rs-module-lexer

ES module parser powered by Rust. Drop-in compatible with es-module-lexer.

Why

es-module-lexer only handles plain JS. Parsing ts / tsx / jsx requires a separate transform step first:

import { parse } from 'es-module-lexer'
import { transformSync } from 'esbuild'

// 1. transform ts to js
const js = transformSync(ts)
// 2. parse imports
const result = parse(js)

rs-module-lexer collapses that into a single call. It parses ts(x) / js(x) directly, powered by SWC.

Just Use It

Tools like Oxc and SWC already handle import parsing well, and sure, you could vibe one up. But even then, you still own the edge cases, spec drift, and the transform pipeline. This library takes care of all that, and it is compatible with the es-module-lexer data structure out of the box.

SWC is not as performance-optimized as Oxc, but import parsing is never the bottleneck in a real build pipeline. Good enough is good enough here.

Install

pnpm i -D rs-module-lexer

Usage

import { parseAsync } from 'rs-module-lexer'
// Sync: import { parse } from 'rs-module-lexer'

const { output } = await parseAsync({
  input: [
    {
      filename: 'index.ts',
      code: `
        export const member = 5
        import { useState } from 'react'
      `,
    },
    // ... other files
  ],
})

// [ { n: 'react', s: 67, e: 72, ss: 41, se: 73, d: -1, a: -1, t: 1 } ]
console.log(output[0].imports)
// [ { n: 'member', ln: 'member', s: 22, e: 28, ls: 22, le: 28 } ]
console.log(output[0].exports)

For the full result shape, see es-module-lexer.

The API accepts multiple files in one call. Syntax is auto-detected from the filename extension.

Migrate from es-module-lexer

- import { init, parse } from 'es-module-lexer'
- await init
- const [imports, exports, facade, hasModuleSyntax] = parse(code)

+ import { parseAsync } from 'rs-module-lexer'
+ const { output } = await parseAsync({ input: [{ filename: 'index.ts', code }] })
+ const { imports, exports, facade, hasModuleSyntax } = output[0]

License

MIT