pack-lambda

Great .zip packager and S3 uploader for AWS Lambda. Bundles only your prod dependencies that are already installed, no need to prune or copy to a temp directory and do a fresh install. And it even works with dependencies installed by pnpm!

Supports:

• Scoped packages
• Same filename as npm pack, except for extension
• prepack package script
• --dry-run
• --pack-destination

Doesn't currently support:

• Packing other packages in args
• postpack package script
• --json
• --workspace
• --workspaces

Filenames

The default filename is [name]-[version].zip where [name] is the package name (scoped names like @foo/bar get converted to foo-bar) and [version] is the package version. If the version ends in -development, a timestamp will be appened like [name]-[version]-[timestamp].zip so that the filename is always different and a CloudFormation stack update will always update the Lambda code.

Configuration

You can add the following config to package.json:

{
  "@jcoreio/pack-lambda": {
    /**
     * Excludes dependencies from getting bundled
     * (aws-sdk is pre-installed in the Lambda runtime so you could do this to save space)
     * This has no effect if your package.json contains bundledDependencies
     */
    "excludeDependencies": ["aws-sdk"]
  }
}

CLI

pack-lambda - create a .zip file

npx --package @jcoreio/pack-lambda pack-lambda

pack .zip file for AWS Lambda

Options:
  --version                       Show version number                  [boolean]
  --help                          Show help                            [boolean]
  --dry-run                       display contents without writing file
                                                      [boolean] [default: false]
  --pack-destination              directory in which to save .zip files [string]

Example Output

> @jcoreio/[email protected] prepack

<... prepack script output ...>

📦  @jcoreio/[email protected]
=== Zip Contents ===
es/index.js.flow
index.js.flow
es/types/npmcli__run-script/index.d.js
types/npmcli__run-script/index.d.js
bin/index.js
es/bin/index.js
es/index.js
index.js
package.json
LICENSE.md
README.md
bin/index.d.ts
es/bin/index.d.ts
es/index.d.ts
index.d.ts
=== Bundled Dependencies ===
@aws-sdk/client-s3
@aws-sdk/lib-storage
@babel/runtime
@npmcli/run-script
archiver
chalk
fs-extra
npm-package-arg
npm-packlist
yargs
@aws-sdk/types
=== Zip Details ===
name:          @jcoreio/pack-lambda
version:       0.0.0-development
filename:      jcoreio-pack-lambda-0.0.0-development-20211205184649022.zip
bundled deps:  11
total files:   922
jcoreio-pack-lambda-0.0.0-development-20211205184649022.zip

upload - upload to S3

npx --package @jcoreio/pack-lambda pack-lambda upload <bucket> [key]

upload .zip to S3

Positionals:
  bucket  S3 Bucket[/Key]                                             [required]
  key     S3 Key

Options:
  --version                       Show version number                  [boolean]
  --help                          Show help                            [boolean]
  --hash     compute hash and skip if already uploaded                 [boolean]

Example Output

> @jcoreio/[email protected] prepack

<... prepack script output ...>

📦  @jcoreio/[email protected]
=== Zip Contents ===
es/index.js.flow
index.js.flow
es/types/npmcli__run-script/index.d.js
types/npmcli__run-script/index.d.js
bin/index.js
es/bin/index.js
es/index.js
index.js
package.json
LICENSE.md
README.md
bin/index.d.ts
es/bin/index.d.ts
es/index.d.ts
index.d.ts
=== Bundled Dependencies ===
@aws-sdk/client-s3
@aws-sdk/lib-storage
@babel/runtime
@npmcli/run-script
archiver
chalk
fs-extra
npm-package-arg
npm-packlist
yargs
@aws-sdk/types
=== Zip Details ===
name:          @jcoreio/pack-lambda
version:       0.0.0-development
filename:      jcoreio-pack-lambda-0.0.0-development-20211205184649022.zip
bundled deps:  11
total files:   922
Uploading to s3://jcore-deploy/lambda/node/@jcoreio/pack-lambda/jcoreio-pack-lambda-0.0.0-development-20211205184649022.zip....done

Node.js API

writeZip

import { writeZip } from '@jcoreio/pack-lambda'

Packs and writes a .zip to disk

import { ManifestResult } from 'pacote'

async function writeZip(options?: {
  /**
   * The directory of the package to pack.  Defaults to process.cwd()
   */
  packageDir?: string
  /**
   * The directory to save .zip file in
   */
  packDestination?: string
  /**
   * If true, will output to stderr but not write to disk
   */
  dryRun?: boolean
}): Promise<{
  /**
   * The files that were packed (relative to packageDir)
   */
  files: string[]
  /**
   * The output .zip filename
   */
  filename: string
  /**
   * The package.json
   */
  manifest: ManifestResult
}>

uploadToS3

import { uploadToS3 } from '@jcoreio/pack-lambda'

Packs and uploads a .zip to S3 (without writing anything to disk)

import { ManifestResult } from 'pacote'

export async function uploadToS3(options: {
  /**
   * The directory of the package to pack.  Defaults to process.cwd()
   */
  packageDir?: string
  /**
   * The S3 bucket to upload to
   */
  Bucket: string
  /**
   * The S3 key to upload to.  Defaults to `lambda/node/${packageName}/${filename}`
   */
  Key?: string
  /**
   * If true, compute hash and append the hash to the filename; check if the S3 key
   * already exists, and if so, skip upload.
   */
  useHash?: boolean
}): Promise<{
  /**
   * The files that were packed (relative to packageDir)
   */
  files: string[]
  /**
   * The output .zip filename
   */
  filename: string
  /**
   * The package.json
   */
  manifest: ManifestResult
  /**
   * The S3 bucket the .zip was uploaded to
   */
  Bucket: string
  /**
   * The S3 key the .zip was uploaded to
   */
  Key: string
}>

createArchive

import { createArchive } from '@jcoreio/pack-lambda'

Creates an archive but doesn't write it to disk or upload it, it's up to you to pipe it somewhere

import { Archiver } from 'archiver'
import { ManifestResult } from 'pacote'

async function createArchive(options: {
  /**
   * The directory of the package to pack
   */
  packageDir: string
}): Promise<{
  /**
   * The Archiver instance to stream to `pipe` to something else.
   * Make sure to await `archive.finalize()` after piping it.
   */
  archive: Archiver
  /**
   * The files that were packed
   */
  files: string[]
  /**
   * The output filename
   */
  filename: string
  /**
   * The package.json
   */
  manifest: ManifestResult
}>