use strict;
use warnings;
use feature 'say';

use HTML::Entities;
use Data::Dump qw/pp/;

my $text;
my %anchor_count;
my @aoh_toc;
my ($min,$max) = (3,6);

my @stack;

$stack[$min-1] = \@aoh_toc;            # root


while (my $line = <DATA>) {
 
   if ( $line =~ m# ^ (\s*) < \s* h([$min-$max]) \s* > (.*) </ \s* h(\
+2) \s* > \s*$ #xi ) {
      #say $line;

      my ($indent,$level,$text) = ($1,$2,$3);

      if ( $indent ) {                  # ignore indented <h*> 
         warn "Skipping $line";
      }
      else {


         my $anchor = text2anchor($text);

         warn "Duplicate '$anchor' " if $anchor_count{$anchor}++;

         $line = "<h$level><a name='$anchor'>$text</a></h$level>\n";

         my $a_sub = [];

         push @{$stack[$level-1]},
           {
            text => $text,
            link => $anchor,
            sub  => $a_sub,
            level => $level,
           };

         $stack[$level] = $a_sub;

      }
      #say $line;
   }

   $text .= $line

}

# pp \@aoh_toc;

say "<ul>";
create_toc(\@aoh_toc);
say "</ul>";



print $text;

sub text2anchor {
   my ($text) = @_;

   # trim
   $text =~ s/^\s+//;
   $text =~ s/\s+$//;

   # make valid indentifier
   my $encoded = encode_entities( $text );

   # replace whitespaces
   $encoded =~ s/\s/_/g;

   return $encoded;
}

sub create_toc {
   my ($aoh_toc) = @_;

   for my $h_line (@$aoh_toc) {

      my $indent = "    " x ($h_line->{level}-$min);
      say qq~$indent<li><a href='#$h_line->{link}'>$h_line->{text}</a>
+</li>~;

      my $a_sub = $h_line->{sub};

      if ( @$a_sub ) {
         say "$indent<ul>";
         create_toc($a_sub);
         say "$indent</ul>";
      }

   }
   
}


__DATA__
...
[download]