unscroll

unscroll is a program to take a pdf file which is the "wrong shape" and output one that's a more useful shape to print with.

This was inspired by the files you create with some interactive whiteboard software, which are laid out on "a4" pages as follows.

      /--------------
      |              |
      |              |
      |    #####     |
      |    #####     |
      |    #####     |
      |    #####     |
      |    #####     |
      |    #####     |
      |    #####     |
      --------------/

This is useless for printing since everything's tiny and there's miles of whitespace everywhere. Fortunately, there are gaps in the black stuff, since it's actually lines of text.

unscroll searches for rectangular hunks of non-white and uses a simple word-wrap algorithm to distribute them among more pages, hopefully ending up with much better use of the available space.

After building the program, the simplest command line is something like

./src/unscroll data/squiggle.pdf out.pdf

This renders squiggle.pdf at 150 dpi and rejigs everything to fit nicely on your system's standard paper size (probably configured at /etc/papersize). It then renders the hunks with a jpeg quality setting of 75%.

To change the defaults, you can use a command like

./src/unscroll -d 200 -Q 30 -s a5 data/squiggle.pdf out.pdf

for 200 dpi, 30% quality and A5 paper.

TODO:

• Implement a "justification" algorithm to spread out hunks on the page more evenly.

• Deal with off-white data better e.g. scanned images. This will require the is_white stuff in page.c to be changed.

• Automatically guess the correct dpi? Is this even possible?

• Vary dpi between pages. This would work well if it was automatically determined. Could be messy as command line arguments.