Skip to content

Latest commit

 

History

History
157 lines (95 loc) · 5.56 KB

Heroku_Deployment_Guide.md

File metadata and controls

157 lines (95 loc) · 5.56 KB
Raw

Heroku Deployment

Prerequisites

Deploying an application

  • In this activity, we will deploy a Pet Pals application to Heroku. The applications takes the name of a pet and plot it's location on the map. The actual code for the application is not nearly as import as the steps for deploying to Heroku. These steps can repeated for your own applications.

  • This process consists of:

    1. Creating a repo for the application.
    2. Preparing the application with additional configuration files (Procfile and requirements.txt)
    3. Creating the Heroku application
    4. Preparing the Heroku database

Part 1: Create a New Repo

  • Files: Pet Pals app

  • From Github create a new repo called Pet_Pals and clone it to your desktop.

  • Add the starter files to this repo.

Part 1: Configuration Files

  • Start by creating a new conda environment just for this app. All of our project dependencies will be installed in this environment. Note: This should only contain python 3.6 and not anaconda.
conda create -n pet_pals_env python=3.6
  • Make sure to activate this new environment before proceeding.
conda activate pet_pals_env
  • Note if you run into issues try the following command instead.
source activate pet_pals_env

  • Next, we install gunicorn with pip install gunicorn. Explain that gunicorn is a high performance web server that can run their Flask app in a production environment.

  • Because this app will use Postgres, we also install psycopg2 with pip install psycopg2.

  • Install the remaining libraries into your new environment.

pip install flask
pip install flask-sqlalchemy
pip install pandas
  • Navigate to the folder that contains initdb.py and run the following to initialize the database.
python initdb.py
  • Create a shell file and add the following code.
FLASK_APP=pet_pals/app.py flask run
  • You can test the application by running the following in your command line.
./run.sh

  • Navigate to 127.0.0.1:5000 to view your webpage.

  • Now that all of the the project dependencies are installed, we need to generate the requirements.txt file. This file is a list of the Python packages required to run the app. Heroku will use this file to install all of the app's dependencies.

pip freeze > requirements.txt
  • The final configuration file that we need is Procfile. This file is used by Heroku to run the app.
touch Procfile
  • Add the following code to the Procfile which instructs Heroku how to run the app.
web: gunicorn pet_pals.app:app

  • pet_pals is the name of the folder that contains your app as a python package (i.e. the name of the folder with the __init__.py file in it).

  • Add, commit and push everything up to your repo.

Part 2: Creating the Heroku App

  • Navigate to Heroku and create an account.

  • Once you are at the main dash click New in the type right and select Create a new app.

    • Give your app an unique name and leave region to default.

  • On Heroku, go to the Deploy section of your app's homepage, and follow the steps to deploy the app.

    Images/deploy05.png

    • In the Deployment method section select GitHub.

    • Once connected search for the pet-pals repo you created that contains your code from the previous step, and connect.

    • With your repo selected, navigate to the Manual deploy section below it and click Deploy Branch.

    • To confirm your app has been successfully deployed navigate to the top of the page and click Open app. This should open a webpage with you pet pals web page. Note the database has not been set up yet, so there is one more step before it is fully functioning.

Part 3: Preparing the Database

  • After creating a new app on Heroku, navigate to Resources:

    Images/deploy01.png

    • Under Add-ons, search Heroku Postgres. Make sure to use the free version then click Provision.

    provision database

  • Once Herokue Postgres is listed on click on it.

  • From the new page, navigate to settings and click on View Credentials.

  • The connection string to the database should now be available in the URI field:

    Images/database_connection.png

  • Heroku will automatically assign this URI string to the DATABASE_URL environment variable that is used within app.py. The code that is already in app.py will be able to use that environment variable to connect to the Heroku database.

    # DATABASE_URL will contain the database connection string:
app.config['SQLALCHEMY_DATABASE_URI'] = os.environ.get('DATABASE_URL', '')
# Connects to the database using the app config
db = SQLAlchemy(app)

  • The final step requires the Heroku CLI. If you do not currently have it installed please follow the instructions for the Heroku CLI.

  • After adding the database, the final step is to initialize the database. To do this, we use the heroku cli. From the terminal, type the following, be sure to replace <name of your app> with the name of your app as it appear in heroku:

heroku run python initdb.py -a <name of your app>
  • Your database is now initialized, and you can open the application using heroku open from the terminal or from Open App on the webpage.