Iris

Content delivery API with dynamic image resizing.

Serves images from an Azure Blob Storage account, the storage account "container" is specified by an environment variable to prevent serving files from other containers within the same storage account.

Currently Iris will happily serve any file within a Blob Storage container, regardless if it is an image or not. There exists a scenario where if the file extension is that of an image, and the file in blob storage is not an image and Iris is instructed to resize the image, it will cause an exception. This is accepted as it requires employees to upload non-image files to Blob Storage with an image filename.

STORAGE_ACCOUNT_CONNECTION_STRING Should be the connection string from the storage account access keys section.

STORAGE_CONTAINER Should be the storage account container name, defaults to media

LINKERD_AWAIT_DISABLED This controls the Docker entrypoint which will wait till the LinkerD proxy is ready. Setting this will stop that functionality.

Installation

To set up the virtual environment

pipenv sync --dev

To update dependencies

pipenv update --dev

Test Server

Export valid Azure storage credentials, then either run the server via python or via docker.

export STORAGE_ACCOUNT_CONNECTION_STRING=BLAH
export STORAGE_CONTAINER=BLAH

pipenv run python -m iris.server
# OR
docker build -t iris .
docker run --rm -it -p 9000:9000 -p 9100:9100 --name iris -e LINKERD_AWAIT_DISABLED=1 -e STORAGE_ACCOUNT_CONNECTION_STRING=$STORAGE_ACCOUNT_CONNECTION_STRING -e STORAGE_CONTAINER=$STORAGE_CONTAINER iris

Example requests:

http get localhost:5000/path/to/image.jpg?width=300&height=200
# or
curl -v http://localhost:5000/path/to/image.jpg?width=300&height=200
# or for production (assuming the image still exists)
curl -v https://api.gb.bink.com/content/media/hermes/bink.jpg

Pycharm Configuration

Add a new Python configuration, change Script path dropdown to Module name then set the module name to iris.server. Then update the environment variables to valid ones that are required.

VSCode Configuration

Not a clue, PyCharm FTW

Tests

Run pytest via

pipenv run pytest