Image Processing

This is a RESTful API able to upload and apply effects to images using Django Rest Framework that has

• Docker to run all the needed services
• Uses Cloudflare R2 with boto3 to store images
• Celery to perform transforms on the background as tasks, without blocking the app
• Redis as Celery broker and results backend, as well as application cache store for recent data.
• JWT for authentication
• Unit tests

The images are private to the users and retrieving image details returns a temporary URL that grants access for a period of time.

Set Up

to be able to use this repo you should first get a Cloudflare R2 account and create a bucket for storing the images. From there you should get the necesary keys to store in an .env file:

    R2_STORAGE_URI=`<your-account-url>`
    R2_ACCOUNT_ID=`<your-account-id>`
    R2_BUCKET_NAME=`<bucket-name>`
    R2_API_TOKEN=`<your-api-token>`
    
    AWS_SECRET_ACCESS_KEY=`<secret-access-key>`
    AWS_ACCESS_KEY_ID=`<secret-key-id>`

you'll need also your postgres credentials

    POSTGRES_DB=`<your-db-name>`
    POSTGRES_USER=`<your-user-name>`
    POSTGRES_PASSWORD=`<your-password>`

a Django secret key

    SECRET_KEY=`<secret>`

and the Redis urls needed for Celery and the in memory cache

    REDIS_CACHE_URL='redis://redis:6379/1'
    
    CELERY_BROKER='redis://redis:6379/0'

Then run from the projects root:

    pipenv install --system

    docker compose up -d --build

    docker compose exec web python manage.py migrate

Usage

Auth

POST hostname:port/api/v1/accounts/register

body: {
  'password',
  'username',
  'email'
}

response: {
  'refresh',
  'access'
}

POST hostname:port/api/v1/accounts/token

body: {
  'password',
  'username',
}

response: {
  'refresh',
  'access'
}

POST hostname:port/api/v1/accounts/token/refresh

body: {
  'password',
  'username',
}

response: {
  'access'
}

POST hostname:port/api/v1/accounts/logout

headers: {
  'Authorization': 'Bearer <accesstoken>'
}

Upload image

POST hostname:port/api/v1/images

headers: {
  'Authorization': 'Bearer: <access-token>'
}

form-data: {
  'image': '<filepath>'
}

Image Detail

specified by UUID which is the image id

GET hostname:port/api/v1/images/<uuid>

headers: {
  'Authorization': 'Bearer: <access-token>'
}

response: {
  'id'
  'file_key', # internal R2 key
  'user_id',
  'created_at',
  'url',
}

Images List

returns a list of paginated image details that the user owns. Query params will fall back to defaults if missing or invalid

GET hostname:port/api/v1/images

query_params: {
  'pages', #optional
  'page_size', #optional
  'limit' #optional
}

headers: {
  'Authorization': 'Bearer: <access-token>'
}

the response is a json array with all the pages with the image details, each image has a access url to be able to use it for a period of time

Delete Image

DELETE hostname:port/api/v1/images/<uuid>

headers: {
  'Authorization': 'Bearer: <access-token>'
}

Image Transform

apply transforms to image and receive the image details as response, all the listed effects are optional, only a one is needed to work

PUT hostname:port/api/v1/images/<uuid>/transform

headers: {
  'Authorization': 'Bearer: <access-token>'
}

body: {
  resize: {
    width,
    height
  }
  crop: {
    width,
    height,
    x, 
    y
  }
  rotate: Float
  flip: Char # either lr (left-right) or tb (top-bottom)
  filters: {
    sepia,
    grayscale
  }
}

response: {
  task_id
}

this will create a background task that you can poll using task_id

GET hostname:port/api/v1/images/tasks/<task_id>

headers: {
  'Authorization': 'Bearer: <access-token>'
}

response: {
  task_id,
  status,
  # if finished
  result: {
    success, 
    error,
    data
  }

}

Tests

with docker container running execute:

docker compose exec web python manage.py test

pytest -m api/v1/images/tests/test_effects.py

https://roadmap.sh/projects/image-processing-service