supprot \x42 HEX numbers in recode

[webpac2] / lib / WebPAC / Input.pm

diff --git a/lib/WebPAC/Input.pm b/lib/WebPAC/Input.pm
index e400e7a..462a303 100644 (file)
--- a/lib/WebPAC/Input.pm
+++ b/lib/WebPAC/Input.pm
@@ -3,12 +3,13 @@ package WebPAC::Input;
 use warnings;
 use strict;
 
-use blib;
+use lib 'lib';
 
 use WebPAC::Common;
 use base qw/WebPAC::Common/;
 use Data::Dump qw/dump/;
 use Encode qw/decode from_to/;
+use YAML;
 
 =head1 NAME
 
@@ -136,7 +137,7 @@ This function will read whole database in memory and produce lookups.
 
 By default, C<input_encoding> is assumed to be C<cp852>.
 
-C<offset> is optional parametar to position at some offset before reading from database.
+C<offset> is optional parametar to skip records at beginning.
 
 C<limit> is optional parametar to read just C<limit> records from database
 
@@ -181,9 +182,7 @@ sub open {
        my $input_encoding = $arg->{'input_encoding'} || $self->{'input_encoding'} || 'cp852';
 
        # store data in object
-       foreach my $v (qw/path offset limit/) {
-               $self->{$v} = $arg->{$v} if ($arg->{$v});
-       }
+       $self->{$_} = $arg->{$_} foreach grep { defined $arg->{$_} } qw(path offset limit);
 
        if ($arg->{load_row} || $arg->{save_row}) {
                $log->logconfess("save_row and load_row must be defined in pair and be CODE") unless (
@@ -207,6 +206,7 @@ sub open {
                        while (@r) {
                                my $from = shift @r;
                                my $to = shift @r;
+                               $from =~ s/^\\x([0-9a-f]{2})/chr(hex($1))/eig;
                                $recode_map->{$from} = $to;
                        }
 
@@ -229,6 +229,8 @@ sub open {
 
        my $class = $self->{module} || $log->logconfess("can't get low-level module name!");
 
+       $arg->{$_} = $self->{$_} foreach qw(offset limit);
+
        my $ll_db = $class->new(
                path => $arg->{path},
                input_config => $arg->{input_config} || $self->{input_config},
@@ -242,6 +244,9 @@ sub open {
                %{ $arg },
        );
 
+       # save for dump and input_module
+       $self->{ll_db} = $ll_db;
+
        unless (defined($ll_db)) {
                $log->logwarn("can't open database $arg->{path}, skipping...");
                return;
@@ -258,10 +263,10 @@ sub open {
        my $to_rec = $size;
 
        if (my $s = $self->{offset}) {
-               $log->debug("skipping to MFN $s");
-               $from_rec = $s;
+               $log->debug("offset $s records");
+               $from_rec = $s + 1;
        } else {
-               $self->{offset} = $from_rec;
+               $self->{offset} = $from_rec - 1;
        }
 
        if ($self->{limit}) {
@@ -270,21 +275,22 @@ sub open {
                $to_rec = $size if ($to_rec > $size);
        }
 
-       # store size for later
-       $self->{size} = ($to_rec - $from_rec) ? ($to_rec - $from_rec + 1) : 0;
-
        my $strict_encoding = $arg->{strict_encoding} || $self->{strict_encoding}; ## FIXME should be 1 really
 
-       $log->info("processing $self->{size}/$size records [$from_rec-$to_rec]",
+       $log->info("processing ", $self->{size} || 'all', "/$size records [$from_rec-$to_rec]",
                " encoding $input_encoding ", $strict_encoding ? ' [strict]' : '',
                $self->{stats} ? ' [stats]' : '',
        );
 
+       $self->{size} = 0;
+
        # read database
        for (my $pos = $from_rec; $pos <= $to_rec; $pos++) {
 
                $log->debug("position: $pos\n");
 
+               $self->{size}++; # XXX I could move this more down if I didn't want empty records...
+
                my $rec = $ll_db->fetch_rec($pos, sub {
                                my ($l,$f_nr,$debug) = @_;
 #                              return unless defined($l);
@@ -292,6 +298,8 @@ sub open {
 
                                return unless ( defined($l) && defined($f_nr) );
 
+                               my $marc_subfields = $l =~ s/\x1F(\w)/\^$1/g; # fix MARC subfiled delimiters to ^
+
                                warn "-=> $f_nr ## |$l|\n" if ($debug);
                                $log->debug("-=> $f_nr ## $l");
 
@@ -319,6 +327,8 @@ sub open {
                                        }
                                }
 
+                               $l =~ s/\^(\w)/\x1F$1/g if $marc_subfields;
+
                                $log->debug("<=- $f_nr ## |$l|");
                                warn "<=- $f_nr ## $l\n" if ($debug);
                                return $l;
@@ -385,12 +395,11 @@ sub open {
        $self->{max_pos} = $to_rec;
        $log->debug("max_pos: $to_rec");
 
-       # save for dump
-       $self->{ll_db} = $ll_db;
-
        return $size;
 }
 
+sub input_module { $_[0]->{ll_db} }
+
 =head2 fetch
 
 Fetch next record from database. It will also displays progress bar.
@@ -410,7 +419,7 @@ sub fetch {
        $log->logconfess("it seems that you didn't load database!") unless ($self->{pos});
 
        if ($self->{pos} == -1) {
-               $self->{pos} = $self->{offset};
+               $self->{pos} = $self->{offset} + 1;
        } else {
                $self->{pos}++;
        }
@@ -468,7 +477,7 @@ because it takes into account C<offset> and C<limit>.
 
 sub size {
        my $self = shift;
-       return $self->{size};
+       return $self->{size}; # FIXME this is buggy if open is called multiple times!
 }
 
 =head2 seek
@@ -561,6 +570,10 @@ sub stats {
 
        $log->debug( sub { dump($s) } );
 
+       my $path = 'var/stats.yml';
+       YAML::DumpFile( $path, $s );
+       $log->info( 'created ', $path, ' with ', -s $path, ' bytes' );
+
        return $out;
 }