Package | Unix / Python 2 | Unix / Python 3 | Windows / Python 2 | Windows / Python 3 | Compatible forking* |
---|---|---|---|---|---|
sh |
âś… Works | âś… Works | đź”´ Not supported | đź”´ Not supported | đź”´ Fails |
pbs |
âś… Works | đź”´ Fails | âś… Works | đź”´ Fails | âś… Works |
pbs3 |
âś… Works | âś… Works | âś… Works | âś… Works | âś… Works |
* By compatible forking we mean a method of creating subprocesses that can work successfully inside a debugging environment such as the one provided by PyCharm.
-
First, in early 2012, the original
pbs
package was conceived by @amoffat. It could launch subprocesses on both Unix and Windows environments with Python 2. -
In late 2012, the
pbs
project was renamed tosh
, lots of functionality was added, but support for Windows was completely dropped. -
For these reasons, the legacy
pbs
package is still used whenever one needs to launch external processes on Windows machines and is still available atpip install pbs
. -
The old
pbs
package is also used for its more compatible way of starting subprocessed. The currentsh
module is not compatible with developing in PyCharm for instance. -
However the latest
pbs
version (v0.110 from Oct 20, 2012) does not work on Python 3. This package,pbs3
fixes this.
-
You can install this version with this command:
pip install pbs3
-
You can use this package with this import statement:
import pbs3 as pbs
-
Or if you are porting from
sh
code:import pbs3 as sh
This would be the traceback that the old pbs v0.110 would produce:
c:\python37\lib\site-packages\pbs.py in __call__(self, *args, **kwargs)
454 cwd=call_args["cwd"], stdin=stdin, stdout=stdout, stderr=stderr)
455
--> 456 return RunningCommand(command_ran, process, call_args, actual_stdin)
457
458
c:\python37\lib\site-packages\pbs.py in __init__(self, command_ran, process, call_args,
stdin)
166 if stdin: stdin = stdin.encode("utf8")
167 self._stdout, self._stderr = self.process.communicate(stdin)
--> 168 self._handle_exit_code(self.process.wait())
169
170 def __enter__(self):
c:\python37\lib\site-packages\pbs.py in _handle_exit_code(self, rc)
233 def _handle_exit_code(self, rc):
234 if rc not in self.call_args["ok_code"]:
--> 235 raise get_rc_exc(rc)(self.command_ran, self._stdout, self._stderr)
236
237 def __len__(self):
c:\python37\lib\site-packages\pbs.py in __init__(self, full_cmd, stdout, stderr)
93
94 msg = "\n\nRan: %r\n\nSTDOUT:\n\n %s\n\nSTDERR:\n\n %s" %\
---> 95 (full_cmd, tstdout.decode(), tstderr.decode())
96 super(ErrorReturnCode, self).__init__(msg)
97
AttributeError: 'str' object has no attribute 'decode'
This would occur when the program called exited with a non-zero return code.
PBS is a unique subprocess wrapper that maps your system programs to Python functions dynamically. PBS helps you write shell scripts in Python by giving you the good features of Bash (easy command calling, easy piping) with all the power and flexibility of Python.
from pbs import ifconfig
print ifconfig("eth0")
PBS is not a collection of system commands implemented in Python.
$> pip install pbs
The easiest way to get up and running is to import pbs directly or import your program from pbs:
import pbs
print pbs.ifconfig("eth0")
from pbs import ifconfig
print ifconfig("eth0")
A less common usage pattern is through PBS Command wrapper, which takes a full path to a command and returns a callable object. This is useful for programs that have weird characters in their names or programs that aren't in your $PATH:
import pbs
ffmpeg = pbs.Command("/usr/bin/ffmpeg")
ffmpeg(movie_file)
The last usage pattern is for trying PBS through an interactive REPL. By default, this acts like a star import (so all of your system programs will be immediately available as functions):
$> python pbs.py
Commands work like you'd expect. Just call your program's name like a function:
# print the contents of this directory
print ls("-l")
# get the longest line of this file
longest_line = wc(__file__, "-L")
# get interface information
print ifconfig("eth0")
Note that these aren't Python functions, these are running the binary commands on your system dynamically by resolving your PATH, much like Bash does. In this way, all the programs on your system are easily available in Python.
You can also call attributes on commands. This translates to the command name followed by the attribute name:
from pbs import git
# resolves to "git branch -v"
print git.branch("-v")
It turns out this is extremely useful for commands whose first argument is often another sub-command (like git, svn, time, sudo, etc). See "Baking" for an advanced usage of this.
Keyword arguments also work like you'd expect: they get replaced with the long-form and short-form commandline option:
# resolves to "curl http://duckduckgo.com/ -o page.html --silent"
curl("http://duckduckgo.com/", o="page.html", silent=True)
# or if you prefer not to use keyword arguments, this does the same thing:
curl("http://duckduckgo.com/", "-o", "page.html", "--silent")
# resolves to "adduser amoffat --system --shell=/bin/bash --no-create-home"
adduser("amoffat", system=True, shell="/bin/bash", no_create_home=True)
# or
adduser("amoffat", "--system", "--shell", "/bin/bash", "--no-create-home")
Piping has become function composition:
# sort this directory by biggest file
print sort(du(glob("*"), "-sb"), "-rn")
# print the number of folders and files in /etc
print wc(ls("/etc", "-1"), "-l")
PBS can redirect the standard and error output streams of a process to a file. This is done with the special _out and _err keyword arguments. You can pass a filename or a file object as the argument value. When the name of an already existing file is passed, the contents of the file will be overwritten.
ls(_out="files.list")
ls("nonexistent", _err="error.txt")
PBS can also redirect the error output stream to the standard output stream, using the special _err_to_out=True keyword argument.
Commands can be run within a "with" context. Popular commands using this might be "sudo" or "fakeroot":
with sudo:
print ls("/root")
If you need to run a command in a with context AND call it, for example, specifying a -p prompt with sudo, you need to use the "_with" keyword argument. This let's the command know that it's being run from a with context so it can behave correctly.
with sudo(p=">", _with=True):
print ls("/root")
Commands can be run in the background with the special _bg=True keyword argument:
# blocks
sleep(3)
print "...3 seconds later"
# doesn't block
p = sleep(3, _bg=True)
print "prints immediately!"
p.wait()
print "...and 3 seconds later"
You can also pipe together background processes!
p = wc(curl("http://github.com/", silent=True, _bg=True), "--bytes")
print "prints immediately!"
print "byte count of github: %d" % int(p) # lazily completes
This lets you start long-running commands at the beginning of your script (like a file download) and continue performing other commands in the foreground.
Foreground processes are processes that you want to interact directly with the default stdout and stdin of your terminal. In other words, these are processes that you do not want to return their output as a return value of their call. An example would be opening a text editor:
vim(file_to_edit)
This will block because pbs will be trying to aggregate the output of the command to python, without displaying anything to the screen. The solution is the "_fg" special keyword arg:
vim(file_to_edit, _fg=True)
This will open vim as expected and let you use it as expected, with all the input coming from the keyboard and the output going to the screen. The return value of a foreground process is an empty string.
"Which" finds the full path of a program, or returns None if it doesn't exist. This command is one of the few commands implemented as a Python function, and therefore doesn't rely on the "which" program actually existing.
print which("python") # "/usr/bin/python"
print which("ls") # "/bin/ls"
print which("some_command") # None
if not which("supervisorctl"): apt_get("install", "supervisor", "-y")
PBS is capable of "baking" arguments into commands. This is similar to the stdlib functools.partial wrapper. An example can speak volumes:
from pbs import ls
ls = ls.bake("-la")
print ls # "/usr/bin/ls -la"
# resolves to "ls / -la"
print ls("/")
The idea is that calling "bake" on a command creates a callable object that automatically passes along all of the arguments passed into "bake". This gets really interesting when you combine this with the attribute access on a command:
from pbs import ssh
# calling whoami on the server. this is tedious to do if you're running
# any more than a few commands.
iam1 = ssh("myserver.com", "-p 1393", "whoami")
# wouldn't it be nice to bake the common parameters into the ssh command?
myserver = ssh.bake("myserver.com", p=1393)
print myserver # "/usr/bin/ssh myserver.com -p 1393"
# resolves to "/usr/bin/ssh myserver.com -p 1393 whoami"
iam2 = myserver.whoami()
assert(iam1 == iam2) # True!
Now that the "myserver" callable represents a baked ssh command, you can call anything on the server easily:
# resolves to "/usr/bin/ssh myserver.com -p 1393 tail /var/log/dumb_daemon.log -n 100"
print myserver.tail("/var/log/dumb_daemon.log", n=100)
Environment variables are available much like they are in Bash:
print HOME
print SHELL
print PS1
You can set enviroment variables the usual way, through the os.environ mapping:
import os
os.environ["TEST"] = "123"
Now any new subprocess commands called from the script will be able to access that environment variable.
Exceptions are dynamically generated based on the return code of the command. This lets you catch a specific return code, or catch all error return codes through the base class ErrorReturnCode:
try: print ls("/some/non-existant/folder")
except ErrorReturnCode_2:
print "folder doesn't exist!"
create_the_folder()
except ErrorReturnCode:
print "unknown error"
exit(1)
Glob-expansion is not done on your arguments. For example, this will not work:
from pbs import du
print du("*")
You'll get an error to the effect of "cannot access '*': No such file or directory". This is because the "*" needs to be glob expanded:
from pbs import du, glob
print du(glob("*"))
You can access commandline arguments similar to Bash's $1, $2, etc by using ARG1, ARG2, etc:
print ARG1, ARG2
# if an argument isn't defined, it's set to None
if ARG10 is None: do_something()
You can access the entire argparse/optparse-friendly list of commandline arguments through "ARGV". This is recommended for flexibility:
import argparse
parser = argparse.ArgumentParser(prog="PROG")
parser.add_argument("-x", default=3, type=int)
ns = parser.parse_args(ARGV)
print ns.x
PBS automatically handles underscore-dash conversions. For example, if you want to call apt-get:
apt_get("install", "mplayer", y=True)
PBS looks for "apt_get", but if it doesn't find it, replaces all underscores with dashes and searches again. If the command still isn't found, a CommandNotFound exception is raised.
Commands with other, less-commonly symbols in their names must be accessed directly through the "Command" class wrapper. The Command class takes the full path to the program as a string:
p27 = Command(which("python2.7"))
print p27("-h")
The Command wrapper is also useful for commands that are not in your standard PATH:
script = Command("/tmp/temporary-script.sh")
print script()
Normally, if a command returns an exit code that is not 0, PBS raises an exception based on that exit code. However, if you have determined that an error code is normal and want to retrieve the output of the command without PBS raising an exception, you can use the "_ok_code" special argument to suppress the exception:
output = pbs.ls("dir_that_exists", "dir_that_doesnt", _ok_code=2)
In the above example, even though you're trying to list a directory that doesn't exist, you can still get the output from the directory that does exist by telling the command that 2 is an "ok" exit code, so don't raise an exception.
_ok_code can also take a list or tuple of numbers for multiple ok exit codes.