The Console Component

The Console component eases the creation of beautiful and testable command line interfaces.

Symfony2 ships with a Console component, which allows you to create command-line commands. Your console commands can be used for any recurring task, such as cronjobs, imports, or other batch jobs.

Installation

You can install the component in many different ways:

  • Use the official Git repository (https://github.com/symfony/Console);
  • Install it via PEAR ( pear.symfony.com/Console);
  • Install it via Composer (symfony/console on Packagist).

Creating a basic Command

To make the console commands available automatically with Symfony2, create a Command directory inside your bundle and create a php file suffixed with Command.php for each command that you want to provide. For example, if you want to extend the AcmeDemoBundle (available in the Symfony Standard Edition) to greet us from the command line, create GreetCommand.php and add the following to it:

// src/Acme/DemoBundle/Command/GreetCommand.php
namespace Acme\DemoBundle\Command;

use Symfony\Bundle\FrameworkBundle\Command\ContainerAwareCommand;
use Symfony\Component\Console\Input\InputArgument;
use Symfony\Component\Console\Input\InputInterface;
use Symfony\Component\Console\Input\InputOption;
use Symfony\Component\Console\Output\OutputInterface;

class GreetCommand extends ContainerAwareCommand
{
    protected function configure()
    {
        $this
            ->setName('demo:greet')
            ->setDescription('Greet someone')
            ->addArgument('name', InputArgument::OPTIONAL, 'Who do you want to greet?')
            ->addOption('yell', null, InputOption::VALUE_NONE, 'If set, the task will yell in uppercase letters')
        ;
    }

    protected function execute(InputInterface $input, OutputInterface $output)
    {
        $name = $input->getArgument('name');
        if ($name) {
            $text = 'Hello '.$name;
        } else {
            $text = 'Hello';
        }

        if ($input->getOption('yell')) {
            $text = strtoupper($text);
        }

        $output->writeln($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 console command by running the following

app/console demo:greet Fabien

This will print the following to the command line:

Hello Fabien

You can also use the --yell option to make everything uppercase:

app/console demo:greet Fabien --yell

This prints:

HELLO FABIEN

Coloring the Output

Whenever you output text, you can surround the text with tags to color its output. For example:

// green text
$output->writeln('<info>foo</info>');

// yellow text
$output->writeln('<comment>foo</comment>');

// black text on a cyan background
$output->writeln('<question>foo</question>');

// white text on a red background
$output->writeln('<error>foo</error>');

Using Command Arguments

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:

$this
    // ...
    ->addArgument('name', InputArgument::REQUIRED, 'Who do you want to greet?')
    ->addArgument('last_name', InputArgument::OPTIONAL, 'Your last name?')
    // ...

You now have access to a last_name argument in your command:

if ($lastName = $input->getArgument('last_name')) {
    $text .= ' '.$lastName;
}

The command can now be used in either of the following ways:

app/console demo:greet Fabien
app/console demo:greet Fabien Potencier

Using Command Options

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 an array 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:

$this
    // ...
    ->addOption('iterations', null, InputOption::VALUE_REQUIRED, 'How many times should the message be printed?', 1)

Next, use this in the command to print the message multiple times:

for ($i = 0; $i < $input->getOption('iterations'); $i++) {
    $output->writeln($text);
}

Now, when you run the task, you can optionally specify a --iterations flag:

app/console demo:greet Fabien

app/console demo:greet Fabien --iterations=5

The first example will only print once, since iterations is empty and defaults to 1 (the last argument of addOption). The second example will print five times.

Recall that options don’t care about their order. So, either of the following will work:

app/console demo:greet Fabien --iterations=5 --yell
app/console demo:greet Fabien --yell --iterations=5

There are 4 option variants you can use:

Option Value
InputOption::VALUE_IS_ARRAY This option accepts multiple values
InputOption::VALUE_NONE Do not accept input for this option (e.g. --yell)
InputOption::VALUE_REQUIRED This value is required (e.g. iterations=5)
InputOption::VALUE_OPTIONAL This value is optional

You can combine VALUE_IS_ARRAY with VALUE_REQUIRED or VALUE_OPTIONAL like this:

$this
    // ...
    ->addOption('iterations', null, InputOption::VALUE_REQUIRED | InputOption::VALUE_IS_ARRAY, 'How many times should the message be printed?', 1)

Asking the User for Information

When creating commands, you have the ability to collect more information from the user by asking him/her questions. For example, suppose you want to confirm an action before actually executing it. Add the following to your command:

$dialog = $this->getHelperSet()->get('dialog');
if (!$dialog->askConfirmation($output, '<question>Continue with this action?</question>', false)) {
    return;
}

In this case, the user will be asked “Continue with this action”, and unless they answer with y, the task will stop running. The third argument to askConfirmation is the default value to return if the user doesn’t enter any input.

You can also ask questions with more than a simple yes/no answer. For example, if you needed to know the name of something, you might do the following:

$dialog = $this->getHelperSet()->get('dialog');
$name = $dialog->ask($output, 'Please enter the name of the widget', 'foo');

Testing Commands

Symfony2 provides several tools to help you test your commands. The most useful one is the Symfony\Component\Console\Tester\CommandTester class. It uses special input and output classes to ease testing without a real console:

use Symfony\Component\Console\Tester\CommandTester;
use Symfony\Bundle\FrameworkBundle\Console\Application;
use Acme\DemoBundle\Command\GreetCommand;

class ListCommandTest extends \PHPUnit_Framework_TestCase
{
    public function testExecute()
    {
        $application = new Application();
        $application->add(new GreetCommand());

        $command = $application->find('demo:greet');
        $commandTester = new CommandTester($command);
        $commandTester->execute(array('command' => $command->getName()));

        $this->assertRegExp('/.../', $commandTester->getDisplay());

        // ...
    }
}

The :method:`Symfony\\Component\\Console\\Tester\\CommandTester::getDisplay` method returns what would have been displayed during a normal call from the console.

You can test sending arguments and options to the command by passing them as an array to the :method:`Symfony\\Component\\Console\\Tester\\CommandTester::getDisplay` method:

use Symfony\Component\Console\Tester\CommandTester;
use Symfony\Bundle\FrameworkBundle\Console\Application;
use Acme\DemoBundle\Command\GreetCommand;

class ListCommandTest extends \PHPUnit_Framework_TestCase
{

    //--

    public function testNameIsOutput()
    {
        $application = new Application();
        $application->add(new GreetCommand());

        $command = $application->find('demo:greet');
        $commandTester = new CommandTester($command);
        $commandTester->execute(
            array('command' => $command->getName(), 'name' => 'Fabien')
        );

        $this->assertRegExp('/Fabien/', $commandTester->getDisplay());
    }
}

Tip

You can also test a whole console application by using Symfony\Component\Console\Tester\ApplicationTester.

Calling an existing Command

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 (for instance, all commands that need to be run when the project’s code has changed on the production servers: clearing the cache, generating Doctrine2 proxies, dumping Assetic assets, ...).

Calling a command from another one is straightforward:

protected function execute(InputInterface $input, OutputInterface $output)
{
    $command = $this->getApplication()->find('demo:greet');

    $arguments = array(
        'command' => 'demo:greet',
        'name'    => 'Fabien',
        '--yell'  => true,
    );

    $input = new ArrayInput($arguments);
    $returnCode = $command->run($input, $output);

    // ...
}

First, you :method:`Symfony\\Component\\Console\\Command\\Command::find` the command you want to execute by passing the command name.

Then, you need to create a new Symfony\Component\Console\Input\ArrayInput with the arguments and options you want to pass to the command.

Eventually, calling the run() method actually executes the command and returns the returned code from the command (0 if everything went fine, any other integer otherwise).

Note

Most of the time, calling a command from code that is not executed on the command line is not a good idea for several reasons. First, the command’s output is optimized for the console. But more important, you can think of a command as being like a controller; it should use the model to do something and display feedback to the user. So, instead of calling a command from the Web, refactor your code and move the logic to a new class.