#!/usr/bin/perl -w

# simple_chunk.pl
# does (very) simple noun phrase chunking

# input:
# many_JJ different_JJ states_NNS have_VBP appealed_VBN for_IN a_DT more_JJR widespread_JJ use_NN of_IN federal_JJ law_NN in_IN alcohol_NN abuse_NN cases_NNS ._. 

# output:
# states_NNS have_VBP appealed_VBN for_IN a_DT more_JJR use_NN of_IN law_NN in_IN cases_NNS ._. 

while (<>)
{
    # chunk NP's,
    s/([^ ]+_DT )?([^ ]+_(NNP|NN|NNS|NNPS|JJ|JJS|CD) )*([^ ]+_(NNP|NN|NNS|NNPS))/$4/g;
    
    # chunk QP's
    s/\$_\$ ([^ ]+_CD )*([^ ]+_CD)/$2/g;
    print $_;

}