#!/usr/bin/perl
# FILE: "/home/evmik/src/my_src/GradeBook/banner2csv.pl"
# LAST MODIFICATION: "Fri, 17 Jan 2014 11:12:43 -0500 (evmik)"
# (C) 2012 by Eugeniy Mikhailov, <evgmik@gmail.com>

#use HTML::Entities;
use HTML::Parser;
use HTML::TableExtract;
use Encode;
#use HTML::TableExtract qw(tree);
use Data::Dumper;
#use strict;

if ( @ARGV < 1 )
{
	print "Converts a WM banner summary list to csv file with students info\n";
	print "\n";
	print "Usage:\n";
	print "  " . $0 . "banner_html_file_to_parse\n";
	print "\n";
	print "Example:\n";
	print "  " . $0 . "banner_class_summary_list.html\n";
	die;
}

my $html_file = shift;
open (F_HTML, "<", $html_file)  or  die "Failed to read file $html_file : $!";

my @fcontent = <F_HTML>;
$html_string=join('', @fcontent);


# this look for the table based on its headers
my $te = HTML::TableExtract->new(
	headers => [qw(Record Student ID)], 
	keep_html => 1, 
	slice_columns => 0
);
 
# find table with students info
$te->parse($html_string);

my $parsed_text="";
my $href="";

$aparser = HTML::Parser->new(api_version => 3,
     start_h => [\&a_start_handler, "self,tagname,attr"],
     report_tags => [qw(a img)],
    );

$spanparser = HTML::Parser->new(api_version => 3,
     start_h => [\&span_start_handler, "self,tagname,attr"],
     report_tags => [qw(span)],
    );

sub span_start_handler
{
    my($self, $tag, $attr) = @_;
    return unless $tag eq "span";

    $self->handler(text  => [], '@{dtext}' );
    $self->handler(start => \&span_start_handler);
    $self->handler(end   => \&spantag_end_handler, "self,tagname");
}

sub a_start_handler
{
    my($self, $tag, $attr) = @_;
    return unless $tag eq "a";
    return unless exists $attr->{href};
    $href=$attr->{href};
    #print "A $attr->{href}\n";

    $self->handler(text  => [], '@{dtext}' );
    $self->handler(start => \&img_handler);
    $self->handler(end   => \&atag_end_handler, "self,tagname");
}

sub img_handler
{
    my($self, $tag, $attr) = @_;
    return unless $tag eq "img";
    push(@{$self->handler("text")}, $attr->{alt} || "[IMG]");
}

sub atag_end_handler
{
    my($self, $tag) = @_;
    my $text = encode('utf8',join("", @{$self->handler("text")}));
    $text =~ s/^\s+//;
    $text =~ s/\s+$//;
    $text =~ s/\s+/ /g;
    #print "T $text\n";
    $parsed_text = $text;

    $self->handler("text", undef);
    $self->handler("start", \&a_start_handler);
    $self->handler("end", undef);
}

sub spantag_end_handler
{
    my($self, $tag) = @_;
    my $text = encode('utf8',join("", @{$self->handler("text")}));
    $text =~ s/^\s+//;
    $text =~ s/\s+$//;
    $text =~ s/\s+/ /g;
    #print "T $text\n";
    $parsed_text = $text;

    $self->handler("text", undef);
    $self->handler("start", \&span_start_handler);
    $self->handler("end", undef);
}


# assumes that we were able to chose a proper table before 
# and really get the proper one
foreach $row ($te->rows) {
	#student name
	$aparser->parse(@$row[1]);
	$name=$parsed_text;

	($lname, $fname)=split(',', $name);
	$fname =~ s/^ //;

	#student id
	#print @$row[2];
	$spanparser->parse(@$row[2]);
	$sid=$parsed_text;

	#student name
	$aparser->parse(@$row[7]);
	$email = $href;
	$email =~ s/mailto://;

	# section
	$section = "unknown";

	print join(',', $sid,  $fname, $lname, $section, $email);
	print "\n";
}