
            PySourceType::Ipynb => Mode::Ipython,
        }
    }
}

/// Returned when a given mode is not valid.
#[derive(Debug)]
pub struct ModeParseError;

impl std::fmt::Display for ModeParseError {
    fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
        write!(f, r#"mode must be "exec", "eval", "ipython", or "single""#)
    }
}