These are the original files. The following commands can be used to produce the 
500-megabyte parts used in the experiments.

  bunzip2 -c fiwiki.bz2 | split_wikipedia.py - 500 revision

  bunzip2 -c enwiki.bz2 | split_wikipedia.py - 500 page

Program split_wikipedia.py can be found in the RLCSA package.