package
A fast, spec-compliant JavaScript and TypeScript parser, producing an ESTree / TypeScript-ESTree AST
npm install yuku-parserGitHub-flavored package documentation.
Across all versions.
Download trend unavailable.
Runtime metadata.
Tagged topics.
Estimated unpacked size across 3 packages.
108.6 kB
Includes dependencies.
A fast, spec-compliant JavaScript and TypeScript parser, part of Yuku.
npm install yuku-parserIt runs on Yuku's native core, installed for your platform. In browsers and edge runtimes, load the WebAssembly core from @yuku-core/wasm and pass it in as core.
import { parse } from "yuku-parser";
const { program, comments, diagnostics } = parse("const x = 1 + 2;");ESTree for JavaScript and JSX, identical to Acorn, and TypeScript-ESTree for TypeScript, matching Oxc for both. On top of the specs, it carries stage 3 decorators, and as on imports and , and the of .
Every node type is exported, listed in the type definitions.
import type { Expression, Identifier, Node, Statement } from "yuku-parser";To walk, build, and check nodes, use yuku-ast. For scopes and bindings, use yuku-analyzer.
parse(source, { path: "src/app.tsx" });| Option | Values | Default | Description |
|---|---|---|---|
path | a file path | none | Names the file in diagnostics. lang and sourceType default from its extension. |
lang | "js", "jsx", "ts", "tsx", "dts" | "js" | The syntax to parse. .d.ts, .tsx, .ts, and .jsx paths select their own. |
sourceType | "module", "script", "commonjs" | "module" | "commonjs" allows top-level return, and .cjs and .cts paths select it. |
preserveParens | true, false | true | Keep nodes. |
An unknown lang or sourceType throws a TypeError. langFromPath(path) and sourceTypeFromPath(path) return the values a path implies.
interface ParseResult {
program: Program;
comments: Comment[];
tokens?: TokenList; // with tokens: true
diagnostics: Diagnostic[];
}The parser recovers from errors, so a result with diagnostics still holds a tree of everything it could read.
Every Yuku package reports diagnostics in one shape.
interface Diagnostic {
severity: "error" | "warning" | "hint" | "info";
message: string;
path: string | null; // the path option
start: number; // UTF-16 offsets, like nodes
end: number;
labels: { start: number; end: number; message: string }[];
help: string | null;
}const { comments } = parse(`// a line comment\nconst x = 1; /* a block comment */`);
// [
// { type: "Line", value: " a line comment", start: 0, end: 17 },
// { type: "Block", value: " a block comment ", start: 31, end: 52 },
// ]attachComments: true also hangs each comment on the node it sits next to, which yuku-codegen prints from, so comments move with their nodes.
const { program } = parse(`// header\nfunction foo() {} // trailing`, { attachComments: true });
program.body[0].comments;
// [
// { type: "Line", position: "before", sameLine: false, value: " header" },
// { type: "Line", position: "after", sameLine: true, value: " trailing" },
// ]position is "before", "after", or "inside" an otherwise empty node, such as function f() { /* hi */ }.
tokens: true keeps every token in a TokenList, a view over the parser's token table where a token is an index.
import { parse, TokenKind } from "yuku-parser";
const { tokens } = parse(source, { tokens: true });
for (let i = 0; i < tokens.length; i++) {
if (tokens.kind(i) === TokenKind.Arrow) console.log(tokens.start(i), tokens.text(i));
}tokens.kind(i) // one of TokenKind
tokens.text(i) // its source text
tokens.start(i) // UTF-16 offsets
tokens.end(i)
tokens.isKeyword(i)
tokens.isReserved(i) // reserved unconditionally or in strict mode
tokens.isUnconditionallyReserved(i)
tokens.isStrictModeReserved(i)
tokens.isIdentifierLike(i) // an identifier or any keyword
tokens.isNumericLiteral(i)
tokens.isBinaryOperator(i)
tokens.isLogicalOperator(i)
tokens.isUnaryOperator(i)
tokens.isAssignmentOperator(i)
tokens.precedence
The queries answer with an index, or -1 when there is none. Tokens are as the parser resolved them, so a regex is one RegexLiteral, and comments are not tokens. The kinds are listed in tokens.d.ts.
MIT
phaseImportExpressionhashbangProgramParenthesizedExpressionsemanticErrors | true, false | false | Also report the early errors that need scopes, such as redeclarations and break outside a loop. |
attachComments | true, false | false | Also attach each comment to its node. See Comments. |
tokens | true, false | false | Keep every token. See Tokens. |
core | a loaded core | native | The core that parses, see @yuku-core/wasm. |