Django Auth Template

This template contains a project, gratitude-back-end, and an app, api, which are set up complete with user authentication and an example resource, Mango, which has an example user ownership implementation.

Preparation

Download this template.
Move the .zip file to your sei/projects/ directory and Unzip it (creating a folder) -- NOTE: if the folder was already unzipped, use the mv command line to move it to the sei/projects/ directory.
Empty README.md and fill with your own content.
Move into the new project and git init.
Create and checkout to a new branch, training, for your work.
Create a .env file
Add a key ENV with the value development exactly.
1. Note: When you deploy, you will create this key on Heroku the value production. This will distinguish the development and production settings set in this template.
Run pipenv shell to start up your virtual environment.
Run pipenv install to install dependencies.
Create a psql database for your project
1. Edit settings.sql then run psql -U postgres -f settings.sql OR:
2. Type psql to get into interactive shell.
3. Run CREATE DATABASE "project_db_name"; where project_db_name is the name you want for your database.
Add the database name to the .env file using the key DB_NAME_DEV.
Replace all instances of gratitude-back-end with your application name. This includes the folder included in this repository.
Generate a secret key using this tool and add it to the .env file using the key SECRET.
Open the repository in Atom with atom .

The `.env` File

After following the steps above, your .env file should look something like the following, replacing project_db_name with your database name and secret_key with your secret key.

ENV=development
DB_NAME_DEV=project_db_name
SECRET=secret_key

Structure

This template includes a project gratitude-back-end which should be renamed as part of the set-up steps. It includes the settings.py file with special settings to be able to run both locally and on production. DO NOT ADD A NEW OR MODIFY THE CURRENT DATABASES DEFINITION UNLESS INSTRUCTED TO DO SO.

There is also an app api which can be renamed if necessary. The api app includes folders for models and view files, which can then be imported into urls.py for use.

Commands

Commands are run with the syntax python3 manage.py <command>:

command	action
`runserver`	Run the server
`makemigrations`	Generate migration files based on changes to models
`migrate`	Run migration files to migrate changes to db
`startapp`	Create a new app

Deployment

Before deploying, make sure you have renamed your project folder and replaced all instances of gratitude-back-end with your app's name.

Once ready, you can follow the steps in the django-heroku-deployment-guide.

Connecting Client

This template is intentionally minimal, and does not override many of Django's defaults. This means connecting either the browser-template or react-auth-template clients to this backend involves updating that client code slightly.

Ultimately, Django and any other backend API framework should be able to build standalone backend APIs that can talk to any client. We just have to make sure the client is following some of the expectations that Django has by default.

Port

When working on our "local" computer, we work on the localhost location. This is paired with a port number to identify where our server is running on our local machine. Our client templates use port 7165, for example, and run at http://localhost:7165.

These templates also talk to a backend at a certain port, which is set to 4147 in the client templates. We need to change the port in the URL the client application uses when running locally.

In the browser-template this means modifying the config.js file, and in the react-auth-template the apiConfig.js file.

This django template uses the port 8000 by default, so any client speaking to this template's default server location would be http://localhost:8000.

URL Syntax

Django defaults to expecting (and requiring) trailing forward slashes / on requests. You'll need to make sure any requests you make from a client to this template look something like http://localhost:8000/books/.

Token Syntax

We've gotten used to the token syntax that the Express framework expects:

Bearer 9944b09199c62bcf9418ad846dd0e4bbdfc6ee4b

However, the DRF TokenAuthentication class in this template is what defines what our token syntax should look like when the client make an authenticated request to our Django application.

When making authenticated requests from any client, make sure your tokens follow this pattern:

Token 9944b09199c62bcf9418ad846dd0e4bbdfc6ee4b

Incoming Request Data

Our JavaScript conventions are to use camelCase when defining almost everything, however Python conventions (and therefore Django conventions) use snake_case instead. We have to keep this in mind when sending data from a client to a Django application.

Debugging

pipenv shell moved me into a different directory!

Pipenv wants to be in the root directory, so if it thinks it's not then it will move you to what it thinks is the root of your repository. Exit out of the shell with exit, then check if the folder it moved you to is a git repository. If you see a .git folder inside of the trainings folder, for example, delete that folder so that trainings is no longer a "git repo." Then, you can change back into your project directory and try running pipenv shell again.

pipenv shell is complaining about my python version not matching

Our python version is defined in the Pipfile. Simply replace the current python_version = "x.x" statement with the appropriate version.

SyntaxError pointing to manage.py when trying to run the server, migrate, etc.

Double-check your python version with python --version. If you see a "2.x.x" version, then you need to use the command python3 when running python scripts. You can also follow these guides to replace your python command so it always uses python3.

Mac: https://stackoverflow.com/questions/49704364/make-python3-as-my-default-python-on-mac/49711594

Linux: https://linuxconfig.org/how-to-change-from-default-to-alternative-python-version-on-debian-linux

Error: No module named when trying to run the server

If Python can't find the module that is your project name, then very likely you forgot a very important piece of the preparation steps. You need to make sure you rename the project folder as well.

I made changes to my models & ran my migrations but it says "No migrations to apply"

Double-check that you generated the migration files before you tried to run them. This means running makemigrations before migrate.

Errors with psycopg2

There's a lot to read about this issue if you want: psycopg/psycopg2#674 https://www.psycopg.org/articles/2018/02/08/psycopg-274-released/

This template uses psycopg2-binary to minimize errors during project development. If you have errors with psycopg2 anyway, notify an instructor.

Additional Resources

Django Rest Framework Tutorial: Authentication

License

All content is licensed under a CCBYNCSA 4.0 license.
All software code is licensed under GNU GPLv3. For commercial use or alternative licensing, please contact legal@ga.co.

robrichardsdpt/gratitude-back-end