aws-lambda-libreoffice

Utility to work with Docker version of LibreOffice in Lambda

Install

$ yarn add @shelf/aws-lambda-libreoffice

Features

• Includes CJK and X11 fonts bundled in the base Docker image!
• Relies on the latest LibreOffice 7.4 version which is not stripped down from features as a previous layer-based version of this package
• Requires node.js 16x runtime (x86_64)

Requirements

Lambda Docker Image

First, you need to create a Docker image for your Lambda function. See the example at libreoffice-lambda-base-image repo.

Example:

FROM public.ecr.aws/shelf/lambda-libreoffice-base:7.4-node16-x86_64

COPY ./ ${LAMBDA_TASK_ROOT}/

RUN yarn install

CMD [ "handler.handler" ]

Lambda Configuration

• At least 3008 MB of RAM is recommended
• At least 45 seconds of Lambda timeout is necessary
• For larger files support, you can extend Lambda's /tmp space using the ephemeral-storage parameter
• Set environment variable HOME to /tmp

Usage (For version 4.x; based on a Lambda Docker Image)

Given you have packaged your Lambda function as a Docker image, you can now use this package:

const {convertTo, canBeConvertedToPDF} = require('@shelf/aws-lambda-libreoffice');

module.exports.handler = async () => {
  // assuming there is a document.docx file inside /tmp dir
  // original file will be deleted afterwards

  // it is optional to invoke this function, you can skip it if you're sure about file format
  if (!canBeConvertedToPDF('document.docx')) {
    return false;
  }

  return convertTo('document.docx', 'pdf'); // returns /tmp/document.pdf
};

Usage (For version 3.x; based on a Lambda Layer)

This version requires Node 12.x or higher.

NOTE: Since version 2.0.0 npm package no longer ships the 85 MB LibreOffice but relies upon libreoffice-lambda-layer instead. Follow the instructions on how to add a lambda layer in that repo.

const {convertTo, canBeConvertedToPDF} = require('@shelf/aws-lambda-libreoffice');

module.exports.handler = async () => {
  // assuming there is a document.docx file inside /tmp dir
  // original file will be deleted afterwards

  if (!canBeConvertedToPDF('document.docx')) {
    return false;
  }

  return convertTo('document.docx', 'pdf'); // returns /tmp/document.pdf
};

Or if you want more control:

const {unpack, defaultArgs} = require('@shelf/aws-lambda-libreoffice');

await unpack(); // default path /tmp/instdir/program/soffice.bin

execSync(
  `/tmp/instdir/program/soffice.bin ${defaultArgs.join(
    ' '
  )} --convert-to pdf file.docx --outdir /tmp`
);

Troubleshooting

• Please allocate at least 3008 MB of RAM for your Lambda function.
• If some file fails to be converted to PDF, try converting it to PDF on your computer first. This might be an issue with LibreOffice itself
  • If you want to include some fonts/plugins to the libreoffice, contribute to the libreoffice-lambda-base-image instead

See Also

• libreoffice-lambda-base-image - a base Docker image for you Lambdas
• libreoffice-lambda-layer - deprecated, not updated anymore, used the Docker image above
• serverless-libreoffice - original implementation
• aws-lambda-tesseract
• aws-lambda-brotli-unpacker
• chrome-aws-lambda

Test

Beside unit tests that could be run via yarn test, there are integration tests.

Smoke test that it works:

cd test
./test.sh

# copy converted PDF file from container to the host to see if it's ok
export CID=$(cat ./cid)
docker cp $CID:/tmp/test.pdf ./test.pdf

Publish

$ git checkout master
$ yarn version
$ yarn publish
$ git push origin master --tags

License

MIT © Shelf