Thumbnail service

This is a flask application that generates thumbnails from data stored on an OCS Science Archive.

Requirements

This script uses amazon S3 to temporarily store images. You'll need a bucket and the correct environmental variables set.

This script also relies on fits2image. Make sure it is installed in your virtualenv.

All other dependencies are listed in the pyproject.toml file and can be installed via poetry.

pip install --upgrade pip
pip install poetry
poetry install

poetry install will create a virtual environment for you, which will have all requirements installed.

Run the tests with poetry run pytest thumbservice/tests.py.

Configuration

This project can be configured using the following environment variables:

Environment Variable	Description	Default
ARCHIVE_API_URL	Archive API URL	http://localhost/
TMP_DIR	Temporary directory for image operations	'/tmp/'
AWS_BUCKET	AWS S3 Bucket to store thumbnails	'changeme'
AWS_ACCESS_KEY_ID	AWS Access Key ID for S3 Bucket	'changeme'
AWS_SECRET_ACCESS_KEY	AWS Secret Access Key for S3 Bucket	'changeme'
STORAGE_URL	URL for bucket storage if different than AWS. Defaults to None to connect to AWS S3	None
REQUIRED_FRAME_VALIDATION_KEYS	Keys from Archive API record required in order to create a thumbnail from the FITS image	'configuration_type,request_id,filename'
VALID_CONFIGURATION_TYPES	Only generate thumbnails from images of these configuration types	'ARC,BIAS,BPM,DARK,DOUBLE,EXPERIMENTAL,EXPOSE,GUIDE,LAMPFLAT,SKYFLAT,SPECTRUM,STANDARD,TARGET,TRAILED'
VALID_CONFIGURATION_TYPES_FOR_COLOR_THUMBS	Only generate color thumbnails from images of these configuration types	'EXPOSE,STANDARD'

Authorization

The API passes through the Authorization header to the archive API. You only need to provide this header if the data you are attempting to get a thumbnail for is proprietary. See the archive documentation for more details.

Endpoints

There are only 2 endpoints: /<frame_id>/ and /<basename>/ where frame_id is the ID of the frame in the archive, and basename is the base part of the filename (no file extension) you wish to make a thumbnail of. Using the frame_id is faster to return if you happen to know it ahead of time.

Both endpoints take the following query parameters:

• width
• height
• label
• image

Width and height are in pixels, label will appear as white text in the lower left had corner of the image.

They both return a url to the thumbnail file that will be good for 1 week unless the image parameter is supplied which will return an image directly.

Example

https://thumbnails.lco.global/3863274/?width=500&height=500&label=So%20many%20stars