trunk2/lib/WebPAC.pm

package WebPAC;

use warnings;
use strict;

use Carp;
use Text::Iconv;
use Config::IniFiles;
use XML::Simple;
use Template;
use Log::Log4perl qw(get_logger :levels);
use Time::HiRes qw(time);

use Data::Dumper;

#my $LOOKUP_REGEX = '\[[^\[\]]+\]';
#my $LOOKUP_REGEX_SAVE = '\[([^\[\]]+)\]';
my $LOOKUP_REGEX = 'lookup{[^\{\}]+}';
my $LOOKUP_REGEX_SAVE = 'lookup{([^\{\}]+)}';

=head1 NAME

WebPAC - base class for WebPAC

=head1 DESCRIPTION

This module implements methods used by WebPAC.

=head1 METHODS

=head2 new

Create new instance of WebPAC using configuration specified by C<config_file>.

 my $webpac = new WebPAC(
        config_file => 'name.conf',
        code_page => 'ISO-8859-2',
        low_mem => 1,
        filter => {
                'lower' => sub { lc($_[0]) },
        },
 );

Default C<code_page> is C<ISO-8859-2>.

Default is not to use C<low_mem> options (see L<MEMORY USAGE> below).

There is optinal parametar C<filter> which specify different filters which
can be applied using C<filter{name}> notation.

This method will also read configuration files
C<global.conf> (used by indexer and Web font-end)
and configuration file specified by C<config_file>
which describes databases to be indexed.

=cut

# mapping between data type and tag which specify
# format in XML file
my %type2tag = (
        'isis' => 'isis',
#       'excel' => 'column',
#       'marc' => 'marc',
#       'feed' => 'feed'
);

sub new {
        my $class = shift;
        my $self = {@_};
        bless($self, $class);

        $self->{'start_t'} = time();

        my $log_file = $self->{'log'} || "log.conf";
        Log::Log4perl->init($log_file);

        my $log = $self->_get_logger();

        # fill in default values
        # output codepage
        $self->{'code_page'} = 'ISO-8859-2' if (! $self->{'code_page'});

        #
        # read global.conf
        #
        $log->debug("read 'global.conf'");

        my $config = new Config::IniFiles( -file => 'global.conf' ) || $log->logcroak("can't open 'global.conf'");

        # read global config parametars
        foreach my $var (qw(
                        dbi_dbd
                        dbi_dsn
                        dbi_user
                        dbi_passwd
                        show_progress
                        my_unac_filter
                        output_template
                )) {
                $self->{'global_config'}->{$var} = $config->val('global', $var);
        }

        #
        # read indexer config file
        #

        $self->{indexer_config_file} = new Config::IniFiles( -file => $self->{config_file} ) || $log->logcroak("can't open '",$self->{config_file},"'");

        # create UTF-8 convertor for import_xml files
        $self->{'utf2cp'} = Text::Iconv->new('UTF-8' ,$self->{'code_page'});

        # create Template toolkit instance
        $self->{'tt'} = Template->new(
                INCLUDE_PATH => ($self->{'global_config_file'}->{'output_template'} || './output_template'),
#               FILTERS => {
#                       'foo' => \&foo_filter,
#               },
                EVAL_PERL => 1,
        );

        # running with low_mem flag? well, use DBM::Deep then.
        if ($self->{'low_mem'}) {
                $log->info("running with low_mem which impacts performance (<32 Mb memory usage)");

                my $db_file = "data.db";

                if (-e $db_file) {
                        unlink $db_file or $log->logdie("can't remove '$db_file' from last run");
                        $log->debug("removed '$db_file' from last run");
                }

                require DBM::Deep;

                my $db = new DBM::Deep $db_file;

                $log->logdie("DBM::Deep error: $!") unless ($db);

                if ($db->error()) {
                        $log->logdie("can't open '$db_file' under low_mem: ",$db->error());
                } else {
                        $log->debug("using file '$db_file' for DBM::Deep");
                }

                $self->{'db'} = $db;
        }

        $log->debug("filters defined: ",Dumper($self->{'filter'}));

        return $self;
}

=head2 open_isis

Open CDS/ISIS database using OpenIsis module and read all records to memory.

 $webpac->open_isis(
        filename => '/data/ISIS/ISIS',
        code_page => '852',
        limit_mfn => 500,
        start_mfn => 6000,
        lookup => [ ... ],
 );

By default, ISIS code page is assumed to be C<852>.

If optional parametar C<start_mfn> is set, this will be first MFN to read
from database (so you can skip beginning of your database if you need to).

If optional parametar C<limit_mfn> is set, it will read just 500 records
from database in example above.

C<lookup> argument is an array of lookups to create. Each lookup must have C<key> and
C<val>. Optional parametar C<eval> is perl code to evaluate before storing
value in index.

 lookup => [
  { 'key' => 'd:v900', 'val' => 'v250^a' },
  { 'eval' => '"v901^a" eq "Podruèje"',
    'key' => 'pa:v561^4:v562^4:v461^1',
    'val' => 'v900' },
 ]

Returns number of last record read into memory (size of database, really).

=cut

sub open_isis {
        my $self = shift;
        my $arg = {@_};

        my $log = $self->_get_logger();

        $log->logcroak("need filename") if (! $arg->{'filename'});
        my $code_page = $arg->{'code_page'} || '852';

        $log->logdie("can't find database ",$arg->{'filename'}) unless (glob($arg->{'filename'}.'.*'));

        # store data in object
        $self->{'isis_filename'} = $arg->{'filename'};
        $self->{'isis_code_page'} = $code_page;

        use OpenIsis;

        #$self->{'isis_code_page'} = $code_page;

        # create Text::Iconv object
        my $cp = Text::Iconv->new($code_page,$self->{'code_page'});

        $log->info("reading ISIS database '",$arg->{'filename'},"'");
        $log->debug("isis code page: $code_page");

        my $isis_db = OpenIsis::open($arg->{'filename'});

        my $maxmfn = OpenIsis::maxRowid( $isis_db ) || 1;
        my $startmfn = 1;

        if (my $s = $self->{'start_mfn'}) {
                $log->info("skipping to MFN $s");
                $startmfn = $s;
        } else {
                $self->{'start_mfn'} = $startmfn;
        }

        $maxmfn = $startmfn + $self->{limit_mfn} if ($self->{limit_mfn});

        $log->info("processing ",($maxmfn-$startmfn)." records...");

        # read database
        for (my $mfn = $startmfn; $mfn <= $maxmfn; $mfn++) {


                $log->debug("mfn: $mfn\n");

                my $rec;

                # read record
                my $row = OpenIsis::read( $isis_db, $mfn );
                foreach my $k (keys %{$row}) {
                        if ($k ne "mfn") {
                                foreach my $l (@{$row->{$k}}) {
                                        $l = $cp->convert($l);
                                        # has subfields?
                                        my $val;
                                        if ($l =~ m/\^/) {
                                                foreach my $t (split(/\^/,$l)) {
                                                        next if (! $t);
                                                        $val->{substr($t,0,1)} = substr($t,1);
                                                }
                                        } else {
                                                $val = $l;
                                        }

                                        push @{$rec->{$k}}, $val;
                                }
                        } else {
                                push @{$rec->{'000'}}, $mfn;
                        }

                }

                $log->confess("record $mfn empty?") unless ($rec);

                # store
                if ($self->{'low_mem'}) {
                        $self->{'db'}->put($mfn, $rec);
                } else {
                        $self->{'data'}->{$mfn} = $rec;
                }

                # create lookup
                $self->create_lookup($rec, @{$arg->{'lookup'}});

                $self->progress_bar($mfn,$maxmfn);

        }

        $self->{'current_mfn'} = -1;
        $self->{'last_pcnt'} = 0;

        $log->debug("max mfn: $maxmfn");

        # store max mfn and return it.
        return $self->{'max_mfn'} = $maxmfn;
}

=head2 fetch_rec

Fetch next record from database. It will also display progress bar (once
it's implemented, that is).

 my $rec = $webpac->fetch_rec;

=cut

sub fetch_rec {
        my $self = shift;

        my $log = $self->_get_logger();

        $log->logconfess("it seems that you didn't load database!") unless ($self->{'current_mfn'});

        if ($self->{'current_mfn'} == -1) {
                $self->{'current_mfn'} = $self->{'start_mfn'};
        } else {
                $self->{'current_mfn'}++;
        }

        my $mfn = $self->{'current_mfn'};

        if ($mfn > $self->{'max_mfn'}) {
                $self->{'current_mfn'} = $self->{'max_mfn'};
                $log->debug("at EOF");
                return;
        }

        $self->progress_bar($mfn,$self->{'max_mfn'});

        if ($self->{'low_mem'}) {
                return $self->{'db'}->get($mfn);
        } else {
                return $self->{'data'}->{$mfn};
        }
}

=head2 mfn

Returns current record number (MFN).

 print $webpac->mfn;

=cut

sub mfn {
        my $self = shift;
        return $self->{'current_mfn'};
}

=head2 progress_bar

Draw progress bar on STDERR.

 $webpac->progress_bar($current, $max);

=cut

sub progress_bar {
        my $self = shift;

        my ($curr,$max) = @_;

        my $log = $self->_get_logger();

        $log->logconfess("no current value!") if (! $curr);
        $log->logconfess("no maximum value!") if (! $max);

        if ($curr > $max) {
                $max = $curr;
                $log->debug("overflow to $curr");
        }

        $self->{'last_pcnt'} ||= 1;

        my $p = int($curr * 100 / $max) || 1;

        # reset on re-run
        if ($p < $self->{'last_pcnt'}) {
                $self->{'last_pcnt'} = $p;
                $self->{'last_t'} = time();
                $self->{'last_curr'} = undef;
        }

        $self->{'last_t'} ||= time();

        if ($p != $self->{'last_pcnt'}) {

                my $last_curr = $self->{'last_curr'} || $curr;
                my $t = time();
                my $rate = ($curr - $last_curr) / (($t - $self->{'last_t'} || 1));
                my $eta = ($max-$curr) / ($rate || 1);
                printf STDERR ("%5d [%-38s] %-5d %0.1f/s %s\r",$curr,"=" x ($p/3)."$p%>", $max, $rate, $self->fmt_time($eta));
                $self->{'last_pcnt'} = $p;
                $self->{'last_t'} = time();
                $self->{'last_curr'} = $curr;
        }
        print STDERR "\n" if ($p == 100);
}

=head2 fmt_time

Format time (in seconds) for display.

 print $webpac->fmt_time(time());

This method is called by L<progress_bar> to display remaining time.

=cut

sub fmt_time {
        my $self = shift;

        my $t = shift || 0;
        my $out = "";

        my ($ss,$mm,$hh) = gmtime($t);
        $out .= "${hh}h" if ($hh);
        $out .= sprintf("%02d:%02d", $mm,$ss);
        $out .= "  " if ($hh == 0);
        return $out;
}

=head2 open_import_xml

Read file from C<import_xml/> directory and parse it.

 $webpac->open_import_xml(type => 'isis');

=cut

sub open_import_xml {
        my $self = shift;

        my $log = $self->_get_logger();

        my $arg = {@_};
        $log->logconfess("need type to load file from import_xml/") if (! $arg->{'type'});

        $self->{'type'} = $arg->{'type'};

        my $type_base = $arg->{'type'};
        $type_base =~ s/_.*$//g;

        $self->{'tag'} = $type2tag{$type_base};

        $log->info("using type '",$self->{'type'},"' tag <",$self->{'tag'},">");

        my $f = "./import_xml/".$self->{'type'}.".xml";
        $log->logconfess("import_xml file '$f' doesn't exist!") if (! -e "$f");

        $log->info("reading '$f'");

        $self->{'import_xml_file'} = $f;

        $self->{'import_xml'} = XMLin($f,
                ForceArray => [ $self->{'tag'}, 'config', 'format' ],
        );

        $log->debug("import xml is ",sub { Dumper($self->{'import_xml'}) });

}

=head2 create_lookup

Create lookup from record using lookup definition.

 $self->create_lookup($rec, @lookups);

Called internally by C<open_*> methods.

=cut

sub create_lookup {
        my $self = shift;

        my $log = $self->_get_logger();

        my $rec = shift || $log->logconfess("need record to create lookup");
        $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o);

        foreach my $i (@_) {
                $log->logconfess("need key") unless defined($i->{'key'});
                $log->logconfess("need val") unless defined($i->{'val'});

                if (defined($i->{'eval'})) {
                        # eval first, so we can skip fill_in for key and val
                        my $eval = $self->fill_in($rec,$i->{'eval'}) || next;
                        if ($self->_eval($eval)) {
                                my $key = $self->fill_in($rec,$i->{'key'}) || next;
                                my @val = $self->fill_in($rec,$i->{'val'}) || next;
                                $log->debug("stored $key = ",sub { join(" | ",@val) });
                                push @{$self->{'lookup'}->{$key}}, @val;
                        }
                } else {
                        my $key = $self->fill_in($rec,$i->{'key'}) || next;
                        my @val = $self->fill_in($rec,$i->{'val'}) || next;
                        $log->debug("stored $key = ",sub { join(" | ",@val) });
                        push @{$self->{'lookup'}->{$key}}, @val;
                }
        }
}

=head2 get_data

Returns value from record.

 my $text = $self->get_data(\$rec,$f,$sf,$i,\$found);

Arguments are:
record reference C<$rec>,
field C<$f>,
optional subfiled C<$sf>,
index for repeatable values C<$i>.

Optinal variable C<$found> will be incremeted if there
is field.

Returns value or empty string.

=cut

sub get_data {
        my $self = shift;

        my ($rec,$f,$sf,$i,$found) = @_;

        if ($$rec->{$f}) {
                return '' if (! $$rec->{$f}->[$i]);
                no strict 'refs';
                if ($sf && $$rec->{$f}->[$i]->{$sf}) {
                        $$found++ if (defined($$found));
                        return $$rec->{$f}->[$i]->{$sf};
                } elsif ($$rec->{$f}->[$i]) {
                        $$found++ if (defined($$found));
                        # it still might have subfield, just
                        # not specified, so we'll dump all
                        if ($$rec->{$f}->[$i] =~ /HASH/o) {
                                my $out;
                                foreach my $k (keys %{$$rec->{$f}->[$i]}) {
                                        $out .= $$rec->{$f}->[$i]->{$k}." ";
                                }
                                return $out;
                        } else {
                                return $$rec->{$f}->[$i];
                        }
                }
        } else {
                return '';
        }
}

=head2 fill_in

Workhourse of all: takes record from in-memory structure of database and
strings with placeholders and returns string or array of with substituted
values from record.

 my $text = $webpac->fill_in($rec,'v250^a');

Optional argument is ordinal number for repeatable fields. By default,
it's assume to be first repeatable field (fields are perl array, so first
element is 0).
Following example will read second value from repeatable field.

 my $text = $webpac->fill_in($rec,'Title: v250^a',1);

This function B<does not> perform parsing of format to inteligenty skip
delimiters before fields which aren't used.

This method will automatically decode UTF-8 string to local code page
if needed.

=cut

sub fill_in {
        my $self = shift;

        my $log = $self->_get_logger();

        my $rec = shift || $log->logconfess("need data record");
        my $format = shift || $log->logconfess("need format to parse");
        # iteration (for repeatable fields)
        my $i = shift || 0;

        # FIXME remove for speedup?
        $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o);

        if (utf8::is_utf8($format)) {
                $format = $self->_x($format);
        }

        my $found = 0;

        my $eval_code;
        # remove eval{...} from beginning
        $eval_code = $1 if ($format =~ s/^eval{([^}]+)}//s);

        my $filter_name;
        # remove filter{...} from beginning
        $filter_name = $1 if ($format =~ s/^filter{([^}]+)}//s);

        # do actual replacement of placeholders
        $format =~ s/v(\d+)(?:\^(\w))?/$self->get_data(\$rec,$1,$2,$i,\$found)/ges;

        if ($found) {
                $log->debug("format: $format");
                if ($eval_code) {
                        my $eval = $self->fill_in($rec,$eval_code,$i);
                        return if (! $self->_eval($eval));
                }
                if ($filter_name && $self->{'filter'}->{$filter_name}) {
                        $log->debug("filter '$filter_name' for $format");
                        $format = $self->{'filter'}->{$filter_name}->($format);
                        return unless(defined($format));
                        $log->debug("filter result: $format");
                }
                # do we have lookups?
                if ($format =~ /$LOOKUP_REGEX/o) {
                        $log->debug("format '$format' has lookup");
                        return $self->lookup($format);
                } else {
                        return $format;
                }
        } else {
                return;
        }
}

=head2 lookup

Perform lookups on format supplied to it.

 my $text = $self->lookup('[v900]');

Lookups can be nested (like C<[d:[a:[v900]]]>).

=cut

sub lookup {
        my $self = shift;

        my $log = $self->_get_logger();

        my $tmp = shift || $log->logconfess("need format");

        if ($tmp =~ /$LOOKUP_REGEX/o) {
                my @in = ( $tmp );

                $log->debug("lookup for: ",$tmp);

                my @out;
                while (my $f = shift @in) {
                        if ($f =~ /$LOOKUP_REGEX_SAVE/o) {
                                my $k = $1;
                                if ($self->{'lookup'}->{$k}) {
                                        foreach my $nv (@{$self->{'lookup'}->{$k}}) {
                                                my $tmp2 = $f;
                                                $tmp2 =~ s/lookup{$k}/$nv/g;
                                                push @in, $tmp2;
                                        }
                                } else {
                                        undef $f;
                                }
                        } elsif ($f) {
                                push @out, $f;
                        }
                }
                $log->logconfess("return is array and it's not expected!") unless wantarray;
                return @out;
        } else {
                return $tmp;
        }
}

=head2 parse

Perform smart parsing of string, skipping delimiters for fields which aren't
defined. It can also eval code in format starting with C<eval{...}> and
return output or nothing depending on eval code.

 my $text = $webpac->parse($rec,'eval{"v901^a" eq "Deskriptor"}descriptor: v250^a', $i);

=cut

sub parse {
        my $self = shift;

        my ($rec, $format_utf8, $i) = @_;

        return if (! $format_utf8);

        my $log = $self->_get_logger();

        $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o);
        $log->logconfess("need utf2cp Text::Iconv object!") if (! $self->{'utf2cp'});

        $i = 0 if (! $i);

        my $format = $self->_x($format_utf8) || $log->logconfess("can't convert '$format_utf8' from UTF-8 to ",$self->{'code_page'});

        my @out;

        $log->debug("format: $format");

        my $eval_code;
        # remove eval{...} from beginning
        $eval_code = $1 if ($format =~ s/^eval{([^}]+)}//s);

        my $filter_name;
        # remove filter{...} from beginning
        $filter_name = $1 if ($format =~ s/^filter{([^}]+)}//s);

        my $prefix;
        my $all_found=0;

        while ($format =~ s/^(.*?)v(\d+)(?:\^(\w))?//s) {

                my $del = $1 || '';
                $prefix ||= $del if ($all_found == 0);

                my $found = 0;
                my $tmp = $self->get_data(\$rec,$2,$3,$i,\$found);

                if ($found) {
                        push @out, $del;
                        push @out, $tmp;
                        $all_found += $found;
                }
        }

        return if (! $all_found);

        my $out = join('',@out);

        if ($out) {
                # add rest of format (suffix)
                $out .= $format;

                # add prefix if not there
                $out = $prefix . $out if ($out !~ m/^\Q$prefix\E/);

                $log->debug("result: $out");
        }

        if ($eval_code) {
                my $eval = $self->fill_in($rec,$eval_code,$i);
                $log->debug("about to eval{$eval} format: $out");
                return if (! $self->_eval($eval));
        }
        
        if ($filter_name && $self->{'filter'}->{$filter_name}) {
                $log->debug("about to filter{$filter_name} format: $out");
                $out = $self->{'filter'}->{$filter_name}->($out);
                return unless(defined($out));
                $log->debug("filter result: $out");
        }

        return $out;
}

=head2 parse_to_arr

Similar to C<parse>, but returns array of all repeatable fields

 my @arr = $webpac->parse_to_arr($rec,'v250^a');

=cut

sub parse_to_arr {
        my $self = shift;

        my ($rec, $format_utf8) = @_;

        my $log = $self->_get_logger();

        $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o);
        return if (! $format_utf8);

        my $i = 0;
        my @arr;

        while (my $v = $self->parse($rec,$format_utf8,$i++)) {
                push @arr, $v;
        }

        $log->debug("format '$format_utf8' returned ",--$i," elements: ", sub { join(" | ",@arr) }) if (@arr);

        return @arr;
}

=head2 fill_in_to_arr

Similar to C<fill_in>, but returns array of all repeatable fields. Usable
for fields which have lookups, so they shouldn't be parsed but rather
C<fill_id>ed.

 my @arr = $webpac->fill_in_to_arr($rec,'[v900];;[v250^a]');

=cut

sub fill_in_to_arr {
        my $self = shift;

        my ($rec, $format_utf8) = @_;

        my $log = $self->_get_logger();

        $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o);
        return if (! $format_utf8);

        my $i = 0;
        my @arr;

        while (my @v = $self->fill_in($rec,$format_utf8,$i++)) {
                push @arr, @v;
        }

        $log->debug("format '$format_utf8' returned ",--$i," elements: ", sub { join(" | ",@arr) }) if (@arr);

        return @arr;
}

=head2 sort_arr

Sort array ignoring case and html in data

 my @sorted = $webpac->sort_arr(@unsorted);

=cut

sub sort_arr {
        my $self = shift;

        my $log = $self->_get_logger();

        # FIXME add Schwartzian Transformation?

        my @sorted = sort {
                $a =~ s#<[^>]+/*>##;
                $b =~ s#<[^>]+/*>##;
                lc($b) cmp lc($a)
        } @_;
        $log->debug("sorted values: ",sub { join(", ",@sorted) });

        return @sorted;
}


=head2 data_structure

Create in-memory data structure which represents layout from C<import_xml>.
It is used later to produce output.

 my @ds = $webpac->data_structure($rec);

This method will also set C<$webpac->{'currnet_filename'}> if there is
<filename> tag in C<import_xml> and C<$webpac->{'headline'}> if there is
<headline> tag.

=cut

sub data_structure {
        my $self = shift;

        my $log = $self->_get_logger();

        my $rec = shift;
        $log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o);

        undef $self->{'currnet_filename'};
        undef $self->{'headline'};

        my @sorted_tags;
        if ($self->{tags_by_order}) {
                @sorted_tags = @{$self->{tags_by_order}};
        } else {
                @sorted_tags = sort { $self->_sort_by_order } keys %{$self->{'import_xml'}->{'indexer'}};
                $self->{tags_by_order} = \@sorted_tags;
        }

        my @ds;

        $log->debug("tags: ",sub { join(", ",@sorted_tags) });

        foreach my $field (@sorted_tags) {

                my $row;

#print "field $field [",$self->{'tag'},"] = ",Dumper($self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}});

                foreach my $tag (@{$self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}}) {
                        my $format = $tag->{'value'} || $tag->{'content'};

                        $log->debug("format: $format");

                        my @v;
                        if ($format =~ /$LOOKUP_REGEX/o) {
                                @v = $self->fill_in_to_arr($rec,$format);
                        } else {
                                @v = $self->parse_to_arr($rec,$format);
                        }
                        next if (! @v);

                        if ($tag->{'sort'}) {
                                @v = $self->sort_arr(@v);
                                $log->warn("sort within tag is usually not what you want!");
                        }

                        # use format?
                        if ($tag->{'format_name'}) {
                                @v = map { $self->apply_format($tag->{'format_name'},$tag->{'format_delimiter'},$_) } @v;
                        }

                        if ($field eq 'filename') {
                                $self->{'current_filename'} = join('',@v);
                                $log->debug("filename: ",$self->{'current_filename'});
                        } elsif ($field eq 'headline') {
                                $self->{'headline'} .= join('',@v);
                                $log->debug("headline: ",$self->{'headline'});
                                next; # don't return headline in data_structure!
                        }

                        # delimiter will join repeatable fields
                        if ($tag->{'delimiter'}) {
                                @v = ( join($tag->{'delimiter'}, @v) );
                        }

                        # default types 
                        my @types = qw(display swish);
                        # override by type attribute
                        @types = ( $tag->{'type'} ) if ($tag->{'type'});

                        foreach my $type (@types) {
                                # append to previous line?
                                $log->debug("type: $type ",sub { join(" ",@v) }, $row->{'append'} || 'no append');
                                if ($tag->{'append'}) {

                                        # I will delimit appended part with
                                        # delimiter (or ,)
                                        my $d = $tag->{'delimiter'};
                                        # default delimiter
                                        $d ||= ", ";

                                        my $last = pop @{$row->{$type}};
                                        $d = "" if (! $last);
                                        $last .= $d . join($d, @v);
                                        push @{$row->{$type}}, $last;

                                } else {
                                        push @{$row->{$type}}, @v;
                                }
                        }


                }

                if ($row) {
                        $row->{'tag'} = $field;

                        # TODO: name_sigular, name_plural
                        my $name = $self->{'import_xml'}->{'indexer'}->{$field}->{'name'};
                        $row->{'name'} = $name ? $self->_x($name) : $field;

                        # post-sort all values in field
                        if ($self->{'import_xml'}->{'indexer'}->{$field}->{'sort'}) {
                                $log->warn("sort at field tag not implemented");
                        }

                        push @ds, $row;

                        $log->debug("row $field: ",sub { Dumper($row) });
                }

        }

        return @ds;

}

=head2 output

Create output from in-memory data structure using Template Toolkit template.

my $text = $webpac->output( template => 'text.tt', data => @ds );

=cut

sub output {
        my $self = shift;

        my $args = {@_};

        my $log = $self->_get_logger();

        $log->logconfess("need template name") if (! $args->{'template'});
        $log->logconfess("need data array") if (! $args->{'data'});

        my $out;

        $self->{'tt'}->process(
                $args->{'template'},
                $args,
                \$out
        ) || confess $self->{'tt'}->error();

        return $out;
}

=head2 output_file

Create output from in-memory data structure using Template Toolkit template
to a file.

 $webpac->output_file(
        file => 'out.txt',
        template => 'text.tt',
        data => @ds
 );

=cut

sub output_file {
        my $self = shift;

        my $args = {@_};

        my $log = $self->_get_logger();

        my $file = $args->{'file'} || $log->logconfess("need file name");

        $log->debug("creating file ",$file);

        open(my $fh, ">", $file) || $log->logdie("can't open output file '$file': $!");
        print $fh $self->output(
                template => $args->{'template'},
                data => $args->{'data'},
        ) || $log->logdie("print: $!");
        close($fh) || $log->logdie("close: $!");
}

=head2 apply_format

Apply format specified in tag with C<format_name="name"> and
C<format_delimiter=";;">.

 my $text = $webpac->apply_format($format_name,$format_delimiter,$data);

Formats can contain C<lookup{...}> if you need them.

=cut

sub apply_format {
        my $self = shift;

        my ($name,$delimiter,$data) = @_;

        my $log = $self->_get_logger();

        if (! $self->{'import_xml'}->{'format'}->{$name}) {
                $log->warn("<format name=\"$name\"> is not defined in ",$self->{'import_xml_file'});
                return $data;
        }

        $log->warn("no delimiter for format $name") if (! $delimiter);

        my $format = $self->_x($self->{'import_xml'}->{'format'}->{$name}->{'content'}) || $log->logdie("can't find format '$name'");

        my @data = split(/\Q$delimiter\E/, $data);

        my $out = sprintf($format, @data);
        $log->debug("using format $name [$format] on $data to produce: $out");

        if ($out =~ m/$LOOKUP_REGEX/o) {
                return $self->lookup($out);
        } else {
                return $out;
        }

}


#
#
#

=head1 INTERNAL METHODS

Here is a quick list of internal methods, mostly useful to turn debugging
on them (see L<LOGGING> below for explanation).

=cut

=head2 _eval

Internal function to eval code without C<strict 'subs'>.

=cut

sub _eval {
        my $self = shift;

        my $code = shift || return;

        my $log = $self->_get_logger();

        no strict 'subs';
        my $ret = eval $code;
        if ($@) {
                $log->error("problem with eval code [$code]: $@");
        }

        $log->debug("eval: ",$code," [",$ret,"]");

        return $ret || undef;
}

=head2 _sort_by_order

Sort xml tags data structure accoding to C<order=""> attribute.

=cut

sub _sort_by_order {
        my $self = shift;

        my $va = $self->{'import_xml'}->{'indexer'}->{$a}->{'order'} ||
                $self->{'import_xml'}->{'indexer'}->{$a};
        my $vb = $self->{'import_xml'}->{'indexer'}->{$b}->{'order'} ||
                $self->{'import_xml'}->{'indexer'}->{$b};

        return $va <=> $vb;
}

=head2 _get_logger

Get C<Log::Log4perl> object with a twist: domains are defined for each
method

 my $log = $webpac->_get_logger();

=cut

sub _get_logger {
        my $self = shift;

        my $name = (caller(1))[3] || caller;
        return get_logger($name);
}

=head2 _x

Convert string from UTF-8 to code page defined in C<import_xml>.

 my $text = $webpac->_x('utf8 text');

=cut

sub _x {
        my $self = shift;
        my $utf8 = shift || return;

        return $self->{'utf2cp'}->convert($utf8) ||
                $self->_get_logger()->logwarn("can't convert '$utf8'");
}

#
#
#

=head1 LOGGING

Logging in WebPAC is performed by L<Log::Log4perl> with config file
C<log.conf>.

Methods defined above have different levels of logging, so
it's descriptions will be useful to turn (mostry B<debug> logging) on
or off to see why WabPAC isn't perforing as you expect it (it might even
be a bug!).

B<This is different from normal Log4perl behaviour>. To repeat, you can
also use method names, and not only classes (which are just few)
to filter logging.


=head1 MEMORY USAGE

C<low_mem> options is double-edged sword. If enabled, WebPAC
will run on memory constraint machines (which doesn't have enough
physical RAM to create memory structure for whole source database).

If your machine has 512Mb or more of RAM and database is around 10000 records,
memory shouldn't be an issue. If you don't have enough physical RAM, you
might consider using virtual memory (if your operating system is handling it
well, like on FreeBSD or Linux) instead of dropping to L<DBD::Deep> to handle
parsed structure of ISIS database (this is what C<low_mem> option does).

Hitting swap at end of reading source database is probably o.k. However,
hitting swap before 90% will dramatically decrease performance and you will
be better off with C<low_mem> and using rest of availble memory for
operating system disk cache (Linux is particuallary good about this).
However, every access to database record will require disk access, so
generation phase will be slower 10-100 times.

Parsed structures are essential - you just have option to trade RAM memory
(which is fast) for disk space (which is slow). Be sure to have planty of
disk space if you are using C<low_mem> and thus L<DBD::Deep>.

However, when WebPAC is running on desktop machines (or laptops :-), it's
highly undesireable for system to start swapping. Using C<low_mem> option can
reduce WecPAC memory usage to around 64Mb for same database with lookup
fields and sorted indexes which stay in RAM. Performance will suffer, but
memory usage will really be minimal. It might be also more confortable to
run WebPAC reniced on those machines.

=cut

1;