It's a Flask CLI extension that helps you migrate, drop, create and seed your SQL database.
After installing it you'll gain access to a flask db
command that will give
you a few database management commands to use.
For the migrations it uses Alembic. Anything you can do with Alembic can be done with this CLI extension. If you're wondering why you might want to use this tool instead of Alembic directly or Flask-Migrate, check out this FAQ item.
db
command is availabledb
command and its sub-commandsHere's a run down on how this extension works and how you can add it to your project:
If you prefer reading instead of video this README file covers installing, configuring and using this extension too.
pip3 install Flask-DB
That's it!
There's no need to even import or initialize anything in your Flask app because it's just a CLI command that gets added to your Flask app.
But if you're curious, a complete example Flask app can be found in the tests/ directory.
db
command is available and everything worksYou'll want to at least set the FLASK_APP
and PYTHONPATH
environment
variables:
# Replace `hello.app` with your app's name.
export FLASK_APP=hello.app
# Due to how Alembic works you'll want to set this to be able to migrate
# your database. It would be expected that you run your migrate commands from
# the root of your project (more on this migrate command later).
export PYTHONPATH="."
export FLASK_DEBUG=true
Then run the flask
binary to see its help menu:
Usage: flask [OPTIONS] COMMAND [ARGS]...
...
Commands:
db Migrate and manage your SQL database.
If all went as planned you should see the new db
command added to your
list of commands.
db
command and its sub-commandsRunning flask db --help
will produce this help menu:
Usage: flask db [OPTIONS] COMMAND [ARGS]...
Migrate and manage your SQL database.
Options:
--help Show this message and exit.
Commands:
init Generate Alembic config files and seeds.py.
reset Drop, create and seed your database (careful in production).
seed Seed the database with your custom records.
migrate Wrap the alembic CLI tool (defaults to running upgrade head).
init
If you've ever used Alembic before, you'll be familiar with alembic init
. If
you've never used it before it creates an alembic.ini
file in the root of
your project along with a few other Alembic files in a yourappname/migrations
directory by default.
You can treat flask db init
as a drop in replacement for alembic init
.
flask db init
does something similar. The 4 main differences are:
It defaults to db/
instead of yourappname/migrations
, but you can modify
this path by running flask db init any/path/you/want
.
It will create the same Alembic config files but it modifies them to be a
bit more generic and portable when it comes to finding your
SQLALCHEMY_DATABASE_URI
.
It also configures Alembic to support auto generating migrations in case you
want to use revision --autogenerate
as a starting point for your
migrations (always review them afterwards!).
It creates a seeds.py
file in the same directory you initialized things to
(more on this soon).
reset
Creates your database if it needs to be created along with doing a
db.drop_all()
and db.create_all()
. That is going to purge all of your
existing data and create any tables based on whatever SQLAlchemy models you
have.
It also automatically calls flask db seed
for you (more on seeding soon).
Options:
--with-testdb Create a test DB in addition to your main DB?
If you run flask db reset --with-testdb
then a second database will also be
created based on whatever your database name is from your
SQLALCHEMY_DATABASE_URI
along with appending _test
to its name.
For example, if you had a db named hello
and you used --with-testdb
then
you would end up with both a hello
and hello_test
database. It's very
useful to have a dedicated test database so you don't clobber your dev data.
That depends on the state of your project.
This is something you'll be running all the time in development as you change your database models.
It's also something you'd typically run once in production the first time you deploy your app. This will handle creating your database and syncing all of your models as tables in your database.
Purging all of your data isn't an option anymore. If you want to make database
changes you should be creating database migrations with flask db migrate
,
which is really running Alembic
commands behind the scenes. That is the official migration tool created by the
same folks who made SQLAlchemy.
We'll go over the migrate
command shortly.
seed
Seeds your database with initial records of your choosing.
When you set up your app for the first time in production chances are you'll want certain things to be created in your database. At the very least probably an initial admin user.
This command will read in and execute a seeds.py
file that exists in the
directory that you initialized with flask db init
. By default that will be
db/seeds.py
.
If you supplied a custom init path you must change your seeds path. You can do
that by setting FLASK_DB_SEEDS_PATH
in your app's config. For example if you
ran flask db init yourappname/migrations
then you'd set FLASK_DB_SEEDS_PATH = "yourappname/migrations/seeds.py"
.
As for the seeds file, you have full control over what you want to do. You can add whatever records that make sense for your app or keep the file empty to not seed anything. It's simply a file that's going to get called and execute any code you've placed into it.
It's a good idea to make your seeds file idempotent. Meaning, if you run it 1 time or 100 times the end result should be the same. The example in the seeds file is idempotent because it first checks to see if the user exists before adding it.
If the user exists, it skips trying to create the user. Without this check then you would end up getting a database uniqueness constraint error on the 2nd run. That's because the example test app added a unique index on the username.
migrate
The flask db migrate
command is an alias to the alembic
command.
Here's a few examples:
flask db migrate revision -m "hi" -> alembic revision -m "hi"
flask db migrate upgrade head -> alembic upgrade head
flask db migrate -> alembic upgrade head (convenience shortcut)
flask db migrate -h -> alembic -h
Every possible thing you can run with alembic
can now be run with flask db migrate
. That means you can follow Alembic's tutorials and documentation
exactly.
You can still use alembic commands directly if you prefer. I just liked the
idea of having all db related commands under the same namespace and the flask db migrate
shortcut is handy.
You'll also be happy to hear that this tool doesn't hard code all of Alembic's
commands, options and arguments with hundreds of lines of code. What that means
is, as Alembic's CLI API changes in the future this flask db migrate
command
will continue to work with all future versions.
All it does is forward everything you pass in, similar to $@
in Bash.
That's also why you need to run -h
to get Alembic's help menu instead of
--help
because that will ultimately call the internal help menu for flask db migrate
.
Here's the breakdown of these tools:
Disclaimer: I'm not here to crap on the work of others, I'm simply giving my opinion on why I chose to create Flask-DB and don't use the Alembic CLI directly, Flask-Migrate or Flask-Alembic.
The official database migration tool for SQLAlchemy written by the same team who made SQLAlchemy.
Alembic generates a bunch of config files and often times you'll want to go in
and edit them to be more dynamic for finding your SQLALCHEMY_DATABASE_URI
as
well as adding a few useful opinions that you'll want in 99.999% of apps.
It got a little tedious making these adjustments in every app.
Besides having config files, Alembic also includes an alembic
CLI tool to
interact with your database such as generating migration files, executing
migrations and more. This tool is fantastic.
Flask-Migrate, Flask-Alembic and Flask-DB all use Alembic behind the scenes, but they are implemented in much different ways.
At the time of writing this FAQ item (November 2020) the approach this library takes it to import the Alembic Python module and then build a custom CLI tool with various commands, options and arguments that lets you access functionality that Alembic already provides.
In my opinion this is very error prone because Alembic already has a battle
hardened alembic
CLI tool that comes part of the Alembic Python package. You
get this out of the box when you pip3 install alembic
. Flask-Migrate is error
prone because it has hundreds upon hundreds of lines of code re-creating the
same alembic
CLI tool that already exists.
Also it's not future proof because if Alembic decides to change its API or add
new features now you'll need to wait for Flask-Migrate to update all of its
code and push a new release instead of just grabbing the new alembic
package.
For example, here's a tiny snippet of code from Flask-Migrate's migrate
command:
@db.command()
@click.option('-d', '--directory', default=None,
help=('Migration script directory (default is "migrations")'))
@click.option('-m', '--message', default=None, help='Revision message')
@click.option('--sql', is_flag=True,
help=('Don\'t emit SQL to database - dump to standard output '
'instead'))
@click.option('--head', default='head',
help=('Specify head revision or <branchname>@head to base new '
'revision on'))
@click.option('--splice', is_flag=True,
help=('Allow a non-head revision as the "head" to splice onto'))
@click.option('--branch-label', default=None,
help=('Specify a branch label to apply to the new revision'))
@click.option('--version-path', default=None,
help=('Specify specific path from config for version file'))
@click.option('--rev-id', default=None,
help=('Specify a hardcoded revision id instead of generating '
'one'))
@click.option('-x', '--x-arg', multiple=True,
help='Additional arguments consumed by custom env.py scripts')
@with_appcontext
def migrate(directory, message, sql, head, splice, branch_label, version_path,
rev_id, x_arg):
"""Autogenerate a new revision file (Alias for 'revision --autogenerate')"""
_migrate(directory, message, sql, head, splice, branch_label, version_path,
rev_id, x_arg)
Needless to say when you do this for a dozen commands with many dozens of flags it's easy for errors to slip by.
Another thing is Flask-Migrate's flask db migrate
command defaults to using
auto-generated migrations. This feature is useful for speeding up the process
of creating migration files but Alembic doesn't detect everything which can be
very confusing if you're not aware of Alembic's limitations with
auto-generate.
There's also no help for doing things like resetting and seeding your database. It's a tool exclusively designed for running database migrations with Alembic.
This tool internally imports the Alembic Python library and creates its own CLI
around that. It's similar to Flask-Migrate in that regard but not of all of the
commands are API compatible with the alembic
CLI.
For example, here's a snippet from Flask-Alembic's revision
command:
@manager.option("message")
@manager.option("--empty", action="store_true", help="Create empty script.")
@manager.option(
"-b", "--branch", default="default", help="Use this independent branch name."
)
@manager.option(
"-p",
"--parent",
default="head",
type=str.split,
help="Parent revision(s) of this revision.",
)
@manager.option("--splice", action="store_true", help="Allow non-head parent revision.")
@manager.option(
"-d", "--depend", type=str.split, help="Revision(s) this revision depends on."
)
@manager.option(
"-l", "--label", type=str.split, help="Label(s) to apply to the revision."
)
@manager.option("--path", help="Where to store the revision.")
def revision(message, empty, branch, parent, splice, depend, label, path):
"""Create new migration."""
base.revision(message, empty, branch, parent, splice, depend, label, path)
It's missing a few flags that alembic revision --help
would provide you.
Personally I'd rather use the official alembic
CLI since it already exists
and it's very well tested and developed by the Alembic team.
Also, it doesn't handle alembic init
or have its own form of init
so you're
left having to generate and modify the default Alembic config files in every
project.
Like Flask-Migrate it also doesn't help with resetting and seeding your database.
Flask-DB is more than a database migration tool that wraps Alembic. It also includes being able to reset and seed your database.
Unlike using Alembic directly for the init
command it modernizes and applies
a few opinions to the default Alembic configuration so that you can usually use
these files as is in your projects without modification. If you do need to
modify them, that's ok. They are generated in your project which you can edit.
As for migrations, it aliases the alembic
CLI but it does it with about 5
lines of code for everything rather than hundreds of lines of code. The lines
of code aren't that important, it's mainly being less error prone and more
future proof. It's taking advantage of the well tested alembic
CLI tool that
Alembic gave us.
That means if a future version of Alembic comes out that adds new features it
will automatically work with Flask-DB without having to update Flask-DB. All
you would do is upgrade your alembic
package version and you're done.
Since about 2015 I used to create my own db
CLI command in each project which
handled resetting and seeding the database. Then I used the alembic
CLI
directly. Anyone who has taken my Build a SAAS App with Flask
course
is probably used to that.
Flask-DB was created as an improvement to that work flow. Now it's as easy as pip installing this tool and you're good to go with ready to go configs, a consistent Alembic CLI API and the added reset + seed functionality.
There's 2 ways to go about this and it's up to you on which one to pick.
In both cases if you're using Flask-Migrate you'll want to pip3 uninstall Flask-Migrate
since both Flask-DB and Flask-Migrate add a db
command to
Flask.
By default Alembic (and therefore Flask-Migrate) will put a migrations/
directory inside of your application.
You can still keep your files there with Flask-DB too. You'll just need to
configure FLASK_DB_SEEDS_PATH
to be yourappname/migrations/seeds.py
. You'll
also want to create that file manually (keeping it empty for now is ok).
That's pretty much all you need to do.
If you're using Flask-Migrate, Flask-DB has more up to date Alembic config
files so you may still want to initialize a new directory with flask db init
.
Then you can pick and choose what you want to keep from your existing Alembic
configs and what's been generated by the Flask-DB's init command.
versions/
into it (recommended)The other option would be to run flask db init
and let it create a db/
directory in the root of your project.
Chances are you already have an alembic.ini
file. The init command will
recognize that and create an alembic.ini.new
file to avoid clobbering your
existing file. Then you can delete your old alembic.init
file and mv alembic.ini.new alembic.ini
to use the new one.
From here you can take your existing migration files in your old versions/
directory and move them into db/versions/
. At this point you can delete your
old yourappname/migrations/
directory as long as you haven't done any drastic
customizations to your alembic related config files.
If you've done a bunch of custom configurations that's ok, you can manually merge in your changes. You're free to edit these files however you see fit.
flask db init
created?Absolutely. The init command is there to set up the initial files that Alembic
expects to be created, along with adding things like the seeds.py
file too.
The Alembic files will likely be good to go for you without having to modify
anything but you can change them however you see fit. One thing worth
mentioning is if your app factory function isn't called create_app
you will
want to change the db/env.py
file's import near the top to use your custom
factory function's name instead.
You also have free reign to add whatever you want to the seeds.py
file. By
default it generates a couple of comments to help show how you can use it
generate an initial user in your database.
Yes! Your alembic.ini
along with your entire db/
directory (default init
directory) including the versions/*
files should all be commit to version
control.
This way you can develop and test your migrations locally on your dev box, make sure everything works then commit and push your code. At that point you can run your migrations in CI and ultimately in production with confidence that it all works.
You might have seen this error: ModuleNotFoundError: No module named 'yourappname'
Chances are you haven't set your PYTHONPATH="."
env variable. If you're using
Docker you can set this in your Dockerfile and you'll be good to go. Here's a
working example from my Build a SAAS App with
Flask
repo.
If you're not using Docker please make sure that your PYTHONPATH
environment
variable is available on your shell by doing any of the things below:
export PYTHONPATH="."
in your terminal and now it'll be set until you close your terminal.env
file and then run source .env
and it'll be set until you close your terminalPYTHONPATH="." flask db migrate
instead of flask db migrate
In all cases it would be expected that you run the flask db migrate
command
from the root of your project. That's the same directory as where your
alembic.ini
file is located.
If you're interested in learning Flask I have a 20+ hour video course called Build a SAAS App with Flask. It's a course where we build a real world SAAS app. Everything about the course and demo videos of what we build is on the site linked above.