It has this crappy DRM that only lets me view the pdf on one computer using only “Adobe Digital Editions”.

If that wasn’t so bad, only a small subset of the text is OCR’d, so most of it isn’t even searchable!

Now I’m pissed, but wait, what do you say? These files are just RSA encrypted, and I have the key?

Some cool guy named i♥cabbages has released code do extract your key, and then decrypt the file to a good ol’ plain pdf. If you want to reproduce my steps you will need to use the PDF decrypter unless you have epubs.

So I use the tool and get a pdf, now I can use one of the most awesome tools in the world: Imagemagick.

Imagemagick can whip this pdf into shape. The first thing I’m going to do is convert each page into a tiff:

$ convert -density 200 input.pdf[1-124] -depth 8 -monochrome %05d.tif

Then I’m going to run tesseract-ocr on them to get the text:

$ for i in $(seq –format=%005.f 1 324)
do
tesseract $i.tif tesseract-$i -l eng
done

Now all I have to do is cat all the text together:

cat *.txt > output.txt

Now I have a fully searchable, plain text file. Exactly what I wanted in the first place!

For the REAL magic, I use agrep to search for strings similar to provided example test questions to help “highlight” the answers. More technical details on that magic on my wiki.