Skip to main content

Call any available system command from python

Project description

ps

Call any available system command from python

To install: pip install ps

Note: The package is aspires to be platform-independent but has only been tested with unix-flavored systems.

Quick Start

Commands() will gather all available commands it can find (searching through the same PATH environment variable that your system does) and makes them available for execution.

>>> from ps import Commands
>>> c = Commands()
>>> c.echo('hello world')
'hello world'
>>> c.ls('-la')
total 6271272
drwxr--r--@ 316 Thor.Whalen  staff      10112 Sep 22 14:01 .
drwxr-xr-x@  17 Thor.Whalen  staff        544 Mar  8  2022 ..
...

A Commands instance is also a Mapping, so you can do things like:

len(c)
# 3097
list(c)
# ['_2to3',
#  'Activate_ps1',
#  'AssetCacheLocatorUtil'
#  ...
# ]
'ls' in c
# True
'this_command_does_not_exist' in c
# False
py = c.get('python3.9', c.get('python3.8', None))
c['ls']
# Command(command='ls')

You can get help like so:

>>> c.ls.help()
LS(1)                     BSD General Commands Manual                    LS(1)

NAME
     ls -- list directory contents
...

Or just get the help string like so:

>>> help_string = c.ls.help_str()

More control

Commands is just a collection of Command instances, which itself is just a callable that will run a shell script with you, in a manner specified by the run function.

So let's have a quick look at those three objects to understand better what powers we have at our disposal.

run

A parametrizable way to run shell commands.

Works somewhat like the subprocess.run function. but with different defaults, as well as the additional arguments on_error and egress.

>>> output = run('pwd')
>>> os.path.isdir(output)  # verify that output is indeed a valid directory path
True

Also very important difference with subprocess.run: You don't specify a LIST of tokenized arguments here: You can specify the full (string) command or parts of it as a sequence of strings:

>>> assert run('echo hello world') == run('echo', 'hello', 'world') == b'hello world'

Note that run will return bytes of the output, stripped of extremal newlines. The argument that does the stripping is egress. You can use this argument to do something else with the output. For example, if you want to to cast the output to a str, strip it, then print it, you could specify this in the egress:

>>> run('echo hello world', egress=lambda x: print(x.decode().strip()))
hello world

run's purpose in life is designed to be curried. That is, you can use functools.partial to make your own specialized functions that use shell scripts as their backend.

>>> from functools import partial
>>> stripped_str = lambda x: x.decode().strip()
>>> pwd = partial(run, 'pwd', egress=stripped_str)
>>> ls_la = partial(run, 'ls', '-la', egress=lambda x: print(stripped_str(x)))
>>> current_dir = pwd()
>>> os.path.isdir(current_dir)
True
>>> ls_la(current_dir)  # doctest: +SKIP
total 56
drwxr-xr-x@  7 Thor.Whalen  staff   224 Sep 23 12:12 .
drwxr-xr-x@ 11 Thor.Whalen  staff   352 Sep 23 11:33 ..
-rw-r--r--@  1 Thor.Whalen  staff    48 Sep 22 12:47 __init__.py
-rw-r--r--@  1 Thor.Whalen  staff  4649 Sep 23 11:33 base.py
-rw-r--r--@  1 Thor.Whalen  staff   348 Sep 22 12:38 raw.py
-rw-r--r--@  1 Thor.Whalen  staff  8980 Sep 23 12:12 util.py

Command

A Command runs a specific shell script for you in a specific manner. The run function is the general function to do that, and we saw that you can curry run to specify what and how to run. Command just wraps such a curried run function (or any compliant run function you provide), and specifies what executable (the command argument) to actually run.

So not much over a curried run.

But what it does do as well is set up the ability to do other things that may be specific to the executable you're running, such as giving your (callable) command instance a signature, some docs, or a help method.

>>> import os
>>> pwd = Command('pwd')
>>> os.path.isdir(pwd())
True
>>> assert pwd.__doc__  # docs exist (and are non-empty)!
>>> # To print the docs:
>>> pwd.help()  # doctest: +SKIP
PWD(1)                    BSD General Commands Manual                   PWD(1)

NAME
     pwd -- return working directory name

SYNOPSIS
     pwd [-L | -P]
...

Commands

A collection of commands.

The general usage is that you can specify a mapping between valid python identifiers (alphanumerical strings (and underscores) that don't start with a number) and functions. If instead of functions you specify a string, a factory comes in play to make a function based on your string. By default, it will consider it as a console command and give you a function that runs it.

>>> import os
>>>
>>> c = Commands({
...     'current_dir': 'pwd',
...     'sys_listdir': 'ls -l',
...     'listdir': os.listdir,
...     'echo': Command('echo', egress=lambda x: print(x.decode().strip())),
... })
>>>
>>> list(c)
['current_dir', 'sys_listdir', 'listdir', 'echo']
>>> current_folder = c.current_dir()
>>> os.path.isdir(current_folder)
True
>>> b = c.sys_listdir()
>>> b[:40]  # doctest: +SKIP
b'total 56\n-rw-r--r--@ 1 Thor.Whalen  staf'
>>> a_list_of_filenames = c.listdir()
>>> isinstance(a_list_of_filenames, list)
True
>>> c.echo('hello world')
hello world

If you don't specify any commands, it will gather all executable names it can find in your local system (according to your PATH environment variable), map those to valid python identifiers if needed, and use that.

Important: Note that finding executable in the PATH doesn't mean that it will work, or is safe -- so use with care!

>>> c = Commands()
>>> assert len(c) > 0
>>> 'ls' in c
True

You can access the 'ls' command as a key or an attribute

>>> assert c['ls'] == c.ls

You can print the .help() (docs) of any command, or just get the help string:

>>> man_page = c.ls.help_str()

Let's see if these docs have a few things we expect it to have for ls:

>>> assert 'BSD General Commands Manual' in man_page
>>> assert 'list directory contents' in man_page

Let's see what the output of ls gives us:

>>> output = c.ls('-la').decode()  # execute "ls -la"
>>> assert 'total' in output  # "ls -l" output usually starts with "total"
>>> assert '..' in output  # the "-a" flag includes '..' as a file

Note that we needed to decode the output here. That's because by default the output of a command will be captured in bytes. If you want to apply a decoder to (attempt to) convert all outputs into strings, you can specify a factory that will do this for you automatically.

The default factory is Command, which has a run argument that defines how an instruction should be run. The default of run is run_command, which conveniently has an egress argument where you can specify a function to call on the output.

So one solution to define a Commands instance that will (attempt to) output strings systematically is to do this:

>>> from functools import partial
>>> from ps.util import run
>>> run_and_cast_to_str = partial(run, egress=bytes.decode)
>>> factory = partial(Command, run=run_and_cast_to_str)
>>> cc = Commands(factory=factory)

So now we have:

>>> output = cc.ls('-la')
>>> isinstance(output, str)  # it's already decoded for us!
True

Notes

Auto docs

Trying to figure out a robust way to get doc strings. Unfortunately, man <command> and <command> --help don't always exist, and the former isn't even always correct (i.e. linked to the same executable as what which <command> says)

See: https://stackoverflow.com/questions/73814043/universal-help-for-terminal-commands

Misc references

Tutorial on options for calling shell commands from python

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

ps-0.1.0.tar.gz (15.0 kB view hashes)

Uploaded Source

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