Flask-Script

Warning

While the maintainers are willing to merge PR’s, they are not actively developing features. As of Flask 0.11, Flask includes a built-in CLI tool, and that may fit your needs better.

+1:2

The Flask-Script extension provides support for writing external scripts in Flask. This includes running a development server, a customised Python shell, scripts to set up your database, cronjobs, and other command-line tasks that belong outside the web application itself.

Flask-Script works in a similar way to Flask itself. You define and add commands that can be called from the command line to a Manager instance:

# manage.py

from flask_script import Manager

from myapp import app

manager = Manager(app)

@manager.command
def hello():
    print "hello"

if __name__ == "__main__":
    manager.run()

Once you define your script commands, you can then run them on the command line:

python manage.py hello
> hello

Source code and issue tracking at GitHub.

Installing Flask-Script

Install with pip and easy_install:

pip install Flask-Script

or download the latest version from version control:

git clone https://github.com/smurfix/flask-script.git
cd flask-script
python setup.py develop

If you are using virtualenv, it is assumed that you are installing Flask-Script in the same virtualenv as your Flask application(s).

Creating and running commands

The first step is to create a Python module to run your script commands in. You can call it anything you like, for our examples we’ll call it manage.py.

You don’t have to place all your commands in the same file; for example, in a larger project with lots of commands you might want to split them into a number of files with related commands.

In your manage.py file you have to create a Manager instance. The Manager class keeps track of all the commands and handles how they are called from the command line:

from flask_script import Manager

app = Flask(__name__)
# configure your app

manager = Manager(app)

if __name__ == "__main__":
    manager.run()

Calling manager.run() prepares your Manager instance to receive input from the command line.

The Manager requires a single argument, a Flask instance. This may also be a function or other callable that returns a Flask instance instead, if you want to use a factory pattern.

The next step is to create and add your commands. There are three methods for creating commands:

  • subclassing the Command class
  • using the @command decorator
  • using the @option decorator

To take a very simple example, we want to create a hello command that just prints out “hello world”. It doesn’t take any arguments so is very straightforward:

from flask_script import Command

class Hello(Command):
    "prints hello world"

    def run(self):
        print "hello world"

Now the command needs to be added to our Manager instance, like the one created above:

manager.add_command('hello', Hello())

This of course needs to be called before manager.run. Now in our command line:

python manage.py hello
> hello world

You can also pass the Command instance in a dict to manager.run():

manager.run({'hello' : Hello()})

The Command class must define a run method. The positional and optional arguments depend on the command-line arguments you pass to the Command (see below).

To get a list of available commands and their descriptions, just run with no command:

python manage.py

To get help text for a particular command:

python manage.py runserver -?

This will print usage plus the docstring of the Command.

This first method is probably the most flexible, but it’s also the most verbose. For simpler commands you can use the @command decorator, which belongs to the Manager instance:

@manager.command
def hello():
    "Just say hello"
    print "hello"

Commands created this way are run in exactly the same way as those created with the Command class:

python manage.py hello
> hello

As with the Command class, the docstring you use for the function will appear when you run with the -? or --help option:

python manage.py -?
> Just say hello

Finally, the @option decorator, again belonging to Manager can be used when you want more sophisticated control over your commands:

@manager.option('-n', '--name', help='Your name')
def hello(name):
    print "hello", name

The @option decorator is explained in more detail below.

New in version 2.0

Help was previously available with --help and -h. This had a couple of less-than-ideal consequences, among them the inability to use -h as a shortcut for --host or similar options.

New in version 2.0.2

If you want to restore the original meaning of -h, set your manager’s help_args attribute to a list of argument strings you want to be considered helpful:

manager = Manager()
manager.help_args = ('-h', '-?', '--help')

You can override this list in sub-commands and -managers:

def talker(host='localhost'):
    pass
ccmd = ConnectCmd(talker)
ccmd.help_args = ('-?', '--help')
manager.add_command("connect", ccmd)
manager.run()

so that manager -h prints help, while manager connect -h fubar.example.com connects to a remote host.

Adding arguments to commands

Most commands take a number of named or positional arguments that you pass in the command line.

Taking the above examples, rather than just print “hello world” we would like to be able to print some arbitrary name, like this:

python manage.py hello --name=Joe
hello Joe

or alternatively:

python manage.py hello -n Joe

To facilitate this you use the option_list attribute of the Command class:

from flask_script import Command, Manager, Option

class Hello(Command):

    option_list = (
        Option('--name', '-n', dest='name'),
    )

    def run(self, name):
        print "hello %s" % name

Positional and optional arguments are stored as Option instances - see the API below for details.

Alternatively, you can define a get_options method for your Command class. This is useful if you want to be able to return options at runtime based on for example per-instance attributes:

class Hello(Command):

    def __init__(self, default_name='Joe'):
        self.default_name=default_name

    def get_options(self):
        return [
            Option('-n', '--name', dest='name', default=self.default_name),
        ]

    def run(self, name):
        print "hello",  name

If you are using the @command decorator, it’s much easier - the options are extracted automatically from your function arguments. This is an example of a positional argument:

@manager.command
def hello(name):
    print "hello", name

You then invoke this on the command line like so:

> python manage.py hello Joe
hello Joe

Or you can do optional arguments:

@manager.command
def hello(name="Fred")
    print "hello", name

These can be called like so:

> python manage.py hello --name=Joe
hello Joe

alternatively:

> python manage.py hello -n Joe
hello Joe

The short form -n is formed from the first letter of the argument, so “name” > “-n”. Therefore it’s a good idea for your optional argument variable names to begin with different letters.

New in version 2.0

Note also that if your optional argument is a boolean, for example:

@manager.command
def verify(verified=False):
    """
    Checks if verified
    """
    print "VERIFIED?", "YES" if verified else "NO"

You can just call it like this:

> python manage.py verify
VERIFIED? NO

> python manage.py verify -v
VERIFIED? YES

> python manage.py verify --verified
VERIFIED? YES

The @command decorator is fine for simple operations, but often you need the flexibility. For more sophisticated options it’s better to use the @option decorator:

@manager.option('-n', '--name', dest='name', default='joe')
def hello(name):
    print "hello", name

You can add as many options as you want:

@manager.option('-n', '--name', dest='name', default='joe')
@manager.option('-u', '--url', dest='url', default=None)
def hello(name, url):
    if url is None:
        print "hello", name
    else:
        print "hello", name, "from", url

This can be called like so:

> python manage.py hello -n Joe -u reddit.com
hello Joe from reddit.com

or alternatively:

> python manage.py hello --name=Joe --url=reddit.com
hello Joe from reddit.com

Adding options to the manager

Options can also be passed to the Manager instance. This is allows you to set up options that are passed to the application rather than a single command. For example, you might want to have a flag to set the configuration file for your application. Suppose you create your application with a factory function:

def create_app(config=None):

    app = Flask(__name__)
    if config is not None:
        app.config.from_pyfile(config)
    # configure your app...
    return app

You want to be able to define the config argument on the command line - for example, if you have a command to set up your database, you most certainly want to use different configuration files for production and development.

In order to pass that config argument, use the add_option() method of your Manager instance. It takes the same arguments as Option:

manager.add_option('-c', '--config', dest='config', required=False)

As with any other Flask-Script configuration you can call this anywhere in your script module, but it must be called before your manager.run() call.

Suppose you have this command:

@manager.command
def hello(name):
    uppercase = app.config.get('USE_UPPERCASE', False)
    if uppercase:
        name = name.upper()
    print "hello", name

You can now run the following:

> python manage.py -c dev.cfg hello joe
hello JOE

Assuming the USE_UPPERCASE setting is True in your dev.cfg file.

Notice also that the “config” option is not passed to the command. In fact, this usage:

> python manage.py hello joe -c dev.cfg

will show an error message because the -c option does not belong to the hello command.

You can attach same-named options to different levels; this allows you to add an option to your app setup code without checking whether it conflicts with a command:

@manager.option('-n', '--name', dest='name', default='joe')
@manager.option('-c', '--clue', dest='clue', default='clue')
def hello(name, clue):
    uppercase = app.config.get('USE_UPPERCASE', False)
    if uppercase:
        name = name.upper()
        clue = clue.upper()
    print "hello {0}, get a {1}!".format(name, clue)

> python manage.py -c dev.cfg hello -c cookie -n frank
hello FRANK, get a COOKIE!

Note that the destination variables (command arguments, corresponding to dest values) must still be different; this is a limitation of Python’s argument parser.

In order for manager options to work you must pass a factory function, rather than a Flask instance, to your Manager constructor. A simple but complete example is available in this gist.

New in version 2.0

Before version 2, options and command names could be interspersed freely. The author decided to discontinue this practice for a number of reasons; the problem with the most impact was that it was not possible to do:

> python manage.py connect -d DEST
> python manage.py import -d DIR

as these options collided.

Getting user input

Flask-Script comes with a set of helper functions for grabbing user input from the command line. For example:

from flask_script import Manager, prompt_bool

from myapp import app
from myapp.models import db

manager = Manager(app)

@manager.command
def dropdb():
    if prompt_bool(
        "Are you sure you want to lose all your data"):
        db.drop_all()

It then runs like this:

> python manage.py dropdb
Are you sure you want to lose all your data ? [N]

See the API below for details on the various prompt functions.

Default commands

runserver

Flask-Script has a couple of ready commands you can add and customise: Server and Shell.

The Server command runs the Flask development server.:

from flask_script import Server, Manager
from myapp import create_app

manager = Manager(create_app)
manager.add_command("runserver", Server())

if __name__ == "__main__":
    manager.run()

and then run the command:

python manage.py runserver

The Server command has a number of command-line arguments - run python manage.py runserver -? for details on these. You can redefine the defaults in the constructor:

server = Server(host="0.0.0.0", port=9000)

Needless to say the development server is not intended for production use.

New in version 2.0.5

The most common use-case for runserver is to run a debug server for investigating problems. Therefore the default, if it is not set in the configuration file, is to enable debugging and auto-reloading.

Unfortunately, Flask currently (as of May 2014) defaults to set the DEBUG configuration parameter to False. Until this is changed, you can safely add DEBUG=None to your Flask configuration. Flask-Script’s runserver will then turn on debugging, but everything else will treat it as being turned off.

To prevent misunderstandings – after all, debug mode is a serious security hole –, a warning is printed when Flask-Script treats a None default value as if it were set to True. You can turn on debugging explicitly to get rid of this warning.

shell

The Shell command starts a Python shell. You can pass in a make_context argument, which must be a callable returning a dict. By default, this is just a dict returning the your Flask application instance:

from flask_script import Shell, Manager

from myapp import app
from myapp import models
from myapp.models import db

def _make_context():
    return dict(app=app, db=db, models=models)

manager = Manager(create_app)
manager.add_command("shell", Shell(make_context=_make_context))

This is handy if you want to include a bunch of defaults in your shell to save typing lots of import statements.

The Shell command will use IPython if it is installed, otherwise it defaults to the standard Python shell. You can disable this behaviour in two ways: by passing the use_ipython argument to the Shell constructor, or passing the flag --no-ipython in the command line:

shell = Shell(use_ipython=False)

There is also a shell decorator which you can use with a context function:

@manager.shell
def make_shell_context():
    return dict(app=app, db=db, models=models)

This enables a shell command with the defaults enabled:

> python manage.py shell

The default commands shell and runserver are included by default, with the default options for these commands. If you wish to replace them with different commands simply override with add_command() or the decorators. If you pass with_default_commands=False to the Manager constructor these commands will not be loaded:

manager = Manager(app, with_default_commands=False)

Sub-Managers

A Sub-Manager is an instance of Manager added as a command to another Manager

To create a submanager:

def sub_opts(app, **kwargs):
    pass
sub_manager = Manager(sub_opts)

manager = Manager(self.app)
manager.add_command("sub_manager", sub_manager)

If you attach options to the sub_manager, the sub_opts procedure will receive their values. Your application is passed in app for convenience.

If sub_opts returns a value other than None, this value will replace the app value that’s passed on. This way, you can implement a sub-manager which replaces the whole app. One use case is to create a separate administrative application for improved security:

def gen_admin(app, **kwargs):
    from myweb.admin import MyAdminApp
    ## easiest but possibly incomplete way to copy your settings
    return MyAdminApp(config=app.config, **kwargs)
sub_manager = Manager(gen_admin)

manager = Manager(MyApp)
manager.add_command("admin", sub_manager)

> python manage.py runserver
[ starts your normal server ]
> python manage.py admin runserver
[ starts an administrative server ]

You can cascade sub-managers, i.e. add one sub-manager to another.

A sub-manager does not get default commands added to itself (by default)

New in version 0.5.0.

Note to extension developers

Extension developers can easily create convenient sub-manager instance within their extensions to make it easy for a user to consume all the available commands of an extension.

Here is an example how a database extension could provide (ex. database.py):

manager = Manager(usage="Perform database operations")

@manager.command
def drop():
    "Drops database tables"
    if prompt_bool("Are you sure you want to lose all your data"):
        db.drop_all()


@manager.command
def create(default_data=True, sample_data=False):
    "Creates database tables from sqlalchemy models"
    db.create_all()
    populate(default_data, sample_data)


@manager.command
def recreate(default_data=True, sample_data=False):
    "Recreates database tables (same as issuing 'drop' and then 'create')"
    drop()
    create(default_data, sample_data)


@manager.command
def populate(default_data=False, sample_data=False):
    "Populate database with default data"
    from fixtures import dbfixture

    if default_data:
        from fixtures.default_data import all
        default_data = dbfixture.data(*all)
        default_data.setup()

    if sample_data:
        from fixtures.sample_data import all
        sample_data = dbfixture.data(*all)
        sample_data.setup()

Then the user can register the sub-manager to their primary Manager (within manage.py):

manager = Manager(app)

from flask_database import manager as database_manager
manager.add_command("database", database_manager)

The commands will then be available:

> python manage.py database

 Please provide a command:

 Perform database operations
  create    Creates database tables from sqlalchemy models
  drop      Drops database tables
  populate  Populate database with default data
  recreate  Recreates database tables (same as issuing 'drop' and then 'create')

Error handling

Users do not like to see stack traces, but developers want them for bug reports.

Therefore, flask_script.commands provides an InvalidCommand error class which is not supposed to print a stack trace when reported.

In your command handler:

from flask_script.commands import InvalidCommand

[… if some command verification fails …]
class MyCommand(Command):
    def run(self, foo=None, bar=None):
        if foo and bar:
            raise InvalidCommand("Options foo and bar are incompatible")

In your main loop:

try:
    MyManager().run()
except InvalidCommand as err:
    print(err, file=sys.stderr)
    sys.exit(1)

This way, you maintain interoperability if some plug-in code supplies Flask-Script hooks you’d like to use, or vice versa.

Accessing local proxies

The Manager runs the command inside a Flask test context. This means that you can access request-local proxies where appropriate, such as current_app, which may be used by extensions.

API

class flask_script.Manager(app=None, with_default_commands=None, usage=None, help=None, description=None, disable_argcomplete=False)

Controller class for handling a set of commands.

Typical usage:

class Print(Command):

    def run(self):
        print "hello"

app = Flask(__name__)

manager = Manager(app)
manager.add_command("print", Print())

if __name__ == "__main__":
    manager.run()

On command line:

python manage.py print
> hello
Parameters:
  • app – Flask instance, or callable returning a Flask instance.
  • with_default_commands – load commands runserver and shell by default.
  • disable_argcomplete – disable automatic loading of argcomplete.
add_command(*args, **kwargs)

Adds command to registry.

Parameters:
  • command – Command instance
  • name – Name of the command (optional)
  • namespace – Namespace of the command (optional; pass as kwarg)
add_option(*args, **kwargs)

Adds a global option. This is useful if you want to set variables applying to the application setup, rather than individual commands.

For this to work, the manager must be initialized with a factory function rather than a Flask instance. Otherwise any options you set will be ignored.

The arguments are then passed to your function, e.g.:

def create_my_app(config=None):
    app = Flask(__name__)
    if config:
        app.config.from_pyfile(config)

    return app

manager = Manager(create_my_app)
manager.add_option("-c", "--config", dest="config", required=False)
@manager.command
def mycommand(app):
    app.do_something()

and are invoked like this:

> python manage.py -c dev.cfg mycommand

Any manager options passed on the command line will not be passed to the command.

Arguments for this function are the same as for the Option class.

command(func)

Decorator to add a command function to the registry.

Parameters:func – command function.Arguments depend on the options.
option(*args, **kwargs)

Decorator to add an option to a function. Automatically registers the function - do not use together with @command. You can add as many @option calls as you like, for example:

@option('-n', '--name', dest='name')
@option('-u', '--url', dest='url')
def hello(name, url):
    print "hello", name, url

Takes the same arguments as the Option constructor.

run(commands=None, default_command=None)

Prepares manager to receive command line input. Usually run inside “if __name__ == “__main__” block in a Python script.

Parameters:
  • commands – optional dict of commands. Appended to any commands added using add_command().
  • default_command – name of default command to run if no arguments passed.
shell(func)

Decorator that wraps function in shell command. This is equivalent to:

def _make_context(app):
    return dict(app=app)

manager.add_command("shell", Shell(make_context=_make_context))

The decorated function should take a single “app” argument, and return a dict.

For more sophisticated usage use the Shell class.

class flask_script.Command(func=None)

Base class for creating commands.

Parameters:func – Initialize this command by introspecting the function.
get_options()

By default, returns self.option_list. Override if you need to do instance-specific configuration.

run()

Runs a command. This must be implemented by the subclass. Should take arguments as configured by the Command options.

class flask_script.Shell(banner=None, make_context=None, use_ipython=True, use_bpython=True, use_ptipython=True, use_ptpython=True)

Runs a Python shell inside Flask application context.

Parameters:
  • banner – banner appearing at top of shell when started
  • make_context – a callable returning a dict of variables used in the shell namespace. By default returns a dict consisting of just the app.
  • use_ptipython – use PtIPython shell if available, ignore if not. The PtIPython shell can be turned off in command line by passing the –no-ptipython flag.
  • use_ptpython – use PtPython shell if available, ignore if not. The PtPython shell can be turned off in command line by passing the –no-ptpython flag.
  • use_bpython – use BPython shell if available, ignore if not. The BPython shell can be turned off in command line by passing the –no-bpython flag.
  • use_ipython – use IPython shell if available, ignore if not. The IPython shell can be turned off in command line by passing the –no-ipython flag.
class flask_script.Server(host='127.0.0.1', port=5000, use_debugger=None, use_reloader=None, threaded=False, processes=1, passthrough_errors=False, ssl_crt=None, ssl_key=None, **options)

Runs the Flask development server i.e. app.run()

Parameters:
  • host – server host
  • port – server port
  • use_debugger – Flag whether to default to using the Werkzeug debugger. This can be overriden in the command line by passing the -d or -D flag. Defaults to False, for security.
  • use_reloader – Flag whether to use the auto-reloader. Default to True when debugging. This can be overriden in the command line by passing the -r/-R flag.
  • threaded – should the process handle each request in a separate thread?
  • processes – number of processes to spawn
  • passthrough_errors – disable the error catching. This means that the server will die on errors but it can be useful to hook debuggers in (pdb etc.)
  • ssl_crt – path to ssl certificate file
  • ssl_key – path to ssl key file
  • optionswerkzeug.run_simple() options.
class flask_script.Option(*args, **kwargs)

Stores positional and optional arguments for ArgumentParser.add_argument.

Parameters:
  • name_or_flags – Either a name or a list of option strings, e.g. foo or -f, –foo
  • action – The basic type of action to be taken when this argument is encountered at the command-line.
  • nargs – The number of command-line arguments that should be consumed.
  • const – A constant value required by some action and nargs selections.
  • default – The value produced if the argument is absent from the command-line.
  • type – The type to which the command-line arg should be converted.
  • choices – A container of the allowable values for the argument.
  • required – Whether or not the command-line option may be omitted (optionals only).
  • help – A brief description of what the argument does.
  • metavar – A name for the argument in usage messages.
  • dest – The name of the attribute to be added to the object returned by parse_args().
class flask_script.Group(*options, **kwargs)

Stores argument groups and mutually exclusive groups for ArgumentParser.add_argument_group <http://argparse.googlecode.com/svn/trunk/doc/other-methods.html#argument-groups> or ArgumentParser.add_mutually_exclusive_group <http://argparse.googlecode.com/svn/trunk/doc/other-methods.html#add_mutually_exclusive_group>.

Note: The title and description params cannot be used with the exclusive or required params.

Parameters:
  • options – A list of Option classes to add to this group
  • title – A string to use as the title of the argument group
  • description – A string to use as the description of the argument group
  • exclusive – A boolean indicating if this is an argument group or a mutually exclusive group
  • required – A boolean indicating if this mutually exclusive group must have an option selected
flask_script.prompt(name, default=None)

Grab user input from command line.

Parameters:
  • name – prompt text
  • default – default value if no input provided.
flask_script.prompt_bool(name, default=False, yes_choices=None, no_choices=None)

Grabs user input from command line and converts to boolean value.

Parameters:
  • name – prompt text
  • default – default value if no input provided.
  • yes_choices – default ‘y’, ‘yes’, ‘1’, ‘on’, ‘true’, ‘t’
  • no_choices – default ‘n’, ‘no’, ‘0’, ‘off’, ‘false’, ‘f’
flask_script.prompt_pass(name, default=None)

Grabs hidden (password) input from command line.

Parameters:
  • name – prompt text
  • default – default value if no input provided.
flask_script.prompt_choices(name, choices, default=None, no_choice=('none', ))

Grabs user input from command line from set of provided choices.

Parameters:
  • name – prompt text
  • choices – list or tuple of available choices. Choices may be single strings or (key, value) tuples.
  • default – default value if no input provided.
  • no_choice – acceptable list of strings for “null choice”