Python subprocess module use

Question

How do you specify what directory the python module subprocess uses to execute commands? I want to run a C program and process it's output in Python in near realtime.

I want to excute the command (this runs my makefile for my C program:

make run_pci

In:

/home/ecorbett/hello_world_pthread

I also want to process the output from the C program in Python. I'm new to Python so example code would be greatly appreciated!

Thanks, Edward

Answer 1

Use the cwd argument to Popen, call, check_call or check_output. E.g.

subprocess.check_output(["make", "run_pci"],
 cwd="/home/ecorbett/hello_world_pthread")

Edit: ok, now I understand what you mean by "near realtime". Try

p = subprocess.Popen(["make", "run_pci"],
 stdout=subprocess.PIPE,
 cwd="/home/ecorbett/hello_world_pthread")
for ln in p.stdout:
 # do processing

• NASA Intern: Thanks! Now how would I capture output from the c program? It simply printfs multiplule lines of text?
• Acorn: If you read the documentation you will see that check_output returns the output of the command. docs.python.org/library/subprocess.html#subprocess.check_output
• NASA Intern: I see. so doing something like this: process = subprocess.check_output(["make","run_pci"],cwd="/home/ecorbett/hello_world_pthread") print process.communicate() Would be unnecessary/incorrect?
• NASA Intern: But back to my original question. If my C program will be running, processing data and producing output that I need to process with my python program simultaneously, will this method still work?
• Fred Foo: @NASAIntern: no, then you need to use Popen instead of check_output. Pass it stdout=subprocess.PIPE, then read from its stdout member.

Answer 2

Read the documentation. You can specify it with the cwd argument, otherwise it uses the current directory of the parent (i.e., the script running the subprocess).