backup_tumblr.py

#!/usr/bin/env python
import urllib, os, time, glob, urlparse, errno, sys, getpass, re
from xml.dom.minidom import parseString as parse

API_PATH = '/api/read'
MAX_NUM_PER_REQUEST = 50

class BadLogin(Exception): pass

def make_api_url(tumblr):
    return 'http://' + tumblr + '.tumblr.com' + API_PATH

def read_from_api(tumblr, email, password, start=0, num=0):
    params = {
        'email': email,
        'password': password,
        'start': start,
        'num': num
    }
    url = make_api_url(tumblr) + '?' + urllib.urlencode(params)
    f = urllib.urlopen(url)
    try:
        rsp = f.read()
        if rsp == 'Invalid credentials.':
            raise BadLogin()
        return rsp
    finally:
        f.close()

def get_total_num_posts(tumblr, email, password):
    dom = parse(read_from_api(tumblr, email, password))
    return int(dom.getElementsByTagName('posts')[0].getAttribute('total'))

def mkdir_p(path):
    try:
        os.makedirs(path)
    except OSError, e:
        if e.errno != errno.EEXIST:
            raise

def backup(tumblr, email, password, output_dir='.', show_progress=True):
    if not output_dir.startswith('.'):
        mkdir_p(output_dir)
    i = 0
    n = MAX_NUM_PER_REQUEST
    total = get_total_num_posts(tumblr, email, password)
    while i < total:
        f = open(os.path.join(output_dir, '%s_tumblr_%s-%s.xml' % (tumblr, i, i+n-1)), 'w')
        try:
            f.write(read_from_api(tumblr, email, password, i, n))
        finally:
            f.close()
        if show_progress:
            sys.stderr.write('.')
            sys.stderr.flush()
        i += n
        time.sleep(2)
    if total > 0:
        print >> sys.stderr
    return total

def usage():
    print >> sys.stderr, 'Usage: %s http://<you>.tumblr.com [output_dir]' % sys.argv[0]
    sys.exit(1)

if __name__ == '__main__':
    if len(sys.argv[1:]) < 1:
        usage()

    m = re.match('http://([\w-]+)\.tumblr\.com', sys.argv[1])
    if not m:
        usage()
    tumblr = m.group(1)

    if len(sys.argv[1:]) == 2:
        output_dir = sys.argv[2]
    else:
        output_dir = '.'

    email, password = None, None

    while not email:
        email = raw_input('Email address: ')

    while not password:
        password = getpass.getpass('Password: ')

    try:
        n = backup(tumblr, email, password, output_dir)
    except BadLogin:
        print >> sys.stderr, 'Your email address and password combination was no good. Try again.'
    else:
        print >> sys.stderr, 'Backed up %s tumblr posts.' % n