Create beautiful and testable command-line interfaces.
Cleo is mostly a higher level wrapper for CliKit, so a lot of the components and utilities comes from it. Refer to its documentation for more information.
To make a command that greets you from the command line, create greet_command.py
and add the following to it:
from cleo import Command
class GreetCommand(Command):
"""
Greets someone
greet
{name? : Who do you want to greet?}
{--y|yell : If set, the task will yell in uppercase letters}
"""
def handle(self):
name = self.argument('name')
if name:
text = 'Hello {}'.format(name)
else:
text = 'Hello'
if self.option('yell'):
text = text.upper()
self.line(text)
You also need to create the file to run at the command line which creates an Application
and adds commands to it:
Test the new command by running the following
This will print the following to the command line:
Hello John
You can also use the --yell
option to make everything uppercase:
This prints:
HELLO JOHN
As you may have already seen, Cleo uses the command docstring to determine the command definition. The docstring must be in the following form :
The signature being in the following form:
The signature can span multiple lines.
Whenever you output text, you can surround the text with tags to color its output. For example:
The closing tag can be replaced by </>
, which revokes all formatting options established by the last opened tag.
It is possible to define your own styles using the add_style()
method:
Available foreground and background colors are: black
, red
, green
, yellow
, blue
, magenta
, cyan
and white
.
And available options are: bold
, underscore
, blink
, reverse
and conceal
.
You can also set these colors and options inside the tag name:
Cleo has four verbosity levels. These are defined in the Output
class:
Mode | Meaning | Console option |
---|---|---|
NA |
Do not output any messages | -q or --quiet |
clikit.VERBOSITY_NORMAL |
The default verbosity level | (none) |
clikit.VERBOSITY_VERBOSE |
Increased verbosity of messages | -v |
clikit.VERBOSITY_VERY_VERBOSE |
Informative non essential messages | -vv |
clikit.VERBOSITY_DEBUG |
Debug messages | -vvv |
It is possible to print a message in a command for only a specific verbosity level. For example:
There are also more semantic methods you can use to test for each of the verbosity levels:
You can also pass the verbosity flag directly to line().
When the quiet level is used, all output is suppressed.
The most interesting part of the commands are the arguments and options that you can make available. Arguments are the strings - separated by spaces - that come after the command name itself. They are ordered, and can be optional or required. For example, add an optional last_name
argument to the command and make the name
argument required:
You now have access to a last_name
argument in your command:
The command can now be used in either of the following ways:
It is also possible to let an argument take a list of values (imagine you want to greet all your friends). For this it must be specified at the end of the argument list:
To use this, just specify as many names as you want:
You can access the names
argument as a list:
There are 3 argument variants you can use:
Mode | Notation | Value |
---|---|---|
clikit.ARGUMENT_REQUIRED |
none (just write the argument name) | The argument is required |
clikit.ARGUMENT_OPTIONAL |
argument? |
The argument is optional and therefore can be omitted |
clikit.ARGUMENT_MULTI_VALUED |
argument* |
The argument can contain an indefinite number of arguments and must be used at the end of the argument list |
You can combine them like this:
If you want to set a default value, you can it like so:
argument=default
The argument will then be considered optional.
Unlike arguments, options are not ordered (meaning you can specify them in any order) and are specified with two dashes (e.g. --yell
- you can also declare a one-letter shortcut that you can call with a single dash like -y
). Options are always optional, and can be setup to accept a value (e.g. --dir=src
) or simply as a boolean flag without a value (e.g. --yell
).
Tip
It is also possible to make an option optionally accept a value (so that --yell
or --yell=loud
work). Options can also be configured to accept a list of values.
For example, add a new option to the command that can be used to specify how many times in a row the message should be printed:
Next, use this in the command to print the message multiple times:
Now, when you run the task, you can optionally specify a --iterations
flag:
The first example will only print once, since iterations
is empty and defaults to 1
. The second example will print five times.
Recall that options don't care about their order. So, either of the following will work:
There are 4 option variants you can use:
Option | Notation | Value |
---|---|---|
clikit.OPTION_MULTI_VALUED |
--option=* |
This option accepts multiple values (e.g. --dir=/foo --dir=/bar ) |
clikit.OPTION_NO_VALUE |
--option |
Do not accept input for this option (e.g. --yell ) |
clikit.OPTION_REQUIRED_VALUE |
--option= |
This value is required (e.g. --iterations=5 ), the option itself is still optional |
clikit.OPTION_OPTIONAL_VALUE |
--option=? |
This option may or may not have a value (e.g. --yell or --yell=loud ) |
You can combine them like this:
Cleo provides several tools to help you test your commands. The most useful one is the CommandTester
class. It uses a special IO class to ease testing without a real console:
import pytest
from cleo import Application
from cleo import CommandTester
def test_execute(self):
application = Application()
application.add(GreetCommand())
command = application.find('demo:greet')
command_tester = CommandTester(command)
command_tester.execute()
assert "..." == tester.io.fetch_output()
The CommandTester.io.fetch_output()
method returns what would have been displayed during a normal call from the console. CommandTester.io.fetch_error()
is also available to get what you have been written to the stderr.
You can test sending arguments and options to the command by passing them as a string to the CommandTester.execute()
method:
import pytest
from cleo import Application
from cleo import CommandTester
def test_execute(self):
application = Application()
application.add(GreetCommand())
command = application.find('demo:greet')
command_tester = CommandTester(command)
command_tester.execute("John")
assert "John" in tester.io.fetch_output()
You can also test a whole console application by using the ApplicationTester
class.
If a command depends on another one being run before it, instead of asking the user to remember the order of execution, you can call it directly yourself. This is also useful if you want to create a "meta" command that just runs a bunch of other commands.
Calling a command from another one is straightforward:
If you want to suppress the output of the executed command, you can use the call_silent()
method instead.
Cleo supports automatic (tab) completion in bash
, zsh
and fish
.
To activate support for autocompletion, pass a complete
keyword when initializing your application:
Now, register completion for your application by running one of the following in a terminal, replacing [program]
with the command you use to run your application:
# BASH - Ubuntu / Debian
[program] completions bash | sudo tee /etc/bash_completion.d/[program].bash-completion
# BASH - Mac OSX (with Homebrew "bash-completion")
[program] completions bash > $(brew --prefix)/etc/bash_completion.d/[program].bash-completion
# ZSH - Config file
mkdir ~/.zfunc
echo "fpath+=~/.zfunc" >> ~/.zshrc
[program] completions zsh > ~/.zfunc/_test
# FISH
[program] completions fish > ~/.config/fish/completions/[program].fish