The below article will cover the intricacies of setting up databases and heroku in respect to a flask app. This is more like a memo and will have out of sequence instructions or solutions to errors so read thoroughly.
You'll need the packages
alembic
Flask-Migrate
Flask-Script
Flask-SQLAlchemy
psycopg2
Now. Information. alembic will install on its own if you install the 3 packages after it. You'll need psycopg2 ONLY IF you plan on deploying to heroku and using their database.
You'll need at least 3 files to use migration and other features of flask-migrate
- models.py
from manage import db,app
class User(db.Model):
__tablename__ = 'users'
id = db.Column(db.Integer, primary_key=True)
nickname = db.Column(db.String(64), index=True, unique=True)
email = db.Column(db.String(120), index=True, unique=True)
def __repr__(self):
return '<User %r>' % (self.nickname)- app.py
from flask import Flask
from flask_sqlalchemy import SQLAlchemy
app = Flask(__name__)
app.config['SQLALCHEMY_DATABASE_URI'] = 'sqlite:///app.db'
db = SQLAlchemy(app)
from models import User
@app.route('/add/')
def webhook():
name = "ram"
email = "[email protected]"
u = User(id = id, nickname = name, email = email)
print("user created", u)
db.session.add(u)
db.session.commit()
return "user created"
@app.route('/delete/')
def delete():
u = User.query.get(i)
db.session.delete(u)
db.session.commit()
return "user deleted"
if __name__ == '__main__':
app.run()- manage.py
from flask_script import Manager
from flask_migrate import Migrate, MigrateCommand
from app import app, db
migrate = Migrate(app, db)
manager = Manager(app)
manager.add_command('db', MigrateCommand)
if __name__ == '__main__':
manager.run()NOTE: Where we are importing what. In app.py we are importing User after we declare db since that is used in models, from where User is imported.
app.config['SQLALCHEMY_DATABASE_URI'] = 'sqlite:///app.db'This sets up the parameter SQLALCHEMY_DATABASE_URI and tells it to point to a sqlite db. This can also be declared in a config.py file and then in app.py
app.config.from_object('config')Initialize the database and create the migrations folder.
Check [https://stackoverflow.com/questions/17174636/can-alembic-autogenerate-column-alterations]
The above one worked. But very importantly also check [https://stackoverflow.com/questions/33944436/flask-migrate-and-changing-column-type]
Heroku has a free hobby development database that we can use happily. Let's see how we can set up the database, link it to our app and do the necessary stuff keeping most of our hair intact.
psycopg2 required from here on.
Go to this [https://devcenter.heroku.com/articles/getting-started-with-python#provision-a-database]
Also check this out [https://blog.miguelgrinberg.com/post/the-flask-mega-tutorial-part-xviii-deployment-on-the-heroku-cloud]
Do note the above link is outdated in regards to the command given for adding a database. It is now hobby-dev not dev. Also we don't need to promote now. It will already be promoted to DATABASE_URL
Run the below command on terminal.
heroku addons:add heroku-postgresql:hobby-dev
2.1. Now we'll face some problem regarding migrating. What we'll do is below in order to bypass the problems.
heroku run python
>> import os
>> os.environ.get('DATABASE_URL')
It will give the link to the development database on Heroku. (It will be something like postgres://bllahblah)
2.2. Replace the value of DATABASE_URI in config.py with this value.
**2.3.**Then do apply your changes to Heroku database.
python manage.py db init
python manage.py db migrate
python manage.py db upgrade
Then push your code as usual. Better push your migrations on Heroku too. Don't leave your migrations folder out of git commit. But keep local sqlite db outside.
web: python manage.py runserver --host 0.0.0.0 --port ${PORT}
init: python manage.py db init
migrate: python manage.py db migrate
upgrade: python manage.py db upgrade
This is how the Procfile MUST look like. We don't care much about the last 3. The first one is what is very very important. Notice in app.py we a calling app.run() without any parameters (host and port). It should be that way because we'll be delivering the host and port from here in Procfile. DON'T PUT THE PARAMETERS INSIDE app.py EVEN IF YOU KNOW THOSE VALUES WELL. If it is not like this, the app won't be able to launch after each push and deploy.
Btw, we call Procfile commands using Heroku CLI from terminal this way
heroku run init
heroku run migrate
heroku run web
This also can be done without calling the aliases
heroku run python manage.py db init
To go into the heroku shell
heroku run bash
ls -ltr shows the directory structure, exit closes the bash terminal.
[https://stackoverflow.com/questions/32311366/alembic-util-command-error-cant-find-identifier]
Do what is said in the SO thread. You'll need to install PostgreSQL first.
Go to this link [https://devcenter.heroku.com/articles/heroku-postgresql#set-up-postgres-on-windows]
Be sure to close everything on mother earth after updating th PATH variable. If everything was correct, you'll be able to call psql on terminal. Then do
heroku pg:psql
You'll then go into the db shell of the heroku database. Now write the commands as instructed in the SO answer and do the rest.
NOTE: If psql asks for password, the password you set up during install won't work. Check the SO thread [https://stackoverflow.com/questions/12562928/psql-exe-password-authentication-failed-in-windows]
Here is another solution I've found reddit to run automaticaly a post-compilation script :
binfolder inside the app root folderpost_compileflask db init && flask db migrate && flask db upgrade