Use html-xml-utils like grep with CSS selectors

cssgrep.sh

#!/bin/bash

if ! which hxselect > /dev/null; then
  echo You need hxselect and hxnormalize
  echo Run this: sudo apt-get install html-xml-utils
  exit 1
fi

hxselect="hxselect -s \n -i"
hxnormalize="hxnormalize -l 10000000 -x "

while [ $# -gt 0 ]; do
    case "$1" in
        -c) 
	  hxselect_args="-c"
	  shift
	  ;;
	*)
	  selector=$1
	  shift
	  break
	  ;;
    esac
done

if [ ! -z $2 ] ; then
  for filename in "$@" ; do
    results=$($hxnormalize "$filename" | $hxselect $hxselect_args "$selector")
    if [ ! -z "$results" ] ; then
      for r in "$results" ; do
        echo "$filename: $r"
      done
    fi
  done
elif [ ! -z $1 ] ; then
  filename=$1
  $hxnormalize "$filename" | $hxselect $hxselect_args "$selector"
else
  $hxnormalize | $hxselect $hxselect_args "$selector"
fi