extract_keynote_slide_text.py

"Extracts the text from every slide in a keynote presentation"

# Limitation: doesn't currently deal with markup nested inside a p tag, e.g. 
# the following:
# <ns2:p ns2:list-level="1" ns2:style="SFWPParagraphStyle-990">Pen knife photo: 
# herzogbr      <ns2:link
#    href="http://www.flickr.com/photos/80516279@N00/2274372747/"><ns2:span 
#    ns2:style="SFWPCharacterStyle-158"
# >http://www.flickr.com/photos/80516279@N00/2274372747/</ns2:span></ns2:link>
# </ns2:p>

import zipfile
from xml.etree import ElementTree as ET

def extract_keynote_slide_text(fp):
    z = zipfile.ZipFile(fp, 'r')
    root = ET.fromstring(z.read('index.apxl'))
    slides = root.find(
        '{http://developer.apple.com/namespaces/keynote2}slide-list'
    )
    slide_texts = []
    for slide in slides:
        slide_text = []
        ps = slide.findall('.//{http://developer.apple.com/namespaces/sf}p')
        for p in ps:
            text = p.text
            if text is None or not text.strip():
                continue
            slide_text.append(text)
        slide_texts.append(slide_text)
    return slide_texts

Maybe a TextIterator helps with the nested markup, something like:
" ".join([t.strip() for t in ET.ElementTextIterator(ps)])