Skip to content

Latest commit

 

History

History
262 lines (217 loc) · 5.54 KB

README.md

File metadata and controls

262 lines (217 loc) · 5.54 KB

django-project

A Webapp for ML computations.

Installation

Go to the folder where you want to start creating your project.

Create conda environment, e.g. via

conda create -n djangoenv python=3.8

Then activate it with

conda activate djangoenv

Install the requirements in the nevironment via

pip install -r requirements.txt

Finally, set the up the new project with

django-admin startproject mysite

In the newly created folder, you will find the following files:

  • settings.py a file containing the generic settings of the project
  • urls.py a file containing the urls of the apps of the project
  • wsgi.py a file for the WSGI deployment

To make the app work

Run django server

Go inside the outer gtdainterface folder and then run:

python manage.py runserver 8080

Run Redis server

Just type

redis-server

Run Redis-RQ worker

Go inside the outer gtdainterface folder and then run:

python manage.py rqworker default

.env file

Create a .env file with the following secret variables:

export SENDGRID_API_KEY = 'xxx'
export RQ_API_TOKEN = 'xxx'
export SECRET_KEY = 'xxx'
export DROPBOX_OAUTH2_TOKEN = 'xxx'
export DATABASE_URL = 'xxx'

In particular:

  • SENDGRID_API_KEY is for sending emails with sendgrid
  • RQ_API_TOKEN is for RQ to contact the Redis server
  • SECRET_KEY this is the django app secret key
  • DROPBOX_OAUTH2_TOKEN tis s to configure Dropbox storage API
  • DATABASE_URL this is for the online PostgreSQL database

Create an app

Go inside the outer mysite folder, and then choose the name -- polls in this case -- and run

python manage.py startapp polls

Inside the newly created app folder you will find:

  • views.py this file is to define the classes or functions associated to the views (basically, the get response). Each function must return the httpResponse
  • admin.py is the file to improve the admin view
  • models.py is teh file with the definition of the classes associated to the database structures
  • tests.py is for the unit and integration tests
  • urls.py specifies all the urls associated to the views
  • apps.py a file with the app configurations

Create tables in DB

To create tables in the databse for auth, session, contenttype, messages, staticfiles write

python manage.py migrate

Connect to PostgreSQL

First you need to install psycopg2 :

pip install psycopg2

and then

Migrating new model

If you create a new DB model, then add its migration via

python manage.py makemigrations polls

If you are using postgreSQL, then use

python manage.py sqlmigrate polls 0001

To actually apply the model tables in your database, use

python manage.py migrate

Django Python interactive interface

To open up the Python shell use

python manage.py shell

Create admin user

Just type

python manage.py createsuperuser

E.G. admin, [email protected], admin

Run tests

To run the tests, that you can write in the polls/tests.py file, write

python manage.py test polls

Run Redis server

Forr ML ocmputation, it is better to have a Redis server that enqueues the jobs and execute them in the background.

Install Redis and Python rq:

brew install redis
pip install django-rq

To star the redis server, run

redis-server

Once you are done, follow the instructions of django-rq. Then run the workers with:

python manage.py rqworker default

You can also add the high and low queues

Use signals

You can use signals that are triggered when a certain event happens. For exammple, every time an entry is saved in a table of the database.

To do so, initalise the connections in the apps.py file and then create a new signals.py module with the receiver functions.

Deploy on Heroku

Check it out

Visit the app site now!

Installation

Need to install postgres Need to download te heroku cli:

brew install heroku/brew/heroku

Login with

heroku login

Then create and deploy the app: first move to the root folder

heroku create
git commit -am "commit files and modifications"
git push heroku main
heroku ps:scale web=1
heroku open

To scale down heroku app processes:

heroku ps:scale web=0

Clear built packages

To clear the build, for example when you modify the Procfile

heroku buildpacks:clear

and the re-coimmit

git commit --allow-empty -m "Adjust buildpacks on Heroku"
git push heroku main

Migrate DB

To migrate the database, type

heroku run python manage.py migrate

Create admin on Heroku

Type

heroku run python manage.py createsuperuser

Collect static files

Simply run

python manage.py collectstatic

Run bash on Heroku

Simply:

heroku run bash

Run django-python shell on Heroku

Simply

heroku run python manage.py shell

Configure env variables

Set secrets using heroku config

heroku config:set VAR = 'xxxx'

Add Redis on heroku

heroku addons:create redistogo:nano

Add this to the Procfile:

worker: python gtdainterface/manage.py rqworker high

And add a worker to heroku:

heroku ps:scale worker=1

Migrate database

Activate PostgreSQL database

heroku addons

and check the heroku config to find the db url. To make the migration

heroku run python manage.py migrate

Heroku Logs streaming

Write in the terminal:

heroku logs --tail