Introduction

This repository deploys the Mushroom Classification Application containers via Docker-Compose. The accompanying docker-compose-prod.yml and docker-compose.yml files in the root of the project can be used respectively for production or to run locally. The docker-compose-light.prod.yml shows how the apps can be deployed without the reverse proxy and certbot images, and be exposed onto an external network (for example, the one being used by your reverse proxy).

Repositories Deployed	DockerHub image	Badges
Mushroom Classification Flask API	mushroom-api
Mushroom Classification Front-End	mushroom-classification-front-end

The production yml file uses nginx as the reverse proxy server and gunicorn as the WSGI server to the serve the application.

For information on how to use Docker and Docker Compose consult the documentation.

Docker Containers Setup

For both docker-compose files a .docker.env file needs to be present in the root of the project. This should contain all the environment variables needed for the Flask API to run. For environment variables that specify paths for the Flask API container the 'working directory' should be assumed to be the src folder.

An example .docker.env file, excluding any sensitive information, would be like:

FLASK_ENV=development FLASK_APP=./api/__init__.py DATASET_DIR=./files CURRENT_MODEL_DIR=./current_model TRAINED_MODELS_DIR=./trained_models FEATURE_DEFINITION_PATH=./files/features-definition.json LOGS_DIRECTORY=./api_logs

Also, a .react.docker.env file must be present in the root of the project, containing the environment variable pointing to the api url, an exmaple would look like: API_BASE_URL=https://mushroomlapi.callums-stuff.net/

For the production docker-compose file an nginx.conf file needs to be present in the nginx directory. For instructions on how to setup this file read the nginx documentation. Or, contact the repository owner for an example.