Download and unpack a tarball from the internet with python.

download.py

# Author: Joseph Lisee <[email protected]>
# License: No rights reserved, code is in the public domain.
# Date: May 2013

__doc__ = """
Downloads and unpacks a tarball from the Internet.
"""

import argparse
import os
import os.path
import sys
import tarfile

from urllib2 import urlopen, URLError, HTTPError

def download_file(url, download_path = '.'):
    """
    Downloads a file, based on StackOverflow:
       http://stackoverflow.com/a/4028894/138948
    """

    success = False

    # Open the url
    try:
        f = urlopen(url)
        print "downloading " + url

        # Formulate the path to our local file
        local_path = os.path.join(download_path, os.path.basename(url))

        # Open our local file for writing
        with open(local_path, "wb") as local_file:
            local_file.write(f.read())

        success = True

    # Handle errors
    except HTTPError, e:
        print "HTTP Error:", e.code, url
    except URLError, e:
        print "URL Error:", e.reason, url

    return success

def unpack_tarball(tar_url, extract_path='.'):
    """
    Extracts a tar file to disk, return root directory (and assumes
    there is one).
    """

    print "extracting " + tar_url

    # Open and extract
    tar = tarfile.open(tar_url, 'r')

    tar.extractall(extract_path)

    # Get root (the shortest path will be the root if there is one)
    return sorted(tar.getnames())[0]

def download_and_unpack(url, output_path):
    """
    Downloads and unpacks the given tarball, and returns you it's root.
    """
    # Download file
    base_path, tar_filename = os.path.split(url)

    download_file(url, output_path)

    # Unpack file
    tar_path = os.path.join(output_path, tar_filename)
    unpack_dir = unpack_tarball(tar_path, output_path)

    return os.path.join(output_path, unpack_dir)

def main(argv = None):
    if argv is None:
        argv = sys.argv

    # Parse arguments
    parser = argparse.ArgumentParser(description='Process some integers.')
    parser.add_argument('url', metavar='URL', type=str,
                        help='File to download and unpack')
    parser.add_argument('--output', dest='output', default=".",
                        help='output directory')

    args = parser.parse_args(argv[1:])

    # Make output directory if needed
    if args.output is not None and not os.path.exists(args.output):
        os.makedirs(args.output)

    # Do our work!
    print download_and_unpack(args.url, args.output)


if __name__ == '__main__':
    sys.exit(main())