Command Line Interface

Installing Flask installs the flask script, a Click command lineinterface, in your virtualenv. Executed from the terminal, this script givesaccess to built-in, extension, and application-defined commands. The —helpoption will give more information about any commands and options.

Application Discovery

The flask command is installed by Flask, not your application; it must betold where to find your application in order to use it. The FLASK_APPenvironment variable is used to specify how to load the application.

Unix Bash (Linux, Mac, etc.):

  1. $ export FLASK_APP=hello
  2. $ flask run

Windows CMD:

  1. > set FLASK_APP=hello
  2. > flask run

Windows PowerShell:

  1. > $env:FLASK_APP = "hello"
  2. > flask run

While FLASK_APP supports a variety of options for specifying yourapplication, most use cases should be simple. Here are the typical values:

  • (nothing)
  • The file wsgi.py is imported, automatically detecting an app(app). This provides an easy way to create an app from a factory withextra arguments.

  • FLASK_APP=hello

  • The name is imported, automatically detecting an app (app) or factory(create_app).

FLASK_APP has three parts: an optional path that sets the current workingdirectory, a Python file or dotted import path, and an optional variablename of the instance or factory. If the name is a factory, it can optionallybe followed by arguments in parentheses. The following values demonstrate theseparts:

  • FLASK_APP=src/hello
  • Sets the current working directory to src then imports hello.

  • FLASK_APP=hello.web

  • Imports the path hello.web.

  • FLASK_APP=hello:app2

  • Uses the app2 Flask instance in hello.

  • FLASK_APP="hello:create_app('dev')"

  • The create_app factory in hello is called with the string 'dev'as the argument.

If FLASK_APP is not set, the command will try to import “app” or“wsgi” (as a “.py” file, or package) and try to detect an applicationinstance or factory.

Within the given import, the command looks for an application instance namedapp or application, then any application instance. If no instance isfound, the command looks for a factory function named create_app ormake_app that returns an instance.

When calling an application factory, if the factory takes an argument namedscript_info, then the ScriptInfo instance is passed as akeyword argument. If the application factory takes only one argument and noparentheses follow the factory name, the ScriptInfo instanceis passed as a positional argument. If parentheses follow the factory name,their contents are parsed as Python literals and passes as arguments to thefunction. This means that strings must still be in quotes.

Run the Development Server

The run command will start the development server. Itreplaces the Flask.run() method in most cases.

  1. $ flask run
  2. * Serving Flask app "hello"
  3. * Running on http://127.0.0.1:5000/ (Press CTRL+C to quit)

Warning

Do not use this command to run your application in production.Only use the development server during development. The development serveris provided for convenience, but is not designed to be particularly secure,stable, or efficient. See Deployment Options for how to run in production.

Open a Shell

To explore the data in your application, you can start an interactive Pythonshell with the shell command. An applicationcontext will be active, and the app instance will be imported.

  1. $ flask shell
  2. Python 3.6.2 (default, Jul 20 2017, 03:52:27)
  3. [GCC 7.1.1 20170630] on linux
  4. App: example
  5. Instance: /home/user/Projects/hello/instance
  6. >>>

Use shell_context_processor() to add other automatic imports.

Environments

Changelog

New in version 1.0.

The environment in which the Flask app runs is set by theFLASK_ENV environment variable. If not set it defaults toproduction. The other recognized environment is development.Flask and extensions may choose to enable behaviors based on theenvironment.

If the env is set to development, the flask command will enabledebug mode and flask run will enable the interactive debugger andreloader.

  1. $ FLASK_ENV=development flask run
  2. * Serving Flask app "hello"
  3. * Environment: development
  4. * Debug mode: on
  5. * Running on http://127.0.0.1:5000/ (Press CTRL+C to quit)
  6. * Restarting with inotify reloader
  7. * Debugger is active!
  8. * Debugger PIN: 223-456-919

Watch Extra Files with the Reloader

When using development mode, the reloader will trigger whenever yourPython code or imported modules change. The reloader can watchadditional files with the —extra-files option, or theFLASK_RUN_EXTRA_FILES environment variable. Multiple paths areseparated with :, or ; on Windows.

  1. $ flask run --extra-files file1:dirA/file2:dirB/
  2. # or
  3. $ export FLASK_RUN_EXTRA_FILES=file1:dirA/file2:dirB/
  4. $ flask run
  5. * Running on http://127.0.0.1:8000/
  6. * Detected change in '/path/to/file1', reloading

Debug Mode

Debug mode will be enabled when FLASK_ENV is development,as described above. If you want to control debug mode separately, useFLASK_DEBUG. The value 1 enables it, 0 disables it.

Environment Variables From dotenv

Rather than setting FLASK_APP each time you open a new terminal, you canuse Flask’s dotenv support to set environment variables automatically.

If python-dotenv is installed, running the flask command will setenvironment variables defined in the files .env and .flaskenv.This can be used to avoid having to set FLASK_APP manually every time youopen a new terminal, and to set configuration using environment variablessimilar to how some deployment services work.

Variables set on the command line are used over those set in .env,which are used over those set in .flaskenv. .flaskenv should beused for public variables, such as FLASK_APP, while .env should notbe committed to your repository so that it can set private variables.

Directories are scanned upwards from the directory you call flaskfrom to locate the files. The current working directory will be set to thelocation of the file, with the assumption that that is the top level projectdirectory.

The files are only loaded by the flask command or callingrun(). If you would like to load these files when running inproduction, you should call load_dotenv() manually.

Setting Command Options

Click is configured to load default values for command options fromenvironment variables. The variables use the patternFLASK_COMMAND_OPTION. For example, to set the port for the runcommand, instead of flask run —port 8000:

  1. $ export FLASK_RUN_PORT=8000
  2. $ flask run
  3. * Running on http://127.0.0.1:8000/

These can be added to the .flaskenv file just like FLASK_APP tocontrol default command options.

Disable dotenv

The flask command will show a message if it detects dotenv files butpython-dotenv is not installed.

  1. $ flask run
  2. * Tip: There are .env files present. Do "pip install python-dotenv" to use them.

You can tell Flask not to load dotenv files even when python-dotenv isinstalled by setting the FLASK_SKIP_DOTENV environment variable.This can be useful if you want to load them manually, or if you’re usinga project runner that loads them already. Keep in mind that theenvironment variables must be set before the app loads or it won’tconfigure as expected.

  1. $ export FLASK_SKIP_DOTENV=1
  2. $ flask run

Environment Variables From virtualenv

If you do not want to install dotenv support, you can still set environmentvariables by adding them to the end of the virtualenv’s activatescript. Activating the virtualenv will set the variables.

Unix Bash, venv/bin/activate:

  1. $ export FLASK_APP=hello

Windows CMD, venv\Scripts\activate.bat:

  1. > set FLASK_APP=hello

It is preferred to use dotenv support over this, since .flaskenv can becommitted to the repository so that it works automatically wherever the projectis checked out.

Custom Commands

The flask command is implemented using Click. See that project’sdocumentation for full information about writing commands.

This example adds the command create-user that takes the argumentname.

  1. import click
  2. from flask import Flask
  3.  
  4. app = Flask(__name__)
  5.  
  6. @app.cli.command("create-user")
  7. @click.argument("name")
  8. def create_user(name):
  9. ...
  1. $ flask create-user admin

This example adds the same command, but as user create, a command in agroup. This is useful if you want to organize multiple related commands.

  1. import click
  2. from flask import Flask
  3. from flask.cli import AppGroup
  4.  
  5. app = Flask(__name__)
  6. user_cli = AppGroup('user')
  7.  
  8. @user_cli.command('create')
  9. @click.argument('name')
  10. def create_user(name):
  11. ...
  12.  
  13. app.cli.add_command(user_cli)
  1. $ flask user create demo

See Testing CLI Commands for an overview of how to test your customcommands.

Registering Commands with Blueprints

If your application uses blueprints, you can optionally register CLIcommands directly onto them. When your blueprint is registered onto yourapplication, the associated commands will be available to the flaskcommand. By default, those commands will be nested in a group matchingthe name of the blueprint.

  1. from flask import Blueprint
  2.  
  3. bp = Blueprint('students', __name__)
  4.  
  5. @bp.cli.command('create')
  6. @click.argument('name')
  7. def create(name):
  8. ...
  9.  
  10. app.register_blueprint(bp)
  1. $ flask students create alice

You can alter the group name by specifying the cli_group parameterwhen creating the Blueprint object, or later withapp.register_blueprint(bp, cli_group='…').The following are equivalent:

  1. bp = Blueprint('students', __name__, cli_group='other')
  2. # or
  3. app.register_blueprint(bp, cli_group='other')
  1. $ flask other create alice

Specifying cli_group=None will remove the nesting and merge thecommands directly to the application’s level:

  1. bp = Blueprint('students', __name__, cli_group=None)
  2. # or
  3. app.register_blueprint(bp, cli_group=None)
  1. $ flask create alice

Application Context

Commands added using the Flask app’s clicommand() decorator will be executed with an applicationcontext pushed, so your command and extensions have access to the app and itsconfiguration. If you create a command using the Click command()decorator instead of the Flask decorator, you can usewith_appcontext() to get the same behavior.

  1. import click
  2. from flask.cli import with_appcontext
  3.  
  4. @click.command()
  5. @with_appcontext
  6. def do_work():
  7. ...
  8.  
  9. app.cli.add_command(do_work)

If you’re sure a command doesn’t need the context, you can disable it:

  1. @app.cli.command(with_appcontext=False)def do_work():

Plugins

Flask will automatically load commands specified in the flask.commandsentry point. This is useful for extensions that want to add commands whenthey are installed. Entry points are specified in setup.py

  1. from setuptools import setup
  2.  
  3. setup(
  4. name='flask-my-extension',
  5. ...,
  6. entry_points={
  7. 'flask.commands': [
  8. 'my-command=flask_my_extension.commands:cli'
  9. ],
  10. },
  11. )

Inside flask_my_extension/commands.py you can then export a Clickobject:

  1. import click
  2.  
  3. @click.command()
  4. def cli():
  5. ...

Once that package is installed in the same virtualenv as your Flask project,you can run flask my-command to invoke the command.

Custom Scripts

When you are using the app factory pattern, it may be more convenient to defineyour own Click script. Instead of using FLASK_APP and letting Flask loadyour application, you can create your own Click object and export it as aconsole script entry point.

Create an instance of FlaskGroup and pass it the factory:

  1. import click
  2. from flask import Flask
  3. from flask.cli import FlaskGroup
  4.  
  5. def create_app():
  6. app = Flask('wiki')
  7. # other setup
  8. return app
  9.  
  10. @click.group(cls=FlaskGroup, create_app=create_app)
  11. def cli():
  12. """Management script for the Wiki application."""

Define the entry point in setup.py:

  1. from setuptools import setup
  2.  
  3. setup(
  4. name='flask-my-extension',
  5. ...,
  6. entry_points={
  7. 'console_scripts': [
  8. 'wiki=wiki:cli'
  9. ],
  10. },
  11. )

Install the application in the virtualenv in editable mode and the customscript is available. Note that you don’t need to set FLASK_APP.

  1. $ pip install -e .
  2. $ wiki run

Errors in Custom Scripts

When using a custom script, if you introduce an error in yourmodule-level code, the reloader will fail because it can no longerload the entry point.

The flask command, being separate from your code, does not havethis issue and is recommended in most cases.

PyCharm Integration

Prior to PyCharm 2018.1, the Flask CLI features weren’t yet fullyintegrated into PyCharm. We have to do a few tweaks to get them workingsmoothly. These instructions should be similar for any other IDE youmight want to use.

In PyCharm, with your project open, click on Run from the menu bar andgo to Edit Configurations. You’ll be greeted by a screen similar tothis:screenshot of pycharm's run configuration settingsThere’s quite a few options to change, but once we’ve done it for onecommand, we can easily copy the entire configuration and make a singletweak to give us access to other commands, including any custom ones youmay implement yourself.

Click the + (Add New Configuration) button and select Python. Givethe configuration a good descriptive name such as “Run Flask Server”.For the flask run command, check “Single instance only” since youcan’t run the server more than once at the same time.

Select Module name from the dropdown (A) then input flask.

The Parameters field (B) is set to the CLI command to execute(with any arguments). In this example we use run, which will runthe development server.

You can skip this next step if you’re using Environment Variables From dotenv. We need toadd an environment variable (C) to identify our application. Clickon the browse button and add an entry with FLASK_APP on the left andthe Python import or file on the right (hello for example).

Next we need to set the working directory (D) to be the folder whereour application resides.

If you have installed your project as a package in your virtualenv, youmay untick the PYTHONPATH options (E). This will more accuratelymatch how you deploy the app later.

Click Apply to save the configuration, or OK to save and close thewindow. Select the configuration in the main PyCharm window and clickthe play button next to it to run the server.

Now that we have a configuration which runs flask run from withinPyCharm, we can copy that configuration and alter the Script argumentto run a different CLI command, e.g. flask shell.