#!/usr/local/bin/perl
# unbib
$version = "0.95BETA";
# by Jon Van Oast (jon@kzsu.stanford.edu)
# see also: http://kzsu.stanford.edu/uwi/about-unbib.html
#
# It is sloppy perl code that checks NCSA httpd(1.4)-made 'referer_log'
# files and builds a list of all/some local web pages and URLs that could
# potentially point at them.
#
# A "data file" is used to store the list for the next usage (when it
# is then read in), and standard out is the lucky recipient of the
# HTML marked-up junk that will contain the back-links.
# Raw data (array) of form: LOCAL_URL<REF_URL1<REF_URL2<...<REF_URLn
# (NOTE: I plan on adding a count field to each REF_URL so that you
# display only back-links with > X back-links [i.e., more valid].)
#
# Example usage:   unbib unbib.dat > unbib.html

$IGNORE = 0;  $STRONG = 7;  $SHOW_COUNT = 0;

print "<!-- unbib $version by jon\@kzsu.stanford.edu -->\n";

# $STORED is the single expected arg -- it is where to save for next time
($STORED,$FILE) = @ARGV;
die "Usage: unbib <data-file-name> <log>\n" if (($STORED eq "") || ($FILE eq ""));

if (open(STORED,$STORED)) {
  @LINKS = <STORED>;
  chop(@LINKS);
  close(STORED);
} else {
  print "<!-- Data-file $STORED will be created (new). -->\n";
}

#fairly useful stats
$prior_to_count = @LINKS + 1 if ($#LINKS > 0);
$to_count = $prior_to_count;
  
#the location/name of the referer_log is hard-coded now.  this sucks.
#this will "get better" in the next version.
#$FILE = "/tmp/htlogs/referer_log";

open(FILE,$FILE) || die "Bad referer_log $FILE!\n";
chop(@REF = <FILE>);
close(FILE);
@REF = grep(/ (\/uwi|\/jon)/,@REF);

foreach $REF (@REF) {

  #this first conditional is a quick weed-out to ignore any local docs you
  #don't care about.  since referer_log entries are of the form
  #REFERER_URL -> LOCAL_FILE, i just look for " <start of path>" in the
  #line to mean that we should process this one.
  #if (/ \/uwi/) {
  {
    ($from,$to) = split(/ \-> /,$REF);
    #these next two lines may be unnecessary, but i was having trouble
    #with url args (the +) causing probs.  the *next* pair of lines
    #should solve the problem, so expect these gone soon.
    #$to =~ tr/\+/\-/;
    #$from =~ tr/\+/\-/;
    #these two lines chop off anything after (and including) a '?' in
    #the url.  i did this cuz i dont care about args... cuz, well, i dont.
    $to = $` if ($to =~ /\?|</);
    $from = $` if ($from =~ /\?|</);

    #this next conditional is a "weed-out" one as well.  you can filter
    #any undesirable paths here as well.  example: refs from your own site!
    if (!($from =~ /^\/uwi/i) && !($from =~ /kzsu\.stanford/i)) {

      match: for ($i = 0 ; $i <= $#LINKS ; $i++) {
        $link = @LINKS[$i];
       
        if ($link =~ /^$to/) {          #is local url ($to) in list already?

          if (!($link =~ /$from/i)) {   #is ref url ($from) in list already?
            @LINKS[$i] = @LINKS[$i] . "<0\{" . $from;
            $from_count++;
          } else {
            #@LINKS[$i] =~ /\<[0-9]*\{$from/i;
            $LINKS[$i] =~ /\<(\d*)\{$from/i;
            $pre = $`;
            $suf = $';
            $c = $1;
            #($c) = split(/\{/,$c);
            #$c =~ /[0-9]+/;
            #$c = $&;
            $c++;
            @LINKS[$i] = "$pre<$c\{$from$suf";
##print "%%@LINKS[$i]%%\n\n";
          }

          $i = $#LINKS + 1;             #don't trust my knowledge of 'last'!
          $from = "";
          last match;
        }
      }

      if ($from ne "") {                #$from eq "" if added above
        push(@LINKS,$to . "<0{" . $from);
        $from_count++;
        $to_count++;
      }

    }
  }
}

if (open(STORED,">".$STORED)) {
  $ok = 1;
} else {
  #(theoretically) if can't write to $STORED dump raw data to stdout
  print "*** UNABLE TO OPEN $STORED AS DATA FILE ***\n";
  print "*** DATA FILE SENT TO STANDARD OUT ***\n\n";
  print @LINKS;
  print "\n\n";
}

#dont think this is necessary --
#$from_count++;

chop($now = `date`);

#makes the html part, to stdout
$c1 = $IGNORE + 1;  $c2 = $STRONG + 1;
print "<title>Unbibliography</title>
  <h1>Unbibliography</h1>
  <h3><a href=\"http://kzsu.stanford.edu/uwi/unbib-about.html\">About</a>
  unbib $version.</h3>
  <b>File updated:</b> $now<br>
  <b>Count of pages linked:</b> $to_count ($prior_to_count from before)<br>
  <b>New from-links added:</b> $from_count<br>
  <b>Ignore</b> &lt;= $c1 &gt; <b>Weak</b> &lt;= $c2 &gt; <b>Strong</b><p>
  <hr>\n";

foreach $link (@LINKS) {
  print STORED "$link\n" if ($ok);
  @EACH = split(/\</,$link);
  $orig = shift(@EACH);
  $count = $#EACH + 1;
  $total = $total + $count;
  print "\n<p><b>$orig</b> ($count) - ";
  foreach $each (@EACH) {
    ($each_ct,$each) = split(/\{/,$each);
    if (($each_ct > $IGNORE) && ($each_ct < $STRONG)) {
      $char = '.';
    } else {
      $char = "#";
    }
    $char = "$each_ct:$char" if ($SHOW_COUNT);
    print "<a href=\"$each\">($char)</a> " if ($each_ct > $IGNORE);
  }
}

print "<p><hr><h2>$total unbibliographical references</h2><p>\n";


close(STORED) if ($ok);