Skip to main content

A toolkit for building command-line interfaces

Project description

Comandante Logo

Comandante is a toolkit for building command-line interfaces in Python.

Build Status Coverage Status License: MIT

Table of Contents

Installation

To get the latest release simply install it with a pip:

pip3 install --upgrade comandante

Getting Started

Some command-line interfaces (like pip, git, go, etc.) have a rich hierarchy of subcommands. Each subcommand may have its own set of arguments and options. Consider for example git commit, git remote add <repo>, git remote rename <old> <new> etc. Comandante is a Python library that makes building rich command-line interfaces (CLI) in Python extremely simple and straightforward.

Example

Here is a simple example:

#!/usr/bin/env python3
import sys, comandante as cli

# Define a new CLI handler as a `cli.Handler` subclass
class CliTool(cli.Handler):

    # define CLI commands as methods decorated with `@cli.command()` 
    @cli.command()
    def repeat(self, message, times: int):
        for i in range(times):
            print(message)

    @cli.command() 
    def sum(self, a: int, b: int):
        result = a + b
        print(result)
        return result

# Then simply pass command-line arguments to the CliTool#invoke
CliTool().invoke(sys.argv[1:])

That's it!

To execute repeat command simply run ./tool repeat with its required arguments:

$ ./tool repeat "Hello world" 2
Hello world
Hello world

The same goes for any defined command-methods:

$ ./tool sum 21 21
42

So in other words to create a command-line interface you simply:

  • define a normal Python class (inherited from the comandante.Handler)
  • equip your class methods with a minimal amount of metadata (via decorators)
  • call invoke(sys.argv[1:]) method on your class instance
  • comandante will inspect metadata and decide how to parse command-line arguments and which method to call.

Just a Normal Classes and Methods

No surprises! Handlers and commands are just a normal classes and methods.

Options

Command options are declared with @comandante.option(...) decorator.

Each command-method receives options as **kwargs parameters (if present).

Each command-method has a convenience method Command#options(kwargs) to merge default values with specified command-line options.

Command Options

Example below demonstrates how to define command option:

#!/usr/bin/env python3
import sys, comandante as cli


class DatabaseCli(cli.Handler):

    @cli.option(name='force', short='f', type=bool, default=False)
    @cli.command()
    def drop(self, database_name, **specified_options):
        # merge specified options with default values
        options = self.drop.options(specified_options)

        # merged options provides attribute-like element access
        if options.force or self.confirm():
            print("Database '{name}' was deleted".format(name=database_name))
        else: 
            print("Aborted")

    @staticmethod        
    def confirm():
        question = 'Are you sure? [y/N]: '
        value = input(question).lower()
        while value not in ['', 'n', 'y']:
            print("Please answer 'y' or 'n'")
            value = input(question)
        return value == 'y'


database_cli = DatabaseCli()
database_cli.invoke(sys.argv[1:])

Then in shell:

$ ./database drop production
Are you sure? [y/N]: y
Database 'production' was deleted

The following two examples are identical:

$ ./database drop -f production
Database 'production' was deleted
$ ./database drop --force production
Database 'production' was deleted

Class Options

Options could be declared on handler itself with Handler#declare_option method.

Options declared on handler will be declared on each of its command and subcommand recursively.

#!/usr/bin/env python3
import sys, comandante as cli

class CliTool(cli.Handler):
    def __init__(self):
        super().__init__()

        # define options
        self.declare_option('verbose', 'v', bool, False)

    @cli.command()
    def first(self, **specified_options):
        if 'verbose' in specified_options:
            print("Hello from the first!")

    @cli.command()
    def second(self, **specified_options):
        if 'verbose' in specified_options:
            print("Hello from the second!")


tool = CliTool()
tool.invoke(sys.argv[1:])

Then in shell:

$ ./tool first --verbose
Hello from the first!
$ ./tool second -v
Hello from the second!

Subcommands

As your CLI becomes more complex and harder to maintain, you might want to have commands that have its own set of commands handled by separate class. In Comandante to compose handlers into hierarchy you may simply use Handler#declare_command method.

Here is a simple example:

#!/usr/bin/env python3

import sys, comandante as cli

class Remote(cli.Handler):
    @cli.command()
    def add(self, name, uri):
        print("Adding repository", name, uri)

    @cli.command()
    def rename(self, old, new):
        print("Renaming repository", old, new) 

class Git(cli.Handler):
    def __init__(self):
        super().__init__() 

        # define subcommands
        self.declare_command(name='remote', handler=Remote())

    @cli.option('message', 'm', str, '')
    @cli.command()
    def commit(self, **specified_options):
        options = self.commit.options(specified_options)
        print("Committing changes with message '{}'".format(options.message))

git = Git()
git.invoke(sys.argv[1:])

Then in shell

$ ./git commit -m "Initial commit"
Committing changes with message 'Initial commit' 
$ ./git remote add origin git@github.com:stepan-anokhin/comandante.git
Adding repository origin git@github.com:stepan-anokhin/comandante.git
$ ./git remote rename origin destination
Renaming repository origin destination

Arguments

In python3 command argument types are declared using annotations:

import comandante as cli

class CliTool(cli.Handler):

    @cli.command()
    def do_something(self, a: int, b: float, c: str):
        print(a + b, c)

Type could be any callable. This callable will be simply called with a command-line string as the only argument. A result will be passed to the command-method as argument. The same is true for option types.

The only exception is bool arguments and options. Bool option doesn't receive any value, if specified it is considered to be True otherwise default value is used.

If no argument type is specified, then str is assumed by default. Type's __name__ attribute is used in automatic help output to represent argument/option type.

Comandante honors default argument values and varargs:

#!/usr/bin/env python3
import sys, comandante as cli

class CliTool(cli.Handler):

    @cli.command()
    def sum(self, *values: int):
        print(sum(values))

    @cli.command()
    def repeat(self, message, times: int = 2):
        for i in range(times):
            print(message)

CliTool().invoke(sys.argv[1:])
$ ./tool sum 1 2 3 4 5
15
$ ./tool repeat "Hello world!"
Hello world!
Hello world!

Type Library

Comandante also provides several higher-order types:

  • comandante.types.choice - to make sure argument value is one of the specified options
  • comandante.types.listof - to parse comma-separated lists (e.g. listof(int) will parse "1,2,3,4" into [1, 2, 3, 4])

You may take a look into the comandante.types to get some additional insights.

Python 2

Python 2 doesn't support parameter annotations. To specify argument types use @comandante.signature()

import comandante as cli

class CliTool(cli.Handler):

    @cli.signature(a=int, b=float)
    @cli.command()
    def do_something(self, a, b, c):
        print(a + b, c)

Printing Help

Comandante provides predefined help command for you which will print formatted help information to the stdout. Command and handler descriptions are taken from the corresponding docstrings.

Example:

import sys
import comandante as cli


class Git(cli.Handler):
    """The stupid content tracker.

    Git is a fast, scalable, distributed revision control system with
    an unusually *rich command* set that provides both high-level operations
    and full access to internals.

    See *gittutorial*(7) to get started, then see *giteveryday*(7) for a useful
    minimum set of commands. The *Git User’s Manual*[1] has a more in-depth
    introduction.
    """

    @cli.option(name='message', short='m', type=str, default='', descr="""
    Use the given <msg> as the commit message. If multiple *-m* options are 
    given, their values are concatenated as separate paragraphs.
    """)
    @cli.command()
    def commit(self):
        """Record changes to the repository

        Create a new commit containing the current contents
        of the index and the given log message describing
        the changes. The new commit is a direct child of HEAD,
        usually the tip of the current branch, and the branch
        is updated to point to it (unless no branch is associated
        with the working tree, in which case HEAD is "detached"
        as described in *git-checkout*(1)).
        """
        print("Committing...")

    @cli.command()
    def clone(self, repository, directory=None):
        """Clone a repository into a new directory

        Clones a repository into a newly created directory, creates
        remote-tracking branches for each branch in the cloned
        repository (visible using git branch *-r*), and creates and
        checks out an initial branch that is forked from the cloned
        repository’s currently active branch.
        """
        print("Cloning...")

Git().invoke(sys.argv[1:])

./git output:

git

./git help clone output:

git help clone

./git help commit output:

git help commit

Error Handling

Successful calls to Handler#invoke and Command#invoke methods return the same value as the corresponding command-method.

By design Comandante doesn't hide any exceptions raised in the course of invoke call. The only special case is subclasses of comandnate.errors.CliSyntaxException which results in help printing before being re-raised.

So it is up to the caller to decide how to handle exceptions.

A reasonable error handling may look like this:

import sys, logging, comandante as cli

# ... initialize handler and logger ... 

try: 
    handler.invoke(sys.argv[1:])
except cli.errors.CliSyntaxException:
    sys.exit(1)
except:
    logger.exception('Unexpected exception')
    sys.exit(1)

Testing Your CLI

Handlers and commands could be tested just like any other classes and methods.

Example cli:

import comandante as cli


class DatabaseCli(cli.Handler):
    def __init__(self, database):
        super().__init__()
        self._database = database

    @cli.option('force', 'f', bool, False)
    @cli.command()
    def drop(self, database_name, **specified_options):
        options = self.drop.options(specified_options)
        if options.force or self.confirm():
            self._database.drop(database_name)

    def confirm(self):
        question = 'Are you sure? [y/N]: '
        value = input(question).lower()
        while value not in ['', 'n', 'y']:
            print("Please answer 'y' or 'n'")
            value = input(question)
        return value == 'y'

Example tests:

from unittest import TestCase
from unittest.mock import MagicMock as Mock


class DatabaseCliTests(TestCase):
    def test_forced_database_drop(self):
        fake_database = Mock()
        database_cli = DatabaseCli(database=fake_database)

        database_cli.drop('production', force=True)

        fake_database.drop.assert_called_with('production')

    def test_database_drop(self):
        fake_database = Mock()
        database_cli = DatabaseCli(database=fake_database)
        database_cli.confirm = Mock(return_value=True)  # confirm

        database_cli.drop('production')

        fake_database.drop.assert_called_with('production')

    def test_rejected_database_drop(self):
        fake_database = Mock()
        database_cli = DatabaseCli(database=fake_database)
        database_cli.confirm = Mock(return_value=False)  # reject

        database_cli.drop('production')

        fake_database.drop.assert_not_called()

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

comandante-0.0.3.tar.gz (19.3 kB view details)

Uploaded Source

Built Distribution

comandante-0.0.3-py3-none-any.whl (25.3 kB view details)

Uploaded Python 3

File details

Details for the file comandante-0.0.3.tar.gz.

File metadata

  • Download URL: comandante-0.0.3.tar.gz
  • Upload date:
  • Size: 19.3 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/1.15.0 pkginfo/1.5.0.1 requests/2.22.0 setuptools/41.2.0 requests-toolbelt/0.9.1 tqdm/4.36.1 CPython/3.5.2

File hashes

Hashes for comandante-0.0.3.tar.gz
Algorithm Hash digest
SHA256 cf279c36c70053360acc6b12cfec5b18d03beb6c1fb7f6ab219f037103495298
MD5 cbdfe366c28393ca746596e42baad075
BLAKE2b-256 270f3dd4a6968af46413dfb2b044d55fa7b20ce89ec1b4d3f790ab9eeee517b8

See more details on using hashes here.

File details

Details for the file comandante-0.0.3-py3-none-any.whl.

File metadata

  • Download URL: comandante-0.0.3-py3-none-any.whl
  • Upload date:
  • Size: 25.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/1.15.0 pkginfo/1.5.0.1 requests/2.22.0 setuptools/41.2.0 requests-toolbelt/0.9.1 tqdm/4.36.1 CPython/3.5.2

File hashes

Hashes for comandante-0.0.3-py3-none-any.whl
Algorithm Hash digest
SHA256 662392fa2a872e8f72831a41e9600d738d0ed49475e509578a07489999b683a5
MD5 3442ab62adc3765faa004e6cf57bc04d
BLAKE2b-256 daad1b93710a37204e9e7ee52292477d04c17fa7d9b795a5c6b02243bf47c70d

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page