The most fundamental of abilities in shell scripts is moving about the file system.

import os
os.chdir('/')
contents = os.listdir(os.getcwd())

To change directories, we use os.chdir with an argument of our destination. To get a listing of the contents of a directory, os.listdir is the command to use. It takes one argument, which is the directory that holds the contents. If we want to know our current location on the filesystem, then os.getcwd will tell us the path.

Instead of moving around, you may want to do the same thing on the files and folders of a tree of directories, a.k.a. a recursive operation. In this next scenario, I have a hypothetical module called mp32vorbis that has a convert function, which converts a specified mp3 file into an ogg vorbis file.

import os
import mp32vorbis

for root, dirs, files in os.walk(os.getcwd()):
    for dir in dirs:
        print "Processing", os.path.join(root, dir)
    for file in files:
        if file.endswith('.mp3'):
            print "Converting", file
            mp32vorbis.convert(file)

As you can see, recursive operations are quick and easy with os.walk. It is a generator that yields each directory, root, plus a collection of directories and files within it. You can process each separate directory and file by looping through them. You may have noticed os.path.join. This is a simple function that connects its arguments with the directory separator of the host OS, which is backslashes for Windows and forward slashes for everything else.

So you are running programs from python now. Perhaps you want to check that a file is readable and executable before trusting it to run? The os module comes to the rescue once again!

import os
filename = 'bin/someprogram'
if not os.path.exists(filename):
    print filename, 'does not exist!'
    raise SystemExit, 2
if not os.access(filename, os.R_OK):
    print filename, 'is not readable!'
    raise SystemExit, 3
if not os.access(filename, os.X_OK):
    print filename, 'is not executable!'
    raise SystemExit, 4

We would like to know that the file exists. This is determined with os.path.exists. Is the file readable and executable? We can find out with os.access and a couple of constants from the os module, os.R_OK and os.X_OK. We could also test if a file is writeable with os.access. We just use the os.W_OK constant. Simple, yes?

Would you like to know the times of creation, modification, and access? What about file size?

import datetime
import os
filename = 'somedir/somefile'
ctime = datetime.datetime.fromtimestamp(os.path.getctime(filename))
mtime = datetime.datetime.fromtimestamp(os.path.getmtime(filename))
atime = datetime.datetime.fromtimestamp(os.path.getatime(filename))
sizeinbytes = os.path.getsize(filename)

It is pretty straight forward. You might be wondering about why I am using datetime.datetime.fromtimestamp here. The os.path.getXtime functions all spit out POSIX timestamps. In python, it is more useful to have these be python datetimes. The datetime.datetime.fromtimestamp function produces a datetime object from a given timestamp.

You might need to make folders and link files from time to time. The os module is used for both of these tasks.

import os
os.mkdir('newdir') # like mkdir
os.makedirs('notexistent/bogus/newdir') # like mkdir -p

os.mkdir and os.makedirs offer little in the way of surprises. os.mkdir works like mkdir and os.makedirs works like mkdir -p, creating any intermediate directories it needs to create the specified directory.

import os
os.link('file', 'newhardlink') # think ln
os.symlink('fileordir', 'newsymlink') # think ln -s

os.link creates hard links to files just like ln. os.symlink creates soft links in the way that ln -s does. I love how python works in such a predictable fashion.