Find undocumented items in valadoc

#!/bin/bash

[ -z "$1" ] && echo "Usage: $0 "'/directory/with/valadoc-generated-html/**
  Processing several dirs at once is supported.
  Make sure you add ** after the names to make the shell list files recursively.'

( # start a subshell so we can grab all the output at once and pipe it
    ( # and another one, for deduplication this time
        for file in "$@"; do
            if (basename "$file" | grep -q ".html"); then
                sed -n -e '/div class="site_content"/,$p' "$file" \
                   | tr --delete '\n' \
                   | sed 's|</li>|&\n|g' \
                   | grep 'li class=' \
                   | grep -v 'class="deprecated"' \
                   | sed --regexp-extended 's|<div class="leaf_brief_description">[[:space:]]*</div>||' \
                   | grep -v 'brief_description'
            fi
        done
    ) | grep "href" | while read line; do
        line="${line##*<li}"
        type="$(echo "$line" | cut -d \" -f 2)"
        name="$(echo ${line#*href=} | cut -d \" -f 2)"
        echo "$type ${name%.htm*}"
    done
# Beware: Dark awk magic ahead!
# Source: http://www.unixcl.com/2008/03/remove-duplicates-without-sorting-file.html
) | grep -v 'virtual_method' | awk ' !x[$0]++' -