User:Ixfd64/revision sizes

This is an R script for visualizing the byte sizes of Wikipedia pages over time. I originally wrote this for an assignment for my Stat 133 class, but I have decided to release it to the public. Other Wikipedians are welcome to improve it as they see fit.

Source code

# current version: 1.2.0
# last update: November 19, 2009

# load the Tcl/Tk library

require(tcltk)
 
# create main GUI

base = tktoplevel()
tkwm.title(base,'Wikipedia revision sizes')
 
# create main frame

nfrm = tkframe(base)
 
# create Tcl variables

revisions.tcl = tclVar('')
page.tcl = tclVar('')
 
# create "Page" text field

f1 = tkframe(nfrm)
tkpack(tklabel(f1, text = 'Page'), side = 'left')
tkpack(tkentry(f1, width = 25, textvariable = page.tcl), side = 'left')
 
# create "Revisions" text field

f2 = tkframe(nfrm)
tkpack(tklabel(f2, text = 'Revisions'), side = 'left')
tkpack(tkentry(f2, width = 8, textvariable = revisions.tcl), side = 'left')
 
# pack GUI elements

tkpack(f1, side='top')
tkpack(f2, side='top')
tkpack(nfrm)
 
# language code
# complete list available at http://meta.wikimedia.org/wiki/List_of_Wikipedias

lang = 'en'
 
# get page revision sizes

getpage = function(...) {
  revisions = as.numeric(tclvalue(revisions.tcl))
  page = as.character(tclvalue(page.tcl))
  page.loc = gsub(' ', '_', page) # fix space parsing
  ge = function(s, g) substring(s, g, g + attr(g, 'match.length') - 1)
  byte.pat = '<span class="history-size">(\\([0-9bytesmtp ,]+\\))</span>'
  historylink = paste('http://', lang, '.wikipedia.org/w/index.php?title=', page.loc, '&limit=', as.character(revisions), '&action=history', sep='')
  history = readLines(historylink)
  if (any(grep('no revision history', history))) {
    tkmessageBox(title = 'Error', message = 'This page does not exist!', type = 'ok')
} else {
  revisiondata = gsub(byte.pat, '\\1', mapply(ge, history[grep(byte.pat, history)], gregexpr(byte.pat, history[grep(byte.pat, history)], ignore.case = TRUE)), ignore.case = TRUE)
  revisiondata = as.character(revisiondata)
  revisiondata = gsub('[\\(\\), ]', '', revisiondata)
  for (a in 1:length(revisiondata)) {
    if (revisiondata[a] == 'empty') {revisiondata[a] = '0'}
  }
  revisiondata = gsub('[bytes]', '', revisiondata)
  revisiondata = as.numeric(revisiondata, ignore.na = TRUE)
  revisiondata = rev(revisiondata)
  plot(revisiondata, xlab = 'Revision index', ylab = paste('Revision size of \"', page, '\" (bytes)', sep = ''), type = 'l')
  }
}
 
# quit program

destroy = function(...) tkdestroy(base)
 
# create buttons

bfrm = tkframe(base)
tkpack(tkbutton(bfrm, text = 'Run', command = getpage), side = 'left')
tkpack(tkbutton(bfrm, text = 'Quit', command = destroy), side = 'right')
 
# pack bottom frame

tkpack(bfrm, side = 'bottom')

Usage

The last 500 revisions of Milky Way

This script requires R to be installed. The best way to use this script is to put it in a text file located in the directory of your R workspace. For example, if you named your file wpgraph.txt, you can run the script by entering source('wpgraph.txt') in the R console.

Limitations

• This script does not work for edits made before April 19, 2007. Such revisions do not indicate the byte size due to an older version of MediaWiki being used at the time. Since this script uses regular expressions to determine the byte sizes, it will not work for revisions that do not have the size indicator.
• Certain characters are not processed correctly. For example, "1+1" must be entered as 1%2B1.