Artwork Image Processor

This is a Django application that leverages a Deep Neural Net to transfer painting styles onto images.

The application also reaches out to the PhotoTagger API to tag any uploaded images with classifcations based on what is in the image (Ex: dog, train, car, etc)

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.

Prerequisites

These are the Python dependencies required to build and run the application:

Python 3.6.6
Django 2.1.2
Bootstrap 4.1.3
Pillow 5.3.0
Numpy 1.14.5
Gunicorn 19.9.0
Whitenoise 4.1
Dj Database URL 0.5.0
requests 2.20.0
imutils 0.5.1
Scipy 1.1.0
OpenCV 3.4.3.18

Installing your Development Environment

1. Install all dependencies noted above
2. Navigate to the project root in your command line / terminal
3. Execute the migrate command to create any necessary database tables

• Ex: python manage.py migrate

4. Execute the collectstatic command to populate the staticfiles directory

• Ex: python manage.py collectstatic

5. Execute the runserver command to start the development server

• Ex: python manage.py runserver <port>
• The runserver command defaults to port 8000

6. In a browser, navigate to http://localhost:<port> to load the root webpage

Configuring PhotoTagger API Endpoint

The PhotoTagger API endpoint can be configured by setting the PHOTOTAGGER_CLASSIFY_API_ENDPOINT variable in settings.py

Automated Test Suite

The project has a suite of automated tests that are built on top of the Django test framework, which is built on top of the unittest module built in to the Python standard library.

Test Files

The project's automated tests are defined in the following files:

Test File	Description
<project_root>/artwork_image_processor/tests/tests_forms.py	Tests validating Django forms in the project
<project_root>/artwork_image_processor/tests/tests_models.py	Tests validating Django models in the project
<project_root>/artwork_image_processor/tests/tests_views.py	Tests validating Django views in the project
<project_root>/artwork_image_processor/tests/tests_style.py	Tests validating the Image Processing logic in the project
<project_root>/artwork_image_processor/tests/tests_tags.py	Tests validating the Photo Tagger API Integration logic in the project

Running the tests

Run the following commands to execute the test suite:

python manage.py test

The test suite is also run automatically as part of our pull-request process. A PR cannot be merged unless all tests are passing.

Generating a Code Coverage Report

A code coverage report can be generated using the Coverage.py utility for python. It monitors your program, noting which parts of the code have been executed, then analyzes the source to identify code that could have been executed but was not.

Coverage.py can be installed using pip or other python dependency management frameworks. Ex: pip install coverage

Run the following commands at the project root to build and view the code coverage report for this project:

coverage run --source='.' manage.py test
coverage report

Deployment

Our current deployment environment is Heroku. We have a CI/CD Pipeline that will automatically deploy changes that are successfully merged into the master branch.

• Our live URL is: https://art-processor.herokuapp.com/

The primary distinction between running the application locally vs in a deployment environment involves the management of the project's static files. We have configured the application to successfully deploy to our Heroku deployment environment, but there may be additional requirements in other deployment environments.

For additional information on how to handle static files when deploying Django applications, see the following links:

• Django - Managing Static Files
• Django - Deploying Static Files
• Heroku - Django and Static Assets

Monitoring

We are currently using two different monitoring solutions for our deployments:

• FreshPing - Uptime Monitoring
• New Relic APM

Contributing

Please read CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests to us.

Authors

• Cecil Beeland - Initial work - beelandc
• Camille Rosewright - Initial work - kamiyokamiru
• Wunyan Wong - Initial work - wunyanw
• Greg Zimmerman - Initial work - gjz999gitit

See also the list of contributors who participated in this project.

License

This project is licensed under the GPL 3.0 License - see the LICENSE.md file for details

Acknowledgments

• For inspiration:
  • Udacity - How to Process Images With TensorFlow - by Mat Leonard
  • How to Upload Files with Django - by Vitor Freitas
  • Neural Style Transfer with OpenCV - by Adrian Rosebrock