Skip to content

README.md

Latest commit

 

History

History
67 lines (46 loc) · 2.45 KB

README.md

File metadata and controls

67 lines (46 loc) · 2.45 KB

debarrel

Removes barrel files (index re-export files) from JavaScript and TypeScript codebases, rewriting imports to point directly at the source modules.

Before / After

// Before — importing through a barrel
import { Button } from "./components";

// After — importing directly from the source
import { Button } from "./components/Button";

Pure barrel files (files that only re-export) are renamed to index.barrel.bak.ts for review.

What it handles

  • Named re-exports (export { X } from './X')
  • Aliased re-exports (export { X as Y } from './X')
  • Default-as-named re-exports (export { default as X } from './X')
  • Namespace re-exports (export * as X from './X')
  • Import-then-reexport patterns
  • Mixed barrel files that contain their own declarations (only the re-exported imports are rewritten; the barrel is kept)

Migration complexity metric

Alongside the rewrite, the codemod emits a barrel-metric event per re-export found in each barrel. Each event carries five cardinalities so you can estimate migration effort before rolling the codemod out at scale:

Cardinality Values
export_style explicit | wildcard
chaining single-level | chained
risk_amplifier none | heavy-usage-or-public-api | cycles-or-side-effects
file workspace-relative path of the affected barrel
migration_complexity_score 0–4 (sum of the other dimensions)

Scoring rules: wildcard adds 1, chained adds 1, heavy-usage-or-public-api adds 1, cycles-or-side-effects adds 2. Roughly: 0–1 = Low, 2 = Medium, 3–4 = High.

Notes:

  • heavy-usage-or-public-api fires only when the barrel matches the advertised entry point of its nearest package.json (main, module, types, or any leaf of exports). A random internal index.ts inside a repo is not flagged.
  • count aggregates raw emission events, not unique re-export sites. Because every file in a barrel's directory emits the same metric tuple (a workaround for per-directory test snapshots), the value is inflated by directory density in proportion to the number of sibling files. To recover an approximate re-export count, divide count by the number of files in the barrel's directory.

Usage

# Run from the registry
codemod run debarrel

# Run locally during development
codemod run -w workflow.yaml

Development

# Run tests
pnpm test

# Type-check
pnpm check-types

License

MIT