Man.. TV is such a good name for a visualiser tool that it'd be excellent if it could pretty-print any content given to it.
Does your framework support the idea of pre-parsing the file's content and selecting an appropriate renderer, or is it fairly tied to CSVs?