Tumblr migration script to move content from a blog to another

scribe.py

import pytumblr
import yaml
import os
import urllib

# stolen from interactive_console.py, because I'm too lazy to oauth
# you'll need to register an app with Tumblr and use interactive_console.py for initial oauth
def create_client():
    yaml_path = os.path.expanduser('~') + '/.tumblr'

    if not os.path.exists(yaml_path):
        raise Exception('Need to authenticate, use interactive_console.py')
    else:
        yaml_file = open(yaml_path, "r")
        tokens = yaml.safe_load(yaml_file)
        yaml_file.close()

    client = pytumblr.TumblrRestClient(
        tokens['consumer_key'],
        tokens['consumer_secret'],
        tokens['oauth_token'],
        tokens['oauth_token_secret']
    )

    return client

# download all posts of a given type, without callers having to worry about the offset+limit
def get_posts(blog, post_type):
    offset = 0

    while True:
        posts = client.posts(blog, post_type, offset=offset)["posts"]
        if not posts:
            break

        for post in posts:
            yield post
        offset += len(posts)

# cleanup blog (very useful while I'm writing this script and figuring out where posting fails)
def cleanup(blog, post_type):
    print("Cleaning up %s posts on %s" % (post_type, blog))
    new_posts = []
    for post in get_posts(blog, post_type):
        new_posts.append(post)
    for post in new_posts:
        client.delete_post(new_blog, post["id"])
    print("Done")

# retrieve relevant data from all posts of given type 
def post_list(blog, post_type, keys):
    post_list = []
    for post in get_posts(blog, post_type):
        post_data = dict()
        for key in keys:
            post_data[key] = post[key]
            try:
                post_data[key] = post_data[key].encode("utf-8")
            except:
                pass
        post_list.append(post_data)

    return post_list

client = create_client()

old_blog = "your-old-blog"
new_blog = "your-new-blog"

# migrate text posts
for post in post_list(old_blog, "text", ["title", "body", "date", "tags"]):
    print(client.create_text(new_blog, title=post["title"], body=post["body"], date=post["date"], tags=post["tags"]))

# migrate quote posts
for post in post_list(old_blog, "quote", ["text", "source", "date", "tags"]):
    print(client.create_quote(new_blog, quote=post["text"], source=post["source"], date=post["date"], tags=post["tags"]))

# migrate audio posts
for post in post_list(old_blog, "audio", ["caption", "audio_url", "date", "tags"]):
    print(client.create_audio(new_blog, caption=post["caption"], external_url=post["audio_url"], date=post["date"], tags=post["tags"]))

# migrate video posts
for post in post_list(old_blog, "video", ["caption", "player", "date", "tags"]):
    print(client.create_video(new_blog, caption=post["caption"], embed=post["player"][0]["embed_code"], date=post["date"], tags=post["tags"]))

# migrate link posts
for post in post_list(old_blog, "link", ["title", "url", "description", "date", "tags"]):
    print(client.create_link(new_blog, title=post["title"], url=post["url"], description=post["description"], date=post["date"], tags=post["tags"]))

# migrate link posts
for i, post in enumerate(post_list(old_blog, "photo", ["photos", "caption", "date", "tags"])):
    # hacky way of downloading all images and re-uploading them
    img_path = "./images/%d.jpg" % (i,)
    if not os.path.exists(img_path):
        url = post[0][0]["original_size"]["url"]
        print("Downloading %s" % (url,))
        urllib.urlretrieve(url, img_path)
    print(client.create_photo(new_blog, caption=post["caption"], data=img_path, date=post["date"], tags=post["tags"]))

# note I didn't have any chat posts to worry about