#!/usr/bin/perl

=head1 NAME

xls2tsv

=head1 SYNOPSIS

 xls2tsv file.xls > file.tsv

=head1 DESCRIPTION

This is a fairly simple-minded tool to extract content from Excel
spreadsheet files and print the cell data in plain-text tab-delimited
format.  The output preserves one or more consecutive empty rows a
single blank line; within non-empty rows, each non-row-final empty
cell is preserved as a tab character.

We loop over the "worksheet" pages in the given .xls file, and output
a ";;; sheetname" line, followed by a blank line, to mark the start of
each one.

When cell data content is found to be unicode, it is printed out in
utf8 encoding.  Leading and trailing spaces are discarded; any cell
found to contain only whitespace characters (including tab,   and
other "wide" space characters available in unicode) is treated as
empty.  Apart from these points, cell contents are output exactly as
found.

Cell and string formatting features are not preserved; changes of
foreground/background color and font style (bold, italic, underline,
etc) are ignored.

=head1 AUTHOR

David Graff  (at ldc.upenn.edu)

=cut

use strict;
use Encode;
use Spreadsheet::ParseExcel;

my $Usage = "$0 file.xls > file.txt\n";
( @ARGV == 1 and -f $ARGV[0] ) or die $Usage;

my $filepath = shift;

my $xl = Spreadsheet::ParseExcel->new;
my $wb = $xl->Parse( $filepath ) or die "$filepath: $!\n";

binmode STDOUT, ":utf8";

for my $sheet ( @{$wb->{Worksheet}} ) {
    printf( ";;; %s\n\n", $sheet->{Name} );
    $sheet->{MaxRow} ||= $sheet->{MinRow};
    my $blankcount = 0;
    for my $row ( $sheet->{MinRow} .. $sheet->{MaxRow} ) {
        $sheet->{MaxCol} ||= $sheet->{MinCol};
        my $rowtxt = "";
        for my $col ( $sheet->{MinCol} .. $sheet->{MaxCol} ) {
            my $cell = $sheet->{Cells}[$row][$col];
            if ( ! $cell ) {
                $rowtxt .= "\t";
                next;
            }
            my $val = $cell->{Val};
            if ( !defined( $val ) or $val eq '' ) {
                $rowtxt .= "\t";
                next;
            }
            $val = decode( "UTF-16BE", $val ) if ( $cell->{Code} eq 'ucs2' );
            $val =~ s/^\s+//; $val =~ s/\s+$//;
            if ( $val eq '' ) {
                $rowtxt .= "\t";
                next;
            }
            $rowtxt .= "$val\t";
        }
        $rowtxt =~ s/\t+$//;
        if ( $rowtxt eq '' ) {
            print "\n" unless ( $blankcount++ );
        }
        else {
            $blankcount = 0;
            print "$rowtxt\n";
        }
    }
    print "\n" unless ( $blankcount );
}