PDF 2016

From Archiveteam
Jump to navigation Jump to search
PDF 2016
Status Online!
Archiving status Unknown
Archiving type Unknown
Project source citeseerxpdf-grab
Project tracker citeseerxpdf
IRC channel #archiveteam-bs (on hackint)
(formerly #pdflush (on EFnet))

PDF (Portable Document Format) is a file format used to present documents in a manner independent of application software, hardware, and operating systems.[1]

PDF 2016 is a codename for an ArchiveTeam project, saving a lot of PDFs.


In March 2016, user davidar informed ArchiveTeam on IRC that he obtained a list of hundreds of millions of links to PDF files from around the Web.[2][3][4] ArchiveTeam decided to make a Warrior project for downloading these files.

Most of the files are open-access scientific documents. Besides being uploaded to the Internet Archive, http://citeseerx.ist.psu.edu/ will also host and index them.


ArchiveTeam is first saving about 1 terabyte of files, then the Internet Archive decides whether they are able to store all downloadable stuff, that is going to be tens or hundreds of terabytes. (The test run also helps making a good estimate on the total size.)

We have 250-300M URLs: ~10% of those are direct links to PDFs, the rest are mostly links to HTML landing pages. 10-15% have a citation_pdf_url meta tag giving a direct link to a PDF and another 10-15% have the PDF linked in an unstructured way.

How can I help?

Running a Warrior

You can start up a Warrior and there select PDF 2016. (If you don't really care what you are archiving, select ArchiveTeam's Choice instead, as at some points ArchiveTeam may prioritize another project.)

Running the script manually

If you use Linux and you're a bit familiar with it, you can try running the script directly.

The instructions can be found at github.com/ArchiveTeam/citeseerxpdf-grab.

Some additional information
Don't forget to replace YOURNICKHERE with your nickname.

The number after --concurrent determines how many threads run at the same time. You can increase this number if your resources (RAM, CPU, bandwidth) are sufficient. However, if you constantly see messages about rate limiting, there is no need to increase the concurrency.

If you want to stop the script, please do it gracefully if possible. To do so, create an empty file named STOP in the folder of the script (terminal command: touch STOP). The script finishes the current item(s) and stops only after that. (If you kill the script immediately, the items get broken, and they will need to be reassigned to another user.) – Before starting the script again, don't forget to remove the STOP file.

If you see "Project code is out of date", kill the script, go to its folder (cd citeseerxpdf-grab) and issue git pull https://github.com/ArchiveTeam/citeseerxpdf-grab. After the updating has finished, re-launch the script.

Donating to the Internet Archive

Content downloaded by the ArchiveTeam will be uploaded to the Internet Archive, where it will be stored and be available – hopefully – forever. However, storing it costs thousands of dollars in the long run. So, if you can afford, please consider donating to the Internet Archive, so that this piece of history can be kept for us all. http://archive.org/donate

Do you like our cause?

If you want to help in other projects, want to learn more about ArchiveTeam, or even help in development in general, navigate to the Main Page of this wiki, from there you can reach a lot of information. The Team consists of volunteers working on the projects in their free time, so helping hands (and resources) are always welcome.