--- trunk/lib/WebPAC/Input/ISIS.pm	2005/12/16 16:23:44	265
+++ trunk/lib/WebPAC/Input/ISIS.pm	2006/09/07 15:01:45	652
@@ -3,318 +3,123 @@
 use warnings;
 use strict;
 
-use WebPAC::Common;
-use base qw/WebPAC::Input WebPAC::Common/;
-use Text::Iconv;
+use WebPAC::Input;
+use Biblio::Isis 0.23;
 
 =head1 NAME
 
-WebPAC::Input::ISIS - support for CDS/ISIS source files
+WebPAC::Input::ISIS - support for CDS/ISIS database files
 
 =head1 VERSION
 
-Version 0.01
+Version 0.07
 
 =cut
 
-our $VERSION = '0.01';
+our $VERSION = '0.07';
 
 
-# auto-configure
-
-my ($have_biblio_isis, $have_openisis) = (0,0);
-
-eval "use Biblio::Isis 0.13;";
-unless ($@) { 
-	$have_biblio_isis = 1
-} else {
-	eval "use OpenIsis;";
-	$have_openisis = 1 unless ($@);
-}
-
 =head1 SYNOPSIS
 
-Open CDS/ISIS, WinISIS or IsisMarc database using Biblio::Isis or OpenIsis
-module and read all records to memory.
+Open CDS/ISIS, WinISIS or IsisMarc database using C<Biblio::Isis>
+and read all records to memory.
 
  my $isis = new WebPAC::Input::ISIS();
- $isis->open( filename => '/path/to/ISIS/ISIS' );
+ $isis->open( path => '/path/to/ISIS/ISIS' );
 
 =head1 FUNCTIONS
 
-=head2 open
-
-This function will read whole database in memory and produce lookups.
-
- $isis->open(
- 	filename => '/data/ISIS/ISIS',
-	code_page => '852',
-	limit_mfn => 500,
-	start_mfn => 6000,
-	lookup => $lookup_obj,
- );
-
-By default, ISIS code page is assumed to be C<852>.
-
-If optional parametar C<start_mfn> is set, this will be first MFN to read
-from database (so you can skip beginning of your database if you need to).
-
-If optional parametar C<limit_mfn> is set, it will read just 500 records
-from database in example above.
-
-Returns size of database, regardless of C<start_mfn> and C<limit_mfn>
-parametars, see also C<$isis->size>.
-
-=cut
-
-sub open {
-	my $self = shift;
-	my $arg = {@_};
-
-	my $log = $self->_get_logger();
-
-	$log->logcroak("need filename") if (! $arg->{'filename'});
-	my $code_page = $arg->{'code_page'} || '852';
-
-	# store data in object
-	$self->{'isis_code_page'} = $code_page;
-	foreach my $v (qw/isis_filename start_mfn limit_mfn/) {
-		$self->{$v} = $arg->{$v} if ($arg->{$v});
-	}
-
-	# 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,$db_size);
-
-	if ($have_openisis) {
-		$log->debug("using OpenIsis perl bindings");
-		$isis_db = OpenIsis::open($arg->{'filename'});
-		$db_size = OpenIsis::maxRowid( $isis_db ) || 1;
-	} elsif ($have_biblio_isis) {
-		$log->debug("using Biblio::Isis");
-		use Biblio::Isis;
-		$isis_db = new Biblio::Isis(
-			isisdb => $arg->{'filename'},
-			include_deleted => 1,
-			hash_filter => sub {
-				my $l = shift || return;
-				$l = $cp->convert($l);
-				return $l;
-			},
-		) or $log->logdie("can't find database ",$arg->{'filename'});
-
-		$db_size = $isis_db->count;
-
-		unless ($db_size) {
-			$log->logwarn("no records in database ", $arg->{'filename'}, ", skipping...");
-			return;
-		}
-
-	} else {
-		$log->logdie("Can't find supported ISIS library for perl. I suggent that you install Bilbio::Isis from CPAN.");
-	}
-
-
-	my $startmfn = 1;
-	my $maxmfn = $db_size;
-
-	if (my $s = $self->{'start_mfn'}) {
-		$log->info("skipping to MFN $s");
-		$startmfn = $s;
-	} else {
-		$self->{'start_mfn'} = $startmfn;
-	}
-
-	if ($self->{limit_mfn}) {
-		$log->info("limiting to ",$self->{limit_mfn}," records");
-		$maxmfn = $startmfn + $self->{limit_mfn} - 1;
-		$maxmfn = $db_size if ($maxmfn > $db_size);
-	}
-
-	# store size for later
-	$self->{'size'} = ($maxmfn - $startmfn) ? ($maxmfn - $startmfn + 1) : 0;
-
-	$log->info("processing ",($maxmfn-$startmfn)." records using ",( $have_openisis ? 'OpenIsis' : 'Biblio::Isis'));
-
-
-	# read database
-	for (my $mfn = $startmfn; $mfn <= $maxmfn; $mfn++) {
-
-		$log->debug("mfn: $mfn\n");
-
-		my $rec;
-
-		if ($have_openisis) {
-
-			# read record using OpenIsis
-			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'}}, $mfn;
-				}
-			}
-
-		} elsif ($have_biblio_isis) {
-			$rec = $isis_db->to_hash($mfn);
-		} else {
-			$log->logdie("hum? implementation missing?");
-		}
-
-		if (! $rec) {
-			$log->warn("record $mfn empty? skipping...");
-			next;
-		}
-
-		# store
-		if ($self->{'low_mem'}) {
-			$self->{'db'}->put($mfn, $rec);
-		} else {
-			$self->{'data'}->{$mfn} = $rec;
-		}
-
-		# create lookup
-		$self->{'lookup'}->add( $rec ) if ($rec && $self->{'lookup'});
-
-		$self->progress_bar($mfn,$maxmfn);
-
-	}
-
-	$self->{'current_mfn'} = -1;
-	$self->{'last_pcnt'} = 0;
+=head2 open_db
 
-	$log->debug("max mfn: $maxmfn");
+Returns handle to database and size in records
 
-	# store max mfn and return it.
-	$self->{'max_mfn'} = $maxmfn;
+  my ($db,$size) = $isis->open_db(
+  	path => '/path/to/LIBRI'
+	filter => sub {
+		my ($l,$field_nr) = @_;
+		# do something with $l which is line of input file
+		return $l;
+	},
+  }
 
-	return $db_size;
-}
+Options:
 
-=head2 fetch
+=over 4
 
-Fetch next record from database. It will also displays progress bar.
+=item path
 
- my $rec = $isis->fetch;
+path to CDS/ISIS database
 
-Record from this function should probably go to C<data_structure> for
-normalisation.
+=back
 
 =cut
 
-sub fetch {
+sub open_db {
 	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 $arg = {@_};
 
-	my $mfn = $self->{'current_mfn'};
+	my $log = $self->_get_logger();
 
-	if ($mfn > $self->{'max_mfn'}) {
-		$self->{'current_mfn'} = $self->{'max_mfn'};
-		$log->debug("at EOF");
-		return;
-	}
+	$log->info("opening ISIS database '$arg->{path}'");
 
-	$self->progress_bar($mfn,$self->{'max_mfn'});
+	$log->debug("using Biblio::Isis");
+	my $isis_db = new Biblio::Isis(
+		isisdb => $arg->{path},
+		include_deleted => 1,
+		hash_filter => $arg->{filter} ? sub { return $arg->{filter}->(@_); } : undef,
+	) or $log->logdie("can't find database $arg->{path}");
 
-	my $rec;
+	my $size = $isis_db->count;
 
-	if ($self->{'low_mem'}) {
-		$rec = $self->{'db'}->get($mfn);
-	} else {
-		$rec = $self->{'data'}->{$mfn};
-	}
+	$self->{_isis_db} = $isis_db;
 
-	$rec ||= 0E0;
+	return ($isis_db, $size);
 }
 
-=head2 pos
-
-Returns current record number (MFN).
+=head2 fetch_rec
 
- print $isis->pos;
+Return record with ID C<$mfn> from database
 
-First record in database has position 1.
+  my $rec = $self->fetch_rec( $mfn, $filter_coderef);
 
 =cut
 
-sub pos {
+sub fetch_rec {
 	my $self = shift;
-	return $self->{'current_mfn'};
-}
-
-
-=head2 size
 
-Returns number of records in database
+	my ($mfn, $filter_coderef) = @_;
 
- print $isis->size;
+	my $rec = $self->{_isis_db}->to_hash({
+		mfn => $mfn,
+		include_subfields => 1,
+		hash_filter => $filter_coderef,
+#		hash_filter => sub {
+#			my ($l,$f) = @_;
+#			warn "## in hash_filter ($l,$f)\n";
+#			my $o = $filter_coderef->($l,$f) if ($filter_coderef);
+#			warn "## out hash_filter = $o\n";
+#			return $o;
+#		},
+	});
 
-Result from this function can be used to loop through all records
-
- foreach my $mfn ( 1 ... $isis->size ) { ... }
-
-because it takes into account C<start_mfn> and C<limit_mfn>.
-
-=cut
-
-sub size {
-	my $self = shift;
-	return $self->{'size'};
+	return $rec;
 }
 
-=head2 seek
-
-Seek to specified MFN in file.
+=head2 dump_rec
 
- $isis->seek(42);
+Return dump of record ID C<$mfn> from database
 
-First record in database has position 1.
+  my $rec = $self->dump_rec( $db, $mfn );
 
 =cut
 
-sub seek {
+sub dump_rec {
 	my $self = shift;
-	my $pos = shift || return;
-
-	my $log = $self->_get_logger();
 
-	if ($pos < 1) {
-		$log->warn("seek before first record");
-		$pos = 1;
-	} elsif ($pos > $self->{'max_mfn'}) {
-		$log->warn("seek beyond last record");
-		$pos = $self->{'max_mfn'};
-	}
+	my $mfn = shift;
 
-	return $self->{'current_mfn'} = (($pos - 1) || -1);
+	return $self->{_isis_db}->to_ascii( $mfn );
 }
 
 =head1 AUTHOR