I am fan of xpath when parsing and extracting data from html file.
Currently, dyer do not support html paring, you have to implement the parser on your own.
I am considering add a xpath feature to the response for better integration with C library libxml2.