The DeepHealth back-end interacts with the front-end, serving various APIs. It receives a configuration from the front-end then runs a deep learning pipeline based on PyECVL and PyEDDL.
API documentation: jenkins-master-deephealth-unix01.ing.unimore.it/backend/swagger
Full DB schema:
- Python3.6+
- git curl --
sudo apt install git curl - Requirements for PyECVL 0.10.0+ and PyEDDL 0.14.0+
Clone and install back-end with:
cd ~
git clone https://github.com/deephealthproject/backend.git
cd backend
virtualenv env
source env/bin/activate
pip install -r requirements.txt
chmod u+x scripts/pyecvl_install.sh
scripts/pyecvl_install.shGenerate a new SECRET_KEY with:
python -c 'from django.core.management.utils import get_random_secret_key;print(get_random_secret_key())'Edit the config file to configure the application (SECRET_KEY, DB and RabbitMQ connection and other optional Django settings).
Install with: sudo apt install rabbitmq-server
and run the celery deamon with: python manage.py celery.
The back-end is a web-server based on Django, so it must be initialized like any Django project.
cd ~/backend
# Apply all the migrations
python manage.py migrate
# Creating an admin user
python manage.py createsuperuser
# Initialize Django fixtures for the DB init
# This will also download ONNX weights of the models
python scripts/init_fixtures.py
# Load fixtures into the DB (default entries)
python manage.py loaddata tasks.json property.json allowedproperty.json dataset.json model.json modelweights.json auth.json
# Start the development server
python manage.py runserver <my-server>:<my-server-port>
# Start celery
python manage.py celeryRead the auth/README.md file for configuring the authentication backend.
Tutorials, get you started with understanding and using the back-end:
The following table contains Postman collections for practicing with the back-end.
| Collection | Postman Link |
|---|---|
| Backend API collection |  |
| 2nd Hackathon - Pneumothorax | |