#!/usr/bin/perl -w # # *** getwebref v0.9 alpha (24 jun 1998) *** # # Copyright (C) 1998 B.W. van Schooten # # Purpose: Find article information on the Web by supplying the title. # Method: Select the title or a substring of the title with your mouse and # run the program. After a while, a results file is generated in the current # directory and you can browse through the information found. # # Details: The last X selection (cutbuffer 0) is neatened up and used as an # exact-match query for the av.yahoo.com database. The documents # corresponding to the first 20 search hits are retrieved and the 7 lines # surrounding where the match occurs are put in a HTML file, together with # the URLs of the documents. Then, Lynx is started to view the results. # # Obvious limitations: # * Should also allow stdin instead of cutbuffer 0 # * # of lines of context is not configurable # * Document generated is not quite valid HTML because of the rather # arbitrary embedding of HTML fragments. # # Bugs: # * Addresses with nondefault ports don't seem to work properly # open file to write html page in open(HTMLOUT,">getwebref-results.html"); # get current X selection open(XSELECTION,"~/bin/xselection -cutbuffer 0|"); @xsel=; close(XSELECTION); print "SELECTION: @xsel\n"; # `neaten up' X selection to be suitable for search engine by: changing # all nonwords to `+' separators, removing leading and trailing nonwords, # making everything lowercase. $_ = "@xsel\n"; while (s/\+/-/) {}; while (s/[^\w\+]+/\+/) {}; s/^\+//; s/\+$//; tr/A-Z/a-z/; $query=$_; print "QUERY: $query\n"; print HTMLOUT "Results of $query\n"; # make connection to av.yahoo.com search engine. open(INDEXOUT,"echo \"GET /bin/query?p=\%22$_\%22\" | telnet av.yahoo.com 80|"); # modify X selection to fit regexp search `+' -> `\W*' while ($query =~ s/\+/\\W\*/) {}; # Now, we collect and parse telnet output. # It is assumed that the URLs of the hits can be identified by searching for: # `
  • ' `){ chop; # check and remove prefix of hit if (s/^\s*\<\s*[lL][iI]\s*\>\s*\<[aA]\s*[hH][rR][eE][fF]="//) { # now, remove postfix of hit so we obtain the stripped url. s/"\>.*//; print"CHECKING $_\n"; print HTMLOUT "

    \n$_

    \n"; # obtain host/port and directory part /http:..([\w\d\.\-:]*)(\/.*)/; $hostport = $1; if (not $hostport =~ s/:/\ /) { # no port given? default port. $hostport .= " 80"; } print "$hostport $2 \n"; open(URLIN,"echo \"GET $2\" | telnet $hostport |"); # init multiline buffer @linebuf=("","","","","","",""); while($nline=) { # strip HTML codes # $nline =~ s/\<\s*\w+.*\>//; # scroll the multiline buffer; shift(@linebuf); push(@linebuf,$nline); # search for string in middle of buffer $_= $linebuf[2] . $linebuf[3] . $linebuf[4]; tr/A-Z/a-z/; if (/$query/) { print "@linebuf"; print HTMLOUT "

      \n@linebuf\n
    "; @linebuf=("","","","","","",""); } } } } print HTMLOUT "\n"; close(HTMLOUT); system("lynx getwebref-results.html");