This week we got started with Django, and began making Djorg, a project for personal organization applications. To close the week, the challenge is - deploy! Getting your application out there is great to learn, shake out bugs, and get feedback as you can share it with others.
Note: the instructions below assume you're on your master
branch in git.
The steps to deploy (at a high level) are:
- Sign up for Heroku
- Install the Heroku CLI
- From your terminal,
heroku login
- Get to your project/repo directory
- Install new dependencies. (If using
virtualenv
, usepip install
as you have been, or migrate topipenv
.)pipenv install gunicorn
- the webserver for Heroku to use (rather than the one built-in to Django)pipenv install psycopg2-binary
- PostgreSQL client binariespipenv install dj-database-url
- enables parameterizing the database connection (so Heroku uses PostgreSQL but local is still SQLite)pipenv install python-decouple
- set important/secret values as environment variablespipenv install whitenoise
- optimizes deployment of static files (you may not have any, but it's good to add this now)- If using
virtualenv
, you need to create arequirements.txt
file in your project root directory with the command:pip freeze > requirements.txt
- Prepare your project
- Copy the
dotenv
file in this repository to.env
in your repository (this should not be checked in) ALLOWED_HOSTS
andDATABASE_URL
are probably already correct for your local environment, but read/understand them- Use the example code (you can just run it in a
python
repl) to generate a new secret key and changeSECRET_KEY
djorg/settings.py
will need new imports (from decouple import config
andimport dj_database_url
)- You can use
config
to load the environment variables you set above, e.g.SECRET_KEY = config('SECRET_KEY')
(ALLOWED_HOSTS
will be a little trickier, but that's why this is a sprint challenge!) - For the database, you want to both load the
DATABASE_URL
and pass it todj_database_url.config
(see documentation) - Make a
Procfile
(example) to tell Heroku what to run to start your app. (Hint: the name of your Django project is probably "djorg", not "gettingstarted".) - Configure
whitenoise
(add a few configuration lines to yoursettings.py
file per the documentation)
- Copy the
heroku create your-app
- makes the project and adds Heroku as a remote to your git repository so you can push to it to deployheroku addons:create heroku-postgresql:hobby-dev
- makes a PostgreSQL database associated with the project (and sets theDATABASE_URL
Heroku config var, equivalent to a local environment variable)- Set the other Heroku config vars, e.g.
ALLOWED_HOSTS=.herokuapp.com
,DEBUG=False
, andSECRET_KEY=somenewsecret
- see the documentation, you can set either via the Heroku CLI or by logging in to the Heroku Dashboard in your browser - Deploy! First, make sure you're all committed. Then
git push heroku master
Once you've got it deployed, you'll probably need to run migrations on Heroku
(since it's using a different database then local). You can do this with
heroku run python manage.py migrate
, and in general
heroku run python manage.py
is the way to do all the helpful Django things,
just "in production." For example, heroku run python manage.py createsuperuser
is a good thing to do, and you can even administer/explore the app/data with
heroku run python manage.py shell
and heroku run python manage.py dbshell
.
Of course, chances are you'll run into things as you make your way through the steps above - and that's okay! Some resource to help:
- Getting Started on Heroku with Python - official tutorial, has a full working example app if you want to step through with that before you try with Djorg
- Lambda School Python/Django resources
You may also find the heroku logs
command useful to diagnose errors, as Heroku
error messages are usually not that descriptive. And, while this is a sprint
challenge, it is okay to chat some about your progress and issues you hit - just
make sure you actually type and understand all of your own code, even if you are
finding it from somewhere else. It's also a good practice to note the resources
you use, as comments in your code and sharing them with others via Slack.
The review lecture will step over the above process, giving you a chance to figure out any parts you miss. But please open a PR before then with:
- A link to your Djorg project repo
- A link to your live site, if you were able to deploy
- A
DeploymentExperiences.md
file where you write summarizing how the process went for you, what went well and what was tricky, and how far you got - A
SoftwareDesign.md
file where you write about:- Your favorite software design pattern you learned this week, why, and a situation you think it'd be useful
- A software design anti-pattern that you've run into, what happened, and what you did to fix the situation
For both writing portions, this is meant to practice your skills at professional writing - in the tech world, that means writing things clearly, concisely, in a way to get the information across efficiently to an audience that may be pretty busy (most people only skim most emails). Suggested length ~1-2 paragraphs for each topic, and bullet lists can be extremely effective.
Note for bash users: if you ever get in a place where your terminal isn't
printing out newlines properly and is not echoing your keypresses, hit CTRL-C
then blindly type in the following command and hit RETURN
:
stty sane