nwi.flt

harvest是一个下载html网页得机器人

#
# Input-filter for the Nordic Web Index record syntax. Output is 'gils-like'.
#
# $Id: nwi.flt,v 1.4 1996/06/11 10:53:56 quinn Exp $
#

/<nwi>/                            { begin record gils }

# Ignore meta tags

/<meta.*>.*$/                    {}

/<ti> */ BODY /$/                  { data -element title $1 }

/<dm> */  BODY /$/                  { data -element dateOfLastModification $1 }

/<ci> */ BODY /$/                  { data -element controlIdentifier $1 }

/<si>/                           { begin element supplementalInformation }
/<lc> */ BODY /$/                  { data -element lastChecked $1 }
/<by> */ BODY /$/                  { data -element bytes $1 }

/<av>/                           { begin element availability }
/<avli> */ BODY /$/                { data -element linkage $1 }
/<ty> */ BODY /$/                  { data -element linkageType $1 }

/<lsi>/                          { begin element localSubjectIndex }
/<dh> */ BODY /$/                  { data -element localSubjectTerm $1 }

# Don't want to have <ip> inside of LocalSubjectIndex
# Since we end localsubjectindex, we consume the end-tag for that as well.

#/<ip>[ \n]*/ BODY /<\/ip>[ \n]*<\/lsi>/  {
#				    end element;
#				    data -element sampleText $1
#				 }

/<ip>[ \n]*/ BODY /<\/ip>/       {
				   end element;
                                   data -element sampleText $1
				 }

/<cr>/                           { begin element crossReference }
/<li> */ BODY /$/                  { data -element linkage $1 }
/<cp> */ BODY /$/                  { data -element title $1 }

/<\/nwi>/                        { end record }

# Generic end-marker

/<\/[^>]*>/                      { end element }
/\n/                             { }
/./ {}