Flask by example part 2

Flask by Example – Setting up Postgres, SQLAlchemy, and Alembic

by Real Python databases devops flask web-dev

In this part we’re going to set up a Postgres database to store the results of our word counts as well as SQLAlchemy, an Object Relational Mapper, and Alembic to handle database migrations.


  • 03/22/2016: Upgraded to Python version 3.5.1 as well as the latest versions of Psycopg2, Flask-SQLAlchemy, and Flask-Migrate. See below for details.
  • 02/22/2015: Added Python 3 support.

Remember: Here’s what we’re building - A Flask app that calculates word-frequency pairs based on the text from a given URL.

  1. Part One: Set up a local development environment and then deploy both a staging and a production environment on Heroku.
  2. Part Two: Set up a PostgreSQL database along with SQLAlchemy and Alembic to handle migrations. (current)
  3. Part Three: Add in the back-end logic to scrape and then process the word counts from a webpage using the requests, BeautifulSoup, and Natural Language Toolkit (NLTK) libraries.
  4. Part Four: Implement a Redis task queue to handle the text processing.
  5. Part Five: Set up Angular on the front-end to continuously poll the back-end to see if the request is done processing.
  6. Part Six: Push to the staging server on Heroku - setting up Redis and detailing how to run two processes (web and worker) on a single Dyno.
  7. Part Seven: Update the front-end to make it more user-friendly.
  8. Part Eight: Create a custom Angular Directive to display a frequency distribution chart using JavaScript and D3.

Need the code? Grab it from the repo.

Install Requirements

Tools used in this part:

  • PostgreSQL (9.4)
  • Psycopg2 (2.6.1) - a Python adapter for Postgres
  • Flask-SQLAlchemy (2.1) - Flask extension that provides SQLAlchemy support
  • Flask-Migrate (1.8.0) - extension that supports SQLAlchemy database migrations via Alembic

To get started, install Postgres on your local computer, if you don’t have it already. Since Heroku uses Postgres, it will be good for us to develop locally on the same database. If you don’t have Postgres installed, Postgres.app is an easy way to get up and running for Mac OS X users. Consult the download page for more info.

Once you have Postgres installed and running, create a database called wordcount_dev to use as our local development database:

$ psql
# create database wordcount_dev;
# \q

In order to use our newly created database within the Flask app we to need to install a few things:

$ cd flask-by-example

cding into the directory should activate the virtual environment and set the environment variables found in the .env file via autoenv, which we set up in part 1.

$ pip install psycopg2==2.6.1 Flask-SQLAlchemy===2.1 Flask-Migrate==1.8.0
$ pip freeze > requirements.txt

If you’re on OS X and having trouble installing psycopg2 check out this Stack Overflow article.

Update Configuration

Add SQLALCHEMY_DATABASE_URI field to the Config() class in your config.py file to set your app to use the newly created database in development (local), staging, and production:

import os

class Config(object):

Your config.py file should now look like this:

import os
basedir = os.path.abspath(os.path.dirname(__file__))

class Config(object):
    DEBUG = False
    TESTING = False
    SECRET_KEY = 'this-really-needs-to-be-changed'

class ProductionConfig(Config):
    DEBUG = False

class StagingConfig(Config):
    DEBUG = True

class DevelopmentConfig(Config):
    DEBUG = True

class TestingConfig(Config):
    TESTING = True

Now when our config is loaded into our app the appropriate database will be connected to it as well.

Similar to how we added an environment variable in the last post, we are going to add a DATABASE_URL variable. Run this in the terminal:

$ export DATABASE_URL="postgresql://localhost/wordcount_dev"

And then add that line into your .env file.

In your app.py file import SQLAlchemy and connect to the database:

from flask import Flask
from flask.ext.sqlalchemy import SQLAlchemy
import os

app = Flask(__name__)
db = SQLAlchemy(app)

from models import Result

def hello():
    return "Hello World!"

def hello_name(name):
    return "Hello {}!".format(name)

if __name__ == '__main__':

Data Model

Set up a basic model by adding a models.py file:

from app import db
from sqlalchemy.dialects.postgresql import JSON

class Result(db.Model):
    __tablename__ = 'results'

    id = db.Column(db.Integer, primary_key=True)
    url = db.Column(db.String())
    result_all = db.Column(JSON)
    result_no_stop_words = db.Column(JSON)

    def __init__(self, url, result_all, result_no_stop_words):
        self.url = url
        self.result_all = result_all
        self.result_no_stop_words = result_no_stop_words

    def __repr__(self):
        return '<id {}>'.format(self.id)

Here we created a table to store the results of the word counts.

We first import the database connection that we created in our app.py file as well as JSON from SQLAlchemy’s PostgreSQL dialects. JSON columns are fairly new to Postgres and are not available in every database supported by SQLAlchemy so we need to import it specifically.

Next we created a Result() class and assigned it a table name of results. We then set the attributes that we want to store for a result-

  • the id of the result we stored
  • the url that we counted the words from
  • a full list of words that we counted
  • a list of words that we counted minus stop words (more on this later)

We then created an __init__() method that will run the first time we create a new result and, finally, a __repr__() method to represent the object when we query for it.

Local Migration

We are going to use Alembic, which is part of Flask-Migrate, to manage database migrations to update a database’s schema.

Create a new file called manage.py:

import os
from flask.ext.script import Manager
from flask.ext.migrate import Migrate, MigrateCommand

from app import app, db


migrate = Migrate(app, db)
manager = Manager(app)

manager.add_command('db', MigrateCommand)

if __name__ == '__main__':

In order to use Flask-Migrate we imported Manager as well as Migrate and MigrateCommand to our manage.py file. We also imported app and db so we have access to them from within the script.

First, we set our config to get our environment - based on the environment variable - created a migrate instance, with app and db as the arguments, and set up a manager command to initialize a Manager instance for our app. Finally, we added the db command to the manager so that we can run the migrations from the command line.

In order to run the migrations initialize Alembic:

$ python manage.py db init
  Creating directory /flask-by-example/migrations ... done
  Creating directory /flask-by-example/migrations/versions ... done
  Generating /flask-by-example/migrations/alembic.ini ... done
  Generating /flask-by-example/migrations/env.py ... done
  Generating /flask-by-example/migrations/README ... done
  Generating /flask-by-example/migrations/script.py.mako ... done
  Please edit configuration/connection/logging settings in
  '/flask-by-example/migrations/alembic.ini' before proceeding.

After you run the database initialization you will see a new folder called “migrations” in the project. This holds the setup necessary for Alembic to run migrations against the project. Inside of “migrations” you will see that it has a folder called “versions”, which will contain the migration scripts as they are created.

Let’s create our first migration by running the migrate command.

$ python manage.py db migrate
  INFO  [alembic.runtime.migration] Context impl PostgresqlImpl.
  INFO  [alembic.runtime.migration] Will assume transactional DDL.
  INFO  [alembic.autogenerate.compare] Detected added table 'results'
    Generating /flask-by-example/migrations/versions/63dba2060f71_.py
    ... done

Now you’ll notice in your “versions” folder there is a migration file. This file is auto-generated by Alembic based on the model. You could generate (or edit) this file yourself; however, for most cases the auto-generated file will do.

Now we’ll apply the upgrades to the database using the db upgrade command:

$ python manage.py db upgrade
  INFO  [alembic.runtime.migration] Context impl PostgresqlImpl.
  INFO  [alembic.runtime.migration] Will assume transactional DDL.
  INFO  [alembic.runtime.migration] Running upgrade  -> 63dba2060f71, empty message

The database is now ready for us to use in our app:

$ psql
# \c wordcount_dev
You are now connected to database "wordcount_dev" as user "michaelherman".
# \dt

                List of relations
 Schema |      Name       | Type  |     Owner
 public | alembic_version | table | michaelherman
 public | results         | table | michaelherman
(2 rows)

# \d results
                                     Table "public.results"
        Column        |       Type        |                      Modifiers
 id                   | integer           | not null default nextval('results_id_seq'::regclass)
 url                  | character varying |
 result_all           | json              |
 result_no_stop_words | json              |
    "results_pkey" PRIMARY KEY, btree (id)

Remote Migration

Finally, let’s apply the migrations to the databases on Heroku. First, though, we need to add the details of the staging and production databases to the config.py file.

To check if we have a database set up on the staging server run:

$ heroku config --app wordcount-stage
=== wordcount-stage Config Vars
APP_SETTINGS: config.StagingConfig

Make sure to replace wordcount-stage with the name of your staging app.

Since we don’t see a database environment variable, we need to add the Postgres addon to the staging server. To do so, run the following command:

$ heroku addons:create heroku-postgresql:hobby-dev --app wordcount-stage
  Creating postgresql-cubic-86416... done, (free)
  Adding postgresql-cubic-86416 to wordcount-stage... done
  Setting DATABASE_URL and restarting wordcount-stage... done, v8
  Database has been created and is available
   ! This database is empty. If upgrading, you can transfer
   ! data from another database with pg:copy
  Use `heroku addons:docs heroku-postgresql` to view documentation.

hobby-dev is the free tier of the Heroku Postgres addon.

Now when we run heroku config --app wordcount-stage again we should see the connection settings for the database:

=== wordcount-stage Config Vars
APP_SETTINGS: config.StagingConfig
DATABASE_URL: postgres://azrqiefezenfrg:Zti5fjSyeyFgoc-U-yXnPrXHQv@ec2-54-225-151-64.compute-1.amazonaws.com:5432/d2kio2ubc804p7

Next we need to commit the changes that you’ve made to git and push to your staging server:

$ git push stage master

Run the migrations that we created to migrate our staging database by using the heroku run command:

$ heroku run python manage.py db upgrade --app wordcount-stage
  Running python manage.py db upgrade on wordcount-stage... up, run.5677
  INFO  [alembic.runtime.migration] Context impl PostgresqlImpl.
  INFO  [alembic.runtime.migration] Will assume transactional DDL.
  INFO  [alembic.runtime.migration] Running upgrade  -> 63dba2060f71, empty message

Notice how we only ran the upgrade, not the init or migrate commands like before. We already have our migration file set up and ready to go; we just need to apply it against the Heroku database.

Let’s now do the same for production.

  1. Set up a database for your production app on Heroku, just like you did for staging: heroku addons:create heroku-postgresql:hobby-dev --app wordcount-pro
  2. Push your changes to your production site: git push pro master Notice how you don’t have to make any changes to the config file - it’s setting the database based on the newly created DATABASE_URL environment variable.
  3. Apply the migrations: heroku run python manage.py db upgrade --app wordcount-pro

Now both our staging and production sites have their databases set up and are migrated - and ready to go!

When you apply a new migration to the production database, there could be down time. If this is an issue, you can set up database replication by adding a “follower” (commonly known as a slave) database. For more on this, check out the official Heroku documentation.


That’s it for part 2. If you’d like to dig deeper into Flask, check out our accompanying video series:

In Part 3 we’re going to build the word counting functionality and have it sent to a task queue to deal with the longer running word count processing.

See you next time. Cheers!

This is a collaboration piece between Cam Linke, co-founder of Startup Edmonton, and the folks at Real Python.

🐍 Python Tricks 💌

Get a short & sweet Python Trick delivered to your inbox every couple of days. No spam ever. Unsubscribe any time. Curated by the Real Python team.

Python Tricks Dictionary Merge

Support Free Python Education...

Real Python brings you free, book-quality tutorials and in-depth articles about Python programming every single week. Everyone on our editorial team gets paid for their work—from our authors, to our editors and proof readers, our designers, community team, and web developers.

We do not believe in spammy ad banners from the big advertising networks. We don’t secretly mine Bitcoin in your browser to cover our hosting costs… And unlike many other publications, we haven’t put up a paywall—we want to keep our educational content as open as we can.

Help make sustainable programming journalism and education a reality by supporting us with a small monthly contribution. For as little as $1, you can support Real Python—and it only takes a minute. Thank you.

VISA Discover American Express Maestro PayPal
Support Real Python →

What Do You Think?

Real Python Comment Policy: The most useful comments are those written with the goal of learning from or helping out other readers—after reading the whole article and all the earlier comments. Complaints and insults generally won’t make the cut here.

Boost Your Python Skills

Master Python 3 and write more Pythonic code with our in-depth books and video courses:

Get Python Books & Courses »

Keep Reading