Reputation: 113
I'm trying to automate a process using python. If I am just in the terminal the workflow looks like:
user:> . /path/to/env1.sh
user:> python something.py
user:> exit
user:> . /path/to/env2.sh
user:> python something2.py
user:> exit
etc for a few more steps. Each env.sh
spawns a new script with a whole slew of environment variables and whatnot set within the current directory. I'm pretty sure I need to use subprocess, but I'm not exactly sure how to go about it. Ideally the workflow would go: open new shell --> run some commands --> exit shell --> repeat as necessary.
EDIT: It seems some clarification is needed. I understand how to use subprocess.Popen()
and subprocess.call()
to call things from within the shell that the Python script was called from. This is not what I need. When one calls env.sh
it sets a whole ton of environment variables and a few other pertinent things and then drops you into a shell to run commands. It is important to note env.sh
does not terminate until one types exit
after running desired commands. Using subprocess.call("./env.sh", shell = True)
opens the shell and stops there. It is just like entering the command ./env.sh
except that when one issues the exit
command, the rest of the python script. So:
subprocess.call(". /path/to/env.sh", shell = True)
subprocess.call("python something.py", shell = True)
Does NOT do what I need it to do, nor does:
p = subprocess.Popen(". /path/to/env.sh", shell = True)
subprocess.call("python something.py", shell = True)
p.kill()
Upvotes: 11
Views: 43364
Reputation: 414207
As I understand you want to run a command and then pass it other commands:
from subprocess import Popen, PIPE
p = Popen("/path/to/env.sh", stdin=PIPE) # set environment, start new shell
p.communicate("python something.py\nexit") # pass commands to the opened shell
Upvotes: 9
Reputation: 309909
subprocess
calls (particular Popen
) accepts an env
argument which is a mapping of environement variables to values. You can use that. e.g.
env = {'FOO': 'Bar', 'HOME': '/path/to/home'}
process = subprocess.Popen(['python', 'something.py'], env=env)
Of course, usually, it's better to just call some functions after *import*ing something.py
instead of spawning a whole new process.
Upvotes: 6
Reputation: 604
p = subprocess.Popen(". /path/to/env.sh", shell = True, stdout=subprocess.PIPE).communicate()
subprocess.call("python something.py", shell = True).communicate()
Upvotes: -2
Reputation: 19264
You can use subprocess
:
>>> import subprocess
>>> subprocess.call('python something.py', shell = True)
Or you can use os
:
>>> import os
>>> os.system('python something.py')
Here is an example (turn on your speakers):
>>> import os
>>> os.system('say Hello')
Upvotes: 8
Reputation: 5972
Here is a help for you.... For running command you could do:
1)
from subprocess import call
call(["ls", "-l"])
2)
import os
os.system("command")
Example:
import os
f = os.popen('date')
now = f.read()
print "Today is ", now
For enabling terminal you can import os
module:
import os
os.system('python script.py')
Or as mentioned you can use import subprocess
Upvotes: 0