/photo-processor

Primary LanguagePython

photo-processor exercise

Description

The aim of the exercise is to orchestrate the generation of thumbnails for a specific set of photos.

Installation

Prerequisites:

  • Docker
  • Ability to run make.

App is bundling Postgres and RabbitMQ instances via Docker, so please stop any local related services to avoid port conflicts. Otherwise you can amend the default port mappings on the docker-compose file.

Start the app:

  • make start

Create or reset the db schema after booting the app:

  • make db-schema

Postgres PSQL can be accessed via:

  • make psql

RabbitMQ management console can be accessed at:
http://localhost:15672/

Web app can be accessed at:
http://localhost:3000/

Available Python libraries on the container

  • Flask (for the web app)
  • Psycopg2 (for accessing postgres)
  • Pillow (for generating thumbnails)

You are welcome to use any Python library of choice, even replacing some of the above defaults.

Tasks to be performed

1. Update db-schema

  • Add on photo_status enum two new statuses named processing and failed.
  • Create a table named photo_thumbnails with columns:
    • uuid (uuid) (auto-generated)
    • photo_uuid (uuid) (FK from photos.uuid) (not null)
    • width (smallint) (not null)
    • height (smallint) (not null)
    • url (text) (not null)
    • created_at (similar to photos.created_at)

2. Add web endpoint for fetching photos of pending status

  • Method: GET
  • URL: /photos/pending

Should return photo records in JSON format.

3. Add web endpoint for triggering the processing of pending photos

  • Method: POST
  • URL: /photos/process

Endpoint should be accepting one or more photo UUIDs as JSON input.
It should be producing one RabbitMQ message on a queue named photo-processor for every photo to be processed.

4. Create RabbitMQ consumer

Create a barebones RabbitMQ consumer that is listening on the photo-proccessor queue, processing one message at a time.
Consumer should be running on the waldo-app container alongside the web app.

5. Process photo

  • Update db photos.status to processing.
  • Download image using photos.url.
  • Generate a thumbnail of max 320x320 dimensions, maintaining the aspect ratio.
  • Store thumbnail file on mounted /waldo-app-thumbs directory.
  • Store a new row on db table photo_thumbnails with the thumbnail details. For the photo_thumbnails.url just use the relative path to the file.
  • On success, update photos.status to completed.
  • On error, update photos.status to failed.

Deliverables

  • Solution should be using Python 3.7.x, and extending the given setup.
  • Git repository with any required instructions for running it, uploaded on GitHub or GitLab or BitBucket.