Posts

@DBot  
I was able to get everything from ‘__tumblr3’, but the other two I have tried the last three weekends to no avail. Not sure if there is something going on or not, but I thought I’d let you know about it. I was excited to see the torrents working again, but I’m hoping I can end up with the others as well. Thanks for all your work so far!

@DBot  

 
So does this mean there are new files which are not in the original three torrents?

According to security logs clients aggressively re-download same files over and over in despair. This should not be the case now with these new torrent files.

@DBot  
If you’re seeing new traffic on them right now, it’s because I’m moving them to my new host. I could never get 100% on #1 and #2 before, maybe it’ll work now.  
But the files can be handled by high performance clients, definitely by transmission, qbittorrent and rutorrent might also work.

It appears that most of users that try to download files using my torrent files struggle to download backups.
 
So here is better way to download them:  
This archive contain all tumblrs torrent files separately (the __tumblr/nsfw, __tumblr2/story and __tumblr3/mixed which didn’t made into top of this thread), so all of these torrents should be handled by any torrent client. (they were created using Linux mktorrent program)
 
If anyone gonna download my torrents (or struggle with old ones) they should use link above and not old ones, which completely overwhelm any torrent ever created.
 
This is a txt file containing every single file (60 mb filesize warning, can hang your browser), it is intended for those who don’t want to use bittorrent client and regular HTTP client, such as curl.

@Darkenetor  
Dropbox is just an intermediary for the day or so until the files are updated here, then I’ll delete them anyway to make space for new images. It’s what I have, and what I know is free, without compression and offers static links.
 
For the API, I guess there’s no other way than to try it out. Making new accounts is extremely easy though, so if you export regularly you should get a decent output.  
You could also work with “width:1280, source_url:tumblr.com, -dead source, my:faves” to get your fav pics first before going for everything else.

Searching in the dashboard could be automated but I was thinking of reblogging everything and hoping the export won’t explode. I’d think post APIs to be much more limited and controlled than lookups.
 
(And I’d suggest against using the Dropbox trick if you’re also using it to horlink, it’ll go down and leave dead links eventually. The Dropbox-website hosts in the article are better but Github and Gitlab are the safest options IMO)

@Darkenetor  
A few good points there, 1280 is the most common size but from what I remember indeed not the only one.
 
Bots might work with non-flagged art but I’m not sure how you could automate the “searching in the dashboard” part. Maybe the API allows an easy workaround though, somebody more talented would have to check.
 
As for being banned, my API key was never banned even when I scraped terabytes worth of blogs so I wouldn’t be too worried about that. As for artists getting in trouble, who would really notice or care in the first place, especially since we’re only talking about a handful of reblogs at most.
 
One other important thing is uploading the resulting images. The best way to do that (that I’m aware of) is hosting the images on some 3rd-party site that does not apply compression again, then link the images to this thread in the mentioned format.  
This way the files themself can be replaced by their better versions without deleting/bumping the original posts, automatically keeping artist uploads etc. and making mass replacement much easier.  
I currently use Dropbox for hosting, here is a quick guide on how to host images with a static link.

I’d never have expected this to work and keep being even more confused on what they’re trying to accomplish there but great find.
 
“How long” for less than 200k posts (lower fixed resolutions than 1280 are still worth getting since they’re really badly resized) if we just prioritize well known art posted here, not much, especially if splitting work.
 
More importantly will we get our dump accounts banned before the export and can this get artists in trouble with accusations of running the bots themselves to boost stats? Anybody more familiar with abusing post API or running bots there?

Posting here since it seems appropriate.
 
As @Barhandar mentioned many months ago there’s a new way to get raw files from tumblr since the old way was blocked.  
It’s simple, you reblog the file(s) you want, then use the export tool to get a full copy of your account and inside is the full res file, not the downsized 1280p version. This also works with restricted blogs which you can only see in your dashboard. The difference now is that the tumblr servers aren’t under enormous stress anymore and the backup is ready very fast.
 
As an example, let’s take >>1809161 (merged) (NSFW and foalcon, sorry but it was the first one I found).  
It’s currently a 1280x943 PNG with 565 kB.  
Now clicking on the source link brings up the “flagged” message since it’s NSFW but we can still view it by clicking the dashboard link and then scrolling down until we find it. Note that these steps are not necessary if the image was not flagged.  
Once we’ve found the post we can reblog it, then go into the account settings and export the whole account.  
In the resulting zip file is the original PNG with a resolution of 4005x2950 and a size of 3.71 MB.
 
Searching through Derpi with “width:1280, source_url:tumblr.com, -dead source” currently brings up 73548 results. Many of them have untagged dead links, especially since a lot of artists deleted their accounts after the recent events. Then there are files that simply are 1280p wide originally and artists with DNPs forbidding higher res versions. However that still leaves a lot of potential images that can be replaced with better versions.  
The question is, how long?
 
So if anyone here has the time, this is a great way to preserve high quality fanart in the face of G4’s death.  
I for one have tested this with Sidekick’s art and it worked great, even more so since a lot of his NSFW art wasn’t found by the Tumblr bot yet.

Welp I’m not really sure, if it’s the right place, but I accidentally found a way to bypass “This post may content adult content and now hidden” (kinda):  
questionable example  
Left — without bypassing; Right — with it.  
 
It only works, when you view a blog through your dashboard. It doesn’t reveal fully hidden posts.
 
How to do bypass on Chrome:  

@CMC Scootaloo  
It’s relatively easy to make standard UI for dumped blog which looks like archive page, because TumblThree seems to save all post data (if you checked it ofc).
 
Recreating custom theme sometimes may be almost impossible and there is no real reason to do it, since theme itself doesn’t contain something useful (and in most cases custom themes is so awful and unusable, that I often used archive page).

@Zanec
 
And do you happen to know how? I could go on longwinded Google searches for this, but I prefer to ask here first, in case anyone here has more knowledge about TumblThree.

@CMC Scootaloo  

 
It’s definitely possible. Not easy though, I’d imagine.

@ElectricGears
 
I am still in this, too, even though I have been quiet lately. There is still a good number of Tumblrs I want to save, because the new automated system might still axe an inactive blog that isn’t against the rules, if it hasn’t already.  
I will upload all I have, too, just want to collect a bit more. And the end game is pretty much that, uploading it all to Derpibooru. There is little use to the archiving efforts if the archived material stays hidden on the HDs of the archivists and if no one else than them can access it.  
Not being able to is not so obvious as you think. You would drop dead rather soon if you would try to upload all of this at once, but that isn’t necessary. Set aside a certain amount of time each day and upload as much as you can during that time. That is going to take a while, but eventually, the pile will all be up here.
 
Other than that, and I asked this before, is it possible to re-create an entire Tumblr blog, with design, layout and posts? TumblThree downloads a lot more than just the pictures themselves and I guess all these files have to be good for something.

I guess it would be a good time to ask what the end-game is here. I have about 2600 blogs totaling about 9TB captured with TumblThree. Obviously I’m not going to be able to publicly host this or manually upload everything. Plus I ended up with a ton of non-pony furry stuff. I’m wondering what everyone else’s plans are for the project. It might make sense to have some kind of special tag tumblrSuicide? for these uploads in case there is a need to sort them out in the future.

So, I’ve been slowly working through my 4tb dump and found one very interesting thing. Nor sure if it’s a common knowledge already but meh. Blog meta object contains a “header_image” field. This is the link to a raw image that was uploaded bu blog owner to be used as the blog’s header image. I just checked multiple blogs manually and found at least one where the header image is a raw image of the other art that was posted and uploaded here.
 
Manual checking using DOG technique revealed that the header image is not an upscale though I doubt my own eyes as of late and if someone has a better method of upscaling detection, I’d be glad to have this “raw” header image checked. You can find both images here.

If anyone still needs someone seeding, just say so here.

@imsesaok  
You’re welcome! I already downloaded xrandomxnessx.tumblr.com via tumblone and the wayback machine (web.archive.org)

@imsesaok  
i’m the one who is downloading. i will reseed it for a while. maybe some other people will join too.

I’m quite late to the party but better than never I guess.
 
I only backed up a handful of blogs(flashytheflashpone, lethanvas, milkmare-of-trottingham, nastylittlepest, rarecandy19, sosoftandtender, xrandomxnessx) but it ain’t small (12GB.) Download link for the torrent is here.
 
I’m in a bad position because I have to run my laptop to seed. I’ll keep it up as long as I can, but if no activity occurs until Saturday, I would not be seeding for the next 3 days when I’m away for the lunar new year.
 
EDIT: Somebody’s been downloading for a while, already in 10%. I’ll keep the thing running and hopefully one or more people gets a full copy. Thanks!

Anyone know the current site traffic of Tumblr since the Snap?  
I don’t know where to find this stuff.

@Lisboeta  
Jan 5 I believe.

@The Smiling Pony  
Pardon me if this is a stupid question, but when exactly they patched it? I archived quite a few things before the deadline, but then seen that access was still possible and kinda relaxed and dropped from the archiving.

So, I think my Tumblr dumping queue just finished getting downloaded, which is supposed to be a good thing considering I’ve fed the whole stage2 list to the dumper. Now I only need to find out how good or bad the stuff I have is and how I can compress or store the stuff efficiently (>IMPLYING not on my server). Thought about using S3 at first but it costs a bloody arm and a leg to use FML.
 
All Tumblr dumps combined weight 4144GB…
 
@Xaxu-Slyph  
My nickname reads TWeaKeR and yeah, after 2010 people tend to misspell it very frequently. As for the archive.org dumper’s speed fluctuations, it can be almost anything ranging from the flaky channel width of archive.org servers to an improper implementation of concurrency in the dumper itself. I skimmed through the code and it seems fine at a first glance but that doesn’t mean it’s completely error-free.