#!/usr/bin/env python2.5
import hashlib
import socket
socket.setdefaulttimeout(5)
import os
import urllib2
import xml.etree.cElementTree as ET

def makedirs():
    try:
        for path in ['xml', 'html']:
            os.mkdir(os.path.join(os.path.dirname(__file__), path))
    except OSError:
        print "store dirs already in place"
        pass
        
def parse_opml(filename):
    tree = ET.parse(filename)
    outlines = tree.findall("//outline")
    for outline in outlines:
        print "Processing %s" % (outline.get('text') or outline.get('htmlUrl'))
        for path in ['xml', 'html']:
            url = outline.get('%sUrl' % path)
            if url is None:
                continue
            if not url.startswith('http://'):
                url = "http://%s" % url
            
            url_hash = hashlib.md5(url).hexdigest()
            file_path = os.path.join(path, url_hash)
            if os.path.exists(file_path):
                continue
            
            try:
                print "\tdownloading: %s ..." % url,
                page = urllib2.urlopen(url).read()
            except (urllib2.HTTPError, urllib2.URLError):
                continue

            f = open(os.path.join(path, url_hash), 'w+')
            f.write(page)
            f.close()
            
            print "\tdone."

def main(opml_file):
    makedirs()
    parse_opml(opml_file)


if __name__ == '__main__':
    import sys
    main(sys.argv[1])