From: Dobrica Pavlinusic <dpavlin@rot13.org>
Date: Thu, 26 May 2011 12:40:29 +0000 (+0200)
Subject: Merge branch 'master' of git.rot13.org:/git/MojoFacets
X-Git-Url: http://git.rot13.org/?p=MojoFacets.git;a=commitdiff_plain;h=6ab2b22ee8e99850a3a8ea97f80465df2c2c8651;hp=3e1ecb5815eb15e076a82ccf9e64cd7e03011162

Merge branch 'master' of git.rot13.org:/git/MojoFacets
---

diff --git a/Makefile.PL b/Makefile.PL
index 29d5e24..a55b88a 100755
--- a/Makefile.PL
+++ b/Makefile.PL
@@ -13,6 +13,7 @@ requires 'HTML::TableExtract';
 requires 'File::Path';
 requires 'Text::Unaccent::PurePerl';
 requires 'Statistics::Descriptive';
+requires 'Text::CSV';
 
 features(
 	'profile' => [
diff --git a/bin/debian-install.sh b/bin/debian-install.sh
index 019f9e6..d2f28e2 100755
--- a/bin/debian-install.sh
+++ b/bin/debian-install.sh
@@ -1,3 +1,3 @@
 #!/bin/sh -x
 
-sudo apt-get install libhtml-tableextract-perl libjson-perl libmodule-install-perl libstatistics-descriptive-perl
+sudo apt-get install libhtml-tableextract-perl libjson-perl libmodule-install-perl libstatistics-descriptive-perl libmodule-install-perl
diff --git a/lib/MojoFacets/Import/CSV.pm b/lib/MojoFacets/Import/CSV.pm
index 6ab3395..260c002 100644
--- a/lib/MojoFacets/Import/CSV.pm
+++ b/lib/MojoFacets/Import/CSV.pm
@@ -5,90 +5,43 @@ use strict;
 
 use base 'Mojo::Base';
 
-use File::Slurp;
+use Text::CSV;
 use Data::Dump qw(dump);
-use Encode;
 
 __PACKAGE__->attr('full_path');
 
-my $null = ''; # FIXME undef?
-
-sub _split_line {
-	my ( $delimiter, $line ) = @_;
-	my @v;
-	while ( $line ) {
-		my $v;
-		if ( $line =~ s/^"// ) {
-			$line =~ s/""/_qq_/gc;
-			$line =~ s/^\s*([^"]*)\s*"\Q$delimiter\E?// || die "can't parse [$line] ",dump(@v);
-			$v = $1;
-		} elsif ( $line =~ s/^\s*([^\Q$delimiter\E]+)\s*\Q$delimiter\E?// ) {
-			$v = $1;
-		} elsif ( $line =~ s/^\s*\Q$delimiter\E// ) {
-			$v = $null;
-		} else {
-			die "can't parse [$line]\n";
-		}
-
-		$v =~ s/^\s*(.+?)\s*$/$1/;
-		$v = $null if $v eq '_qq_'; # "" field which is not first one
-		$v =~ s/_qq_/"/g;
-		$v =~ s/_LF_/\n/g;
-		push @v, $v;
-	}
-
-	return @v;
-}
-
 sub data {
 	my $self = shift;
 
 	my $path = $self->full_path;
 
-	my $data = read_file $path, { binmode => ':raw' }; # FIXME configurable!
 	my $encoding = 'utf-8';
 	if ( $path =~ m/\.(\w+).csv/i ) {
 		$encoding = $1;
 	}
-	warn "decoding ", length($data), " bytes using $encoding\n";
-	$data = decode($encoding, $data);
 
-	# multi-line strings
-	while ( $data =~ s/(,"[^"]*)[\n\r]+([^"]*)/$1_LF_$2/sg ) {
-		 warn "multi-line quoted CSV data found";
-	}
+	my $data = { items => [] };
+	my @header;
 
-	my @lines = split(/\r?\n/, $data);
-	$data = { items => [] };
+	my $csv = Text::CSV->new ( { binary => 1, eol => $/ } )
+		or die "Cannot use CSV: ".Text::CSV->error_diag ();
 
-	my $delimiter = ',';
-
-	if ( $lines[0] !~ m/,/ ) {
-		if ( $lines[0] =~ m/;/ ) {
-			$delimiter = ';';
-		} elsif ( $lines[0] !~ /;/ && $lines[1] =~ /;/ ) {
-			shift @lines; # FIXME skip non-header line
-			$delimiter = ';';
+	open my $fh, "<:encoding($encoding)", $path or die "$path: $!";
+	while ( my $row = $csv->getline( $fh ) ) {
+		if ( ! @header ) {
+			@header = @$row;
+			next;
 		}
-	}
-
-	warn "$path ", $#lines + 1, " lines encoding: $encoding delimiter:",dump($delimiter);
-
-	my $header_line = shift @lines;
-
-	my @header = _split_line( $delimiter, $header_line );
-	warn "# header ",dump( @header );
-
-	while ( my $line = shift @lines ) {
-		chomp $line;
-		my @v = _split_line($delimiter, $line);
 		my $item;
-		foreach my $i ( 0 .. $#v ) {
-			$item->{ $header[$i] || "f_$i" } = [ $v[$i] ];
+		foreach my $i ( 0 .. $#{$row} ) {
+			$item->{ $header[$i] || "f_$i" } = [ $row->[$i] ];
 		}
 		push @{ $data->{items} }, $item;
 	}
 
+	$csv->eof or $csv->error_diag();
+	close $fh;
+
 	$data->{header} = [ @header ];
 	
 	return $data;