Tumblr’s users who wish to move their content over to a self-hosted WordPress blog now have a new solution. WordPress developer Pete Mall used the existing exporter PHP scripts written by Hao Chen and its “enhancement fork” created by Ben Ward to create a new “Tumblr 2 WordPress” exporter: http://tumblr2wp.com/.
[UPDATE – August 2, 2012] I disabled the link to Hao Chen’s blog because it was listed as suspicious by Google. One can check the actual status of this website by going to haochen.me/tumblr/
This script differs from its previous iteration in at least two ways. First, it’s able to generate titles for most post types created in Tumblr (be it “quote”, “text”, “photo” etc. I haven’t tested all of them). It’s one of the biggest challenge one faces while trying to move away from Tumblr to WordPress. Most of Tumblr’s post types don’t make use of title within Tumblr. Pete Mall’s script generates titles on the fly by selecting a certain amount of characters or words at the beginning of each post.
Unfortunately, there’s no way to control the title extraction behavior (that is: to control the parameters for the extraction). Which brings me to the second aspect: unless I’m mistaking, Pete Mall’s script isn’t open source. It’s was created for WooThemes, a developer of premium, paying WordPress themes, most likely as a marketing tool. It’s designed –among other things– to bring new customers to buy a “tumblog” theme for WordPress.
As for media, I really don’t know. It’s another big issue when moving from Tumblr to WordPress: one doesn’t want his images to be hot-linked to Tumblr’s servers. If you’re running your own self-hosted WordPress installation, you’ll most likely want to store images on your own server. According to a Mashable article about the WooThemes exporter, one should be able to import images… to a WooThemes WordPress theme. I’ve try to “download and import file attachments” using WordPress’ import tool with a non-WooThemes theme: it failed. I wasn’t able to import any images to my WordPress installation. It’s no big deal since there’s a workaround for this specific problem: after the importation, one could install and run the Hot Linked Image Cacher plugin for WordPress.
In the end, I guess the WooThemes exporter (as we may call it) could be an interesting solution for two kinds of people: 1) Those who are already making use of a WooThemes WordPress theme or who won’t mind paying for such a theme and 2) Those who won’t mind not being able to precisely control how the exporter formats their posts.
Personally, I’ll wait for version 0.5 of Ben Ward’s exporter (which is now version 0.4).
For more about the issues related to moving from Tumblr to WordPress, one could read this piece on Stack Overflow: “Customizing a script to extract titles from Tumblr post (stripping HMTL tags from text input)” (it’s an inquiry about how to code a proper title extractor in PHP). Also, there’s a discussion about “How to import Tumblr blog into WordPress” on Peter’s blog Blog Tips (read the comments as well).