X-Git-Url: http://git.rot13.org/?a=blobdiff_plain;f=lib%2FMojoFacets%2FImport%2FCSV.pm;h=e8fa19624cc809f514e3438226b3d8dde7f63d3f;hb=ce2e4a650a6a3c52ddccc2382488c6ff4a2ed763;hp=fc9b1ebc61500d686cf2f9e3fd9a7955561596e2;hpb=b3637d471c17dfa513a774a093adf5fcc51d8b35;p=MojoFacets.git

diff --git a/lib/MojoFacets/Import/CSV.pm b/lib/MojoFacets/Import/CSV.pm
index fc9b1eb..e8fa196 100644
--- a/lib/MojoFacets/Import/CSV.pm
+++ b/lib/MojoFacets/Import/CSV.pm
@@ -5,56 +5,74 @@ use strict;
 
 use base 'Mojo::Base';
 
-use File::Slurp;
+use Text::CSV;
 use Data::Dump qw(dump);
-use Encode;
 
-__PACKAGE__->attr('path');
-__PACKAGE__->attr('full_path'); # FIXME remove full_path
+__PACKAGE__->attr('full_path');
+
+sub ext { '\.[ct]sv$' };
 
 sub data {
 	my $self = shift;
 
-	my $path = $self->full_path || $self->path;
+	my $path = $self->full_path;
 
-	my $data = read_file $path, { binmode => ':raw' }; # FIXME configurable!
 	my $encoding = 'utf-8';
-	if ( $path =~ m/\.(\w+).csv/i ) {
+	if ( $path =~ m/\.([\w\-]+).[ct]sv/i ) {
 		$encoding = $1;
 	}
-	warn "decoding ", length($data), " bytes using $encoding\n";
-	$data = decode($encoding, $data);
 
-	my @lines = split(/\r?\n/, $data);
-	$data = { items => [] };
+	my $data = { items => [] };
+	my @header;
 
-	my $delimiter = qr/,/;
+	open my $fh, "<:encoding($encoding)", $path or die "$path: $!";
+	my $first = <$fh>;
+	my $possible_delimiters;
+	while ( $first =~ s/(\W)// ) {
+		$possible_delimiters->{$1}++;
+	}
+	warn "# possible_delimiters = ",dump($possible_delimiters);
+	seek $fh,0,0; # rewind for Text::CSV
+
+	my @sep_by_usage = sort { $possible_delimiters->{$b} <=> $possible_delimiters->{$a} } keys %$possible_delimiters;
+	my $sep_char = shift @sep_by_usage;
+	while ( $sep_char =~ m/^\s$/ ) {
+		last if $sep_char eq "\t" && $path =~ m/\.tsv$/i;
+		warn "## skip whitespace separator ",dump($sep_char);
+		$sep_char = shift @sep_by_usage;
+	}
 
-	if ( $lines[0] !~ /;/ && $lines[1] =~ /;/ ) {
-		shift @lines; # FIXME ship non-header line
-		$delimiter = qr/;/;
+	while ( $sep_char =~ m/^\"$/ ) {
+		warn "## skip quote separator ",dump($sep_char);
+		$sep_char = shift @sep_by_usage;
 	}
 
-	warn "$path ", $#lines + 1, " lines encoding: $encoding delimiter:",dump($delimiter);
+	if ( $sep_char !~ m/,/ && $possible_delimiters->{','} && $path =~ m/\.csv/i ) {
+		$sep_char = ',';
+		warn "## csv file detected so prefer , as separator";
+	}
 
-	my $header_line = shift @lines;
+	warn "sep_char = [$sep_char] for $path\n";
 
-	my @header = map { s/^"(.+)"$/$1/; s/^\s*(.+?)\s*$/$1/; $_ } split( $delimiter, $header_line );
-	warn "# header ",dump( @header );
+	my $csv = Text::CSV->new ( { binary => 1, eol => $/, sep_char => $sep_char } )
+		or die "Cannot use CSV: ".Text::CSV->error_diag ();
 
-	while ( my $line = shift @lines ) {
-		chomp $line;
-		my @v = split($delimiter, $line);
+	while ( my $row = $csv->getline( $fh ) ) {
+		if ( ! @header ) {
+			@header = @$row;
+			$header[0] =~ s/^#// if $path =~ m/\.tsv/i; # remove hash from 1st column
+			next;
+		}
 		my $item;
-		foreach my $i ( 0 .. $#v ) {
-			my $v = $v[$i];
-			$v =~ s/^"(.+)"$/$1/;
-			$v =~ s/^\s*(.+?)\s*$/$1/;
-			$item->{ $header[$i] || "f_$i" } = [ $v ];
+		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;