# --------------------------------------------------------------------------------- 
# Author: Dean Stringer
# Description/Purpose:
#
#	Takes an CSV formatted file as input, reads its 1st line expecting a row
#	that lists column names, then uses these as element names to be output to
#	STDOUT as XML records.
#
#	Perl Solutions considered for parsing the source CSV...
#		AnyData::Format::CSV
#		Text::CSV_XS
#		XML::SAXDriver::CSV
#		XML::CSV
#		XML::Excel
#
#	Excel option seems the most fragile, as we can't rely on MS keeping XLS in the
#	same format (in fact they are planning to move to XML). The SAXDriver option would
#	be best for serial handling of *large* CSV files.
#	Text::CSV_XS is an extenstion of Text::CSV by an author who's published a number
#	of modules on CPAN, and it seems the most mature of the above.
#
# Dependencies:
#	Perl Modules...
#		Text::CSV_XS	parsing the input CSV data
#		XML::Writer	writing the XML output
#
# Expected Parameters: 
#	filename to parse
#
# Returns: 
#	XML document
# 
# Sample Invocation:
#		perl <filename.pl> filename.txt > output.xml
#
# Error situations:
#	Will bomb-out w error if sourcefile not found or if not able to parse the
#	CSV line data for some reason.
#
#	There is currently no error handling for missing fields either, so if someone
#	removes a field in a row, that data will appear in the wrong output column.
# --------------------------------------------------------------------------------- 
use strict;
use XML::Writer;
use Text::CSV_XS;
use HTML::Entities;

my $file2Read = $ARGV[0];
my $syntax = "\nSyntax:  $0 <file2read>\n";
unless ((-e $file2Read) && (-f $file2Read)) { die "No such file $file2Read $syntax"; }

my $parentTagName = "phonebook";
my $nodeTagName = "person";

my $csv = Text::CSV_XS->new( {
     'quote_char'  => '"',
     'escape_char' => '"',
     'sep_char'    => ',',
     'binary'      => '1' }
);

open(INFILE, "<" . $file2Read) || die "Cant open file!";

# Check to see if the file contains a set of field header/names in the 1st row
my $line = <>;
# $line =~ tr/\r\n|\n//d;	# If a Windows machine source file, remove CR/LF pairs off end of line
my @fields = split(/,/,$line);
my $fieldCount = @fields;

# Cleanup field name formats, remove spaces so we just have A-Z chars
for (my $i=1; $i<$fieldCount; $i++) {
	$fields[$i] =~ s/[^a-z]//igs;
}

my $writer = new XML::Writer( NEWLINES => 1 );
$writer->xmlDecl();	# Send declaration so we can post-process w XSLT later
$writer->startTag($parentTagName);

while ($line = <>){

	if ($csv->parse($line)) {

		my @field = $csv->fields;
		my $count = 0;
		$writer->startTag($nodeTagName);
		for my $column (@field) {
			$writer->startTag($fields[$count]);
			$writer->characters(HTML::Entities::encode($column));
			$writer->endTag($fields[$count]);
			$count++;		
		}
		$writer->endTag($nodeTagName);
	} else {
		my $err = $csv->error_input;
		print STDERR "parse() failed on argument: ", $err, "\n";
		}
}
close(INFILE);

$writer->endTag($parentTagName);
$writer->end();

exit;