#!/usr/bin/perl
# ************************************************************************** #
# filesize.prl written by Chris Evans (C.Evans@sghms.ac.uk)                  #
# This program finds all your HTML files (if you give it some pointers, see  #
# below!) and places a note of the size of the locals files pointed to at    #
# end of the HREF.  Currently it only handles local calls with absolute      #
# filenames, i.e. it won't work on calls like <A HREF="http:// .... even if  #
# they are to the local server, nor will it handle <A HREF="../file.htm"     #
# The latter is actually surprisingly difficult to handle really well,       #
# allowing for all possible relative forms and I haven't got the time to add #
# these things now.  I will probably do so in the near future.               #
#                                                                            #
# The program takes as input in the variables declared immediately below     #
#     $filemask   -- file mask (appropriate to the OS) for the find command  #
#     $outputdir  -- the directory into which to put these files             #
#     $menu       -- filename for the menu to all other reports              #
#     $report     -- filename for the directory order output                 #
#     $time_rep   -- filename for the modification time sorted output        #
#     $other_rep  -- filename for the files other than *.htm?                #
#     $startdir   -- directory from which to start the recursive search      #
#     $www_offset -- any additional directory reference that accounts for    #
#          a possible difference between the root used by unix for the       #
#          search versus that shown to the world by the HTTP server in use   #
#     This last variable may seem odd but on our local machine the root      #
#          shown to the outside world by the server is at:                   #
#              /usr1/www/pages/                                              #
#          but my pages (from which I want the search for *.htm files to     #
#          start is at:                                                      #
#              /usr1/www/pages/mhs/psychotherapy/                            #
#          giving the value for $www_offset of /usr1/www/pages               #
#                                                                            #
# file written by Chris Evans (C.Evans@sghms.ac.uk) 21.iii.96                #
# copyright Chris Evans _BUT_ feel free to distribute this subject to the    #
# following requirements:                                                    #
# 1) you do not make a profit on the distribution                            #
# 2) you retain this header information in full                              #
# I would also very much appreciate feedback (I know the programming is      #
# awful, but I've written very little Perl work    and I'm a psychodynamic   #
# psychotherapist with no intention of giving up my day job so no flames if  #
# you can resist the temptation) and a copy of enhanced versions if you do   #
# hack this into something better                                            #
#                                                                            #
# Chris Evans (C.Evans@sghms.ac.uk) Section of Psychotherapy,                #
#      St. George's Hospital Medical School, Cranmer Terrace,                #
#      London, SW17 0RE, Britain                                             #
# ************************************************************************** #

# you need to tell the program how the directory (.) from which it searches
# is shown as an HREF.  For me this is /mhs/psychotherapy, if your searching
# from a server root directory it would be just "/"
$startdir   = "/mhs/psychotherapy"; 

# now you need to tell the program the location of the server root as 
# seen by the operating system (note, this is very Unix dependent so
# I should probably be writing "Unix" not "OS" and you shouldn't expect
# much without a lot of fiddling on non-Unix platforms)
# for me the Unix location of the server root is:
$www_offset = "/usr1/www/pages";

# and finally you need to tell the program what files you want it to handle
# you may need to add *.html and *.HTML to this list but I don't use those
# extensions 'cos I like to see my Unix directories via Samba
@filemask   = ("*.htm", "*.HTM"); 

# the next two lines were used for debugging only
# open (TMP,">tmp.lst"); # sometimes used for debugging!
# @filemask   = ("intro.htm"); # again, used for debugging only!

# the program plays very safe and makes backup copies of all the files
# it changes by appending this backup extension to the filename
# NOTE: intro.htm becomes intro.htm.bak NOT intro.bak
# I didn't have time to change this to the more sensible form
$bak        = ".bak";             # backup extension

# **** That's the end of user input I hope ****************************


&getdata;


# now run through all subdirectories looking for files
foreach $mask (@filemask) {
   open(FIND, "find . -name \"$mask\" -print |") || die "Couldn't run find: $!\n";
   while ($filename = <FIND>)
   {
      chop($filename);
      $touched{$filename} = 1;
      print "\nProcessing file $filename...\n";
      $backname = $filename . $bak;
      rename($filename, $backname) || die "Couldn't rename $filename as $backname\n";
      open (NEW, ">$filename") || die "Couldn't open new file as $filename\n";
      open (OLD, "$backname") || die "Couldn't open $filename\n";

      while (<OLD>) {
         if (/<A HREF="(\/mhs\/psychotherapy\/[-%\w#.\/]+)/) {
            $href = $1;
            $href =~ s/%20/ /g; # handles spaces in filenames which are
                                # repesented by %20 in URLs, note the 
                                # g on the substitution to catch repeated spaces
            if ($href =~ s/#\w+//) {
               # theres a named anchor point in the file
               $name = 1;  # flag this up for handling in the
               # the light of testing the fileref itself (next!)
            } else {
               $name = 0;
            }
            if (!$size{$href}) {
               # don't have file size for that href
               # might be 'cos it's a directory
               $href =~ s/\/$//; # chop final slash off if there is one
               $file = $www_offset.$href;
               if (-d($file)) {
                  $size = "dir";               
               } else {
                  # if it's not a directory then it probably a duff HREF
                  # flag it as highly suspect
                  $size = "?bad link?";
               }
            } else {
               # it's a straightforward file HREF and 
               # the file exists and you know its size
               if ($name) {
                  $size = "in " . $size{$href};
               } else {
                  $size = $size{$href};
               }
            }
            if (/}<\/A>/) {  # test to see if the link has already been given a size
               s/({[^<]+)/{$size}/;    # if so, replace that
            } else {
               s/<\/A>/ {$size}<\/A>/; # if not, create a size label
            }
         } # end of if (/<A HREF=
      print NEW; # print it out to the new file
      }
      close NEW;
      close OLD;
   }
}

# now reset all the times
foreach $filename (keys(%touched)) {
   system("touch -c $date{$filename} $filename");
}





# ****************************************************#
# subroutines                                         #
# ****************************************************#

sub getdata {
   # stores all size data in one associative array
   open(FIND, "find . -name \"*.*\" -print |") || die "Couldn't run find: $!\n";
   while ($filename = <FIND>)
   {
      undef $size;
      chop $filename;
      # get data on the file
      ($nsize,$time) = (stat($filename))[7,9]; #use the 7th & 9th items in the stat array
      # reformat the size
      if (-d _) {
         $nsize = "dir";
      } elsif ($nsize <= 1000) {
         $size = "<1k";
      } elsif ($nsize < 1000000) {
         $size = int($nsize/1000);
         $size = $size."kb";
      } else {
         $size = int($nsize/1000000);
         $size = $size."Mb";
      }

      # reformat the date and time data
      ($s, $m, $h, $day, $mon, $yr) = (localtime($time))[0,1,2,3,4,5];
      $tot_size += $nsize;
      $h = &two_digits($h);
      $m = &two_digits($m);
      $s = &two_digits($s);
      $day = &two_digits($day);
      $ntime = $h.":".$m.":".$s;
      $montxt = (Jan,Feb,Mar,Apr,May,Jun,Jul,Aug,Sep,Oct,Nov,Dec)[$mon];
      $mon++;
      $mon = &two_digits($mon);
      $touch = $mon.$day.$h.$m.$yr;

      # reformat the filename to the http server form
      $href = $filename;
      $href =~ s|.|$startdir|o;

      # put this into two associative arrays for future use
      $size{$href} = $size;
      $date{$filename} = $touch;
   }
   close TIME;
} # &getdata


sub two_digits {
   local ($_) = @_;
   if (length($_) == 0) {
      $_ = "00";
   } elsif (length($_) == 1) {
      $_ = "0".$_;
   } elsif (length($_) > 2) {
      die "Supplied a number of more than 2 digits to two_digits\n";
   }
   $_;
}