This is fascinating. I did a backup using tumblr-utils, it came out to about 3.6GB, it completed ever. This is not compressed, it’s just a bunch of HTML and images.
Tumblr’s exporter produced a zip file which is >5GB, and still running.
Comparing a blog exported by tumblr’s exporter to one exported by tumblr-utils, they are producing very similar results, with some images slightly different between them. Like, of three images on one blog, two were apparently identical, one changed size and I don’t know why.
What’s weird is that it’s not at all obvious why the export of this blog is over 5GB using tumblr’s exporter. Is it including more stuff? We can’t tell, because there’s no way to look at any part of the file unless the whole file makes it down, because zip files don’t have their primary table of contents until the end of the archive. But people are reporting corrupt zips. But intuitively, it shouldn’t be larger than the uncompressed data, right? Well, who knows. This is tumblr.
What is Tumblr-utlis and where can I find it, considering the fresh hell this blue nonsense is putting people through?
Python script to grab things. Grabbed my blog in ~24 minutes, contrast with it taking something like an hour for tumblr to decide it had created a file, and another hour to actually download it. Seems significantly cleaner, although honestly I suspect tumblr used the code from this one because of similarities in the generated output. So I think they used it and then modified it and broke it a bit.
Published
just moving along like everyone else 