blogger_import.py

import argparse
import xml.etree.cElementTree as ET
import os


def is_post(node):
    t = node.find('{http://www.w3.org/2005/Atom}category').get('term')
    d = node.find('{http://purl.org/atom/app#}control')
    return (not d) and (t == 'http://schemas.google.com/blogger/2008/kind#post')

def get_posts_id(blogger_dump):
    title_map = {}
    tree = ET.parse(blogger_dump)
    elem = tree.getroot()
    feed = elem.findall('{http://www.w3.org/2005/Atom}entry')
    for f in filter(is_post, feed):
        post_id = f.find('{http://www.w3.org/2005/Atom}id')
        id_number = post_id.text.split('-')[2]
        post_title = f.find('{http://www.w3.org/2005/Atom}title').text
        title_map[post_title] = id_number
    return title_map

def inject_redirect(file_name, title_map):
    yaml_separator = 0
    old_file = file_name + 'old'
    print file_name
    print old_file
    os.rename(file_name, old_file)
    f = open(old_file, 'r')
    n = open(file_name, 'w+')
    for line in f:
        if line == '---':
            yaml_separator = yaml_separator + 1
        if yaml_separator == 2:
            n.write(line)
            continue
        if line.startswith('title'):
            title = line.split('"')[1]
            try:
                post_id = title_map[title]
                n.write('redirect_from:\n')
                n.write('  - /blog/%s/\n'%post_id)
            except KeyError:
                print 'id not found for ' + title
        n.write(line)

def set_redirect(posts_path, titles_map):
    for file_name in os.listdir(posts_path):
        if (file_name.endswith('html')):
            inject_redirect(os.path.join(posts_path, file_name), titles_map)

if __name__ == '__main__':
    parser = argparse.ArgumentParser()
    parser.add_argument('-p', '--posts', dest='posts_path', help='path of the converted _posts')
    parser.add_argument('-b', '--blogger', dest='blogger_dump', help='blogger dump file')
    args = parser.parse_args()

    titles_map = get_posts_id(args.blogger_dump)
    set_redirect(args.posts_path, titles_map)