here are the token definitions I've tried
$control: 1
lex:
{
define: WHT "[ \t]"
define: NL "{WHT}*\n"
token: WORD "[A-Za-z0-9]+" #word
token: WORD "[0-9]+\\.[0-9]+" #word
>> token: WORD "[##0-9]*" #word
>> token: WORD "[#0-9]*" #word
>> token: WORD "[\#0-9]*" #word
>> token: WORD "[\\#0-9]*" #word
token: EOS "[.?!]" #end of sentence
token: NEWLINE "{NL}" #single end-of-line
token: PARA "{NL}{NL}" #end of paragraph
token: WHITE "{WHT}" #whitespace
token: PUNCT "." #all other text
}