Google Cloud Function for Thumbnailization

This is a Node.js / Cloud Function implementation of 'JPEG preview thumbnails' technology used nowadays by many quality mobile / web applications on the market.

For an idea how this technology works, please see https://goo.gl/vuf9xG

This function tracks a Google Cloud Storage bucket and any image stored in the bucket is processed and the extracted thumbnail data is written to the output bucket as a file. The service using the thumbnails can similarly track the output bucket for the thumbnail data files and update their data storage objects accordingly. In addition to the thumbnail generation, the dominant color of the image is written to the output GCS file as an attribute dominantColor whose value is the color as a hex string.

Note that the exports.foo = .. part is the function entry point. The exported name must match with the name specified in the gcloud beta functions deploy command.

Released under the MIT license.

Set up your environment

export STAGING_BUCKET=<your-staging-bucket>
export INPUT_BUCKET=<your-input-bucket>
export OUTPUT_BUCKET=<your-output-bucket>
gcloud config set project <project-id>

config.js

You must create a file in the main directory with the name config.js with the following contents:

exports.OUTPUT_BUCKET = '<your-output-bucket>';

Create your buckets

Obviously you only do this once:

gsutil mb -c regional -l europe-west1 gs://$INPUT_BUCKET
gsutil mb -c regional -l europe-west1 gs://$OUTPUT_BUCKET
gsutil mb -c regional -l europe-west1 gs://$STAGING_BUCKET

Deploy the Cloud Function

gcloud beta functions deploy thumbnails --stage-bucket $STAGING_BUCKET --trigger-bucket $INPUT_BUCKET

Upload an image

gsutil cp <some-local-image-path> gs://$INPUT_BUCKET

Read the function logs

gcloud beta functions logs read --limit 20

Reassembly

To reassemble the images, use the JPEG header IM_JPEG_header.bin found in the root directory of this project.

This file has been generated by running the following and extracting the header part:

convert <input> -define jpeg:optimize-coding=false -compress JPEG -strip -quality 30 <output>

So it is important that any JPEG thumbnail data get generated by the Node.js code using the same settings to match the header properly.