Python Lambda LibreOffice Conversion Service

This project contains a Python Lambda function that converts documents to PDF using LibreOffice. The conversion service runs inside a custom AWS Lambda Docker container. It can then be hosted on ECR and integrated with Terraform to work in a CI/CD pipeline to quickly deploy Lambdas to convert files.

Docker Images

This project uses two Dockerfiles:

1. base.dockerfile: This Dockerfile is used to create the base image that installs and configures LibreOffice.
2. lambda.dockerfile: This Dockerfile is used to create the Lambda image that contains the conversion service.

The conversion service runs as a Python script (main.py). The script downloads a document from an S3 bucket, converts it to PDF using LibreOffice, and uploads the resulting PDF back to S3.

How to Use

Building Docker Images

The base image can be built using the command:

docker build --no-cache -t <image name> -f base.dockerfile .

The Lambda image can be built using the command:

docker build --no-cache -t lambda-image -f lambda.dockerfile .

Deploying to AWS

1. Tag the Docker images:

docker tag <image name>:latest <account id>.dkr.ecr.<region>.amazonaws.com/<image name>:base
docker tag lambda-image:latest <account_id>.dkr.ecr.<region>.amazonaws.com/libreoffice-image:lambda

2. Log in to AWS ECR:

aws ecr get-login-password --region <region> | docker login --username AWS --password-stdin <account id>.dkr.ecr.<region>.amazonaws.com

3. Push the Docker images to ECR:

docker push <account id>.dkr.ecr.<region>.amazonaws.com/<image name>:base
docker push <account_id>.dkr.ecr.<region>.amazonaws.com/libreoffice-image:lambda

Please see the comments in the Dockerfiles for more detailed instructions and useful Docker commands.

Environment Variables

The Lambda function uses the following environment variables:

• OBJECT_KEY: The key of the S3 object to be converted.
• INPUT_BUCKET_NAME: The name of the S3 bucket where the document to be converted is stored.
• OUTPUT_BUCKET_NAME: The name of the S3 bucket where the converted PDF should be uploaded.

Running the Lambda Function Locally

To run the Lambda function locally, use the command:

docker run -d --name test lambda-image
docker exec -it test /bin/sh

License

This project is open source and available under the MIT License.