This is the script used to get the links out of Two Scoops of Django 1.6 LaTeX files. It's been modified to generate HTML instead of a new appendix. It currently has a minor problem with duplicates, but as the rendered content is not going into the book, I'm not worrying about it.

getlinks.py

import glob
import re

# Regex to find all the links defined in LaTeX
regex = re.compile(r'\\link{.*}', re.IGNORECASE)


def strip_http_things(link):
    """Used to clean links for better sorting and for 'cleaner' 
        final display:

        e.g. http://www.djangopackages.com is djangopackages.com
        """
    link = link.replace("https://", "").replace("http://", "")
    return link.replace("www.", "")


def cleaned_url_comparison(a, b):
    """This custom sort ensures that the domain is the leading
        sort, not the various protocol prefixes.

        http://docs.python.org/2/library/functions.html#sorted
    """
    a = strip_http_things(a)
    b = strip_http_things(b)

    if a > b:
        return 1
    if a == b:
        return 0
    return -1


def main():
    tmp = []  # temporary holder for links
    links = []  # results list

    # Find all the LaTeX files in the chapters directory
    for filename in glob.glob("chapters/*.tex"):

        # Open the LaTeX file
        with open(filename) as f:
            text = f.read()

        # Find all the links in the LaTeX file and add them to tmp
        tmp += re.findall(regex, text)

    # Loops through all the links in tmp and do light cleanup
    for link in tmp:
        if ':8000' not in link:  # Ignore the links to port 8000
            # Remove the unnecessary 'www.'
            # Nearly duplicates strip_http_things()
            link = link.replace("www.2scoops.co", "2scoops.co")
            # Lazy way to correct an intermittent bug in my regex.
            links.append(link[0:link.index('}') + 1])

    # Remove MOST of the duplicates
    links = list(set(links))

    # Sort the links. See cleaned_url_comparison() above.
    links = sorted(links, cmp=cleaned_url_comparison)

    return links


if __name__ == "__main__":
    # Context manager! Yeah!
    with open("links.html", "w") as f:

        # Check out my old-school string handling!
        text = "<ul>\n"

        for link in main():
            # strip LaTeX elements
            link = link.replace("\link{", "").replace("}", "")

            # Add to the unordered list
            text += """  <li><a href="{}">{}</a></li>\n""".format(
                link,
                strip_http_things(link)
            )

        # Bare metal rockstar method of closing out a string.
        text += "</ul>\n"
        f.write(text)