How do I 'unbook' a pdf

Question

I have a pdf file that was made into a booklet format. It is meant to print on A4 paper in landscape orientation; here are two pages in the pdf, which should correspond to four pages in the actual book.

 -------------------------------
|              |                |
|              |                |
|              |                |
|    (1)       |      (3)       |
|              |                |
|              |                |
|              |                |
 ------------------------------- 

 -------------------------------
|              |                |
|              |                |
|              |                |
|    (4)       |      (2)       |
|              |                |
|              |                |
|              |                |
 -------------------------------

The bracketed numbers correspond to the order of the individual pages.

I know that there are all sorts of commands (pdfbook, pdfnup, etc.) on linux (which was probably used to make this booklet in the first place). How do I 'unbook' it---that is, I'd like to make a pdf document from this where each individual page of the final product is an individual page of the pdf, ordered in the usual way.

Edit

Thanks to Gilles, I managed to use the following code:

import copy, sys
from pyPdf import PdfFileWriter, PdfFileReader
input = PdfFileReader(sys.stdin)
output = PdfFileWriter()
second_half = []
for i in range(0,input.getNumPages()):
    p = input.getPage(i)
    q = copy.copy(p)

    print >> sys.stderr, 'splitting page',i
    print >> sys.stderr, '\tlowerLeft:',p.mediaBox.lowerLeft
    print >> sys.stderr, '\tupperRight:',p.mediaBox.upperRight

    p.mediaBox.upperRight = (ur[0], (bl[1]+ur[1])/2)
    p.mediaBox.lowerLeft = bl

    q.mediaBox.upperRight = ur
    q.mediaBox.lowerLeft = (bl[0], (bl[1]+ur[1])/2)

    if i % 2 == 0:
        output.addPage(p)
        qold = q
    else:
        output.addPage(q)
        output.addPage(qold)
        output.addPage(p)
output.write(sys.stdout)

Answer 1

Here's a small Python script using the PyPdf library that does the job. (Derived from un2up.) Save it in a script called unbook, make it executable (chmod +x unbook), and run it as a filter (unbook <book.pdf >1up.pdf).

I tested this script on the output of pdfbook --signature=N. For another method, you may not need to reverse every other input page, and the page order may be different (it depends on the orientation of the landscape pages). The page disposition doesn't match up with your question; the 13,42 doesn't make sense to me (in a 4-page book, 3 should be next to 2, not next to 1).

#!/usr/bin/env python
import copy, sys
from pyPdf import PdfFileWriter, PdfFileReader
input = PdfFileReader(sys.stdin)
output = PdfFileWriter()
second_half = []
for i in range(0,input.getNumPages()):
    p = input.getPage(i)
    q = copy.copy(p)
    (w, h) = p.mediaBox.upperRight
    p.mediaBox.upperRight = (w/2, h)
    q.mediaBox.upperLeft = (w/2, h)
    if i % 2 == 0:
        p.rotateClockwise(180)
        q.rotateClockwise(180)
        output.addPage(p)
        second_half.append(q)
    else:
        output.addPage(p)
        second_half.append(q)
second_half.reverse()
for q in second_half: output.addPage(q)
output.write(sys.stdout)

_{Ignore any deprecation warning; only the PyPdf maintainers need be concerned with these.}