--- trunk/IsisDB.pm 2005/01/05 21:23:04 33 +++ trunk/lib/Biblio/Isis.pm 2005/03/12 21:05:29 41 @@ -1,4 +1,4 @@ -package IsisDB; +package Biblio::Isis; use strict; use Carp; @@ -9,7 +9,7 @@ BEGIN { use Exporter (); use vars qw ($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS); - $VERSION = 0.09; + $VERSION = 0.13; @ISA = qw (Exporter); #Give a hoot don't pollute, do not export more than needed by default @EXPORT = qw (); @@ -20,13 +20,13 @@ =head1 NAME -IsisDB - Read CDS/ISIS, WinISIS and IsisMarc database +Biblio::Isis - Read CDS/ISIS, WinISIS and IsisMarc database =head1 SYNOPSIS - use IsisDB; + use Biblio::Isis; - my $isis = new IsisDB( + my $isis = new Biblio::Isis( isisdb => './cds/cds', ); @@ -81,7 +81,7 @@ Open ISIS database - my $isis = new IsisDB( + my $isis = new Biblio::Isis( isisdb => './cds/cds', read_fdt => 1, include_deleted => 1, @@ -147,7 +147,10 @@ push @must_exist, "fdt" if ($self->{read_fdt}); foreach my $ext (@must_exist) { - croak "missing ",uc($ext)," file in ",$self->{isisdb} unless ($self->{$ext."_file"}); + unless ($self->{$ext."_file"}) { + carp "missing ",uc($ext)," file in ",$self->{isisdb}; + return; + } } print STDERR "## using files: ",join(" ",@isis_files),"\n" if ($self->{debug}); @@ -192,12 +195,12 @@ # NXTMFB* last block allocated to master file # NXTMFP offset to next available position in last block # MFTYPE always 0 for user db file (1 for system) - seek($self->{'fileMST'},4,0) || carp "can't seek to offset 0 in MST: $!"; + seek($self->{'fileMST'},4,0) || croak "can't seek to offset 0 in MST: $!"; my $buff; - read($self->{'fileMST'}, $buff, 4) || carp "can't read NXTMFN from MST: $!"; - $self->{'NXTMFN'}=unpack("V",$buff) || carp "NXTNFN is zero"; + read($self->{'fileMST'}, $buff, 4) || croak "can't read NXTMFN from MST: $!"; + $self->{'NXTMFN'}=unpack("V",$buff) || croak "NXTNFN is zero"; print STDERR Dumper($self),"\n" if ($self->{debug}); @@ -221,80 +224,6 @@ return $self->{'NXTMFN'} - 1; } -=head2 read_cnt - -Read content of C<.CNT> file and return hash containing it. - - print Dumper($isis->read_cnt); - -This function is not used by module (C<.CNT> files are not required for this -module to work), but it can be useful to examine your index (while debugging -for example). - -=cut - -sub read_cnt { - my $self = shift; - - croak "missing CNT file in ",$self->{isisdb} unless ($self->{cnt_file}); - - # Get the index information from $db.CNT - - open(my $fileCNT, $self->{cnt_file}) || carp "can't read '$self->{cnt_file}': $!"; - binmode($fileCNT); - - my $buff; - - read($fileCNT, $buff, 26) || carp "can't read first table from CNT: $!"; - $self->unpack_cnt($buff); - - read($fileCNT, $buff, 26) || carp "can't read second table from CNT: $!"; - $self->unpack_cnt($buff); - - close($fileCNT); - - return $self->{cnt}; -} - -=head2 unpack_cnt - -Unpack one of two 26 bytes fixed length record in C<.CNT> file. - -Here is definition of record: - - off key description size - 0: IDTYPE BTree type s - 2: ORDN Nodes Order s - 4: ORDF Leafs Order s - 6: N Number of Memory buffers for nodes s - 8: K Number of buffers for first level index s - 10: LIV Current number of Index Levels s - 12: POSRX Pointer to Root Record in N0x l - 16: NMAXPOS Next Available position in N0x l - 20: FMAXPOS Next available position in L0x l - 24: ABNORMAL Formal BTree normality indicator s - length: 26 bytes - -This will fill C<$self> object under C with hash. It's used by C. - -=cut - -sub unpack_cnt { - my $self = shift; - - my @flds = qw(ORDN ORDF N K LIV POSRX NMAXPOS FMAXPOS ABNORMAL); - - my $buff = shift || return; - my @arr = unpack("vvvvvvVVVv", $buff); - - print STDERR "unpack_cnt: ",join(" ",@arr),"\n" if ($self->{'debug'}); - - my $IDTYPE = shift @arr; - foreach (@flds) { - $self->{cnt}->{$IDTYPE}->{$_} = abs(shift @arr); - } -} - =head2 fetch Read record with selected MFN @@ -335,7 +264,15 @@ # read XRFMFB abd XRFMFP read($self->{'fileXRF'}, $buff, 4); - my $pointer=unpack("V",$buff) || carp "pointer is null"; + my $pointer=unpack("V",$buff); + if (! $pointer) { + if ($self->{include_deleted}) { + return; + } else { + warn "pointer for MFN $mfn is null\n"; + return; + } + } # check for logically deleted record if ($pointer & 0x80000000) { @@ -464,7 +401,7 @@ my $mfn = shift || croak "need MFN"; - my $rec = $self->fetch($mfn); + my $rec = $self->fetch($mfn) || return; my $out = "0\t$mfn"; @@ -532,7 +469,7 @@ # init record to include MFN as field 000 my $rec = { '000' => [ $mfn ] }; - my $row = $self->fetch($mfn); + my $row = $self->fetch($mfn) || return; foreach my $k (keys %{$row}) { foreach my $l (@{$row->{$k}}) { @@ -576,6 +513,81 @@ return $self->{'TagName'}->{$tag} || $tag; } + +=head2 read_cnt + +Read content of C<.CNT> file and return hash containing it. + + print Dumper($isis->read_cnt); + +This function is not used by module (C<.CNT> files are not required for this +module to work), but it can be useful to examine your index (while debugging +for example). + +=cut + +sub read_cnt { + my $self = shift; + + croak "missing CNT file in ",$self->{isisdb} unless ($self->{cnt_file}); + + # Get the index information from $db.CNT + + open(my $fileCNT, $self->{cnt_file}) || croak "can't read '$self->{cnt_file}': $!"; + binmode($fileCNT); + + my $buff; + + read($fileCNT, $buff, 26) || croak "can't read first table from CNT: $!"; + $self->unpack_cnt($buff); + + read($fileCNT, $buff, 26) || croak "can't read second table from CNT: $!"; + $self->unpack_cnt($buff); + + close($fileCNT); + + return $self->{cnt}; +} + +=head2 unpack_cnt + +Unpack one of two 26 bytes fixed length record in C<.CNT> file. + +Here is definition of record: + + off key description size + 0: IDTYPE BTree type s + 2: ORDN Nodes Order s + 4: ORDF Leafs Order s + 6: N Number of Memory buffers for nodes s + 8: K Number of buffers for first level index s + 10: LIV Current number of Index Levels s + 12: POSRX Pointer to Root Record in N0x l + 16: NMAXPOS Next Available position in N0x l + 20: FMAXPOS Next available position in L0x l + 24: ABNORMAL Formal BTree normality indicator s + length: 26 bytes + +This will fill C<$self> object under C with hash. It's used by C. + +=cut + +sub unpack_cnt { + my $self = shift; + + my @flds = qw(ORDN ORDF N K LIV POSRX NMAXPOS FMAXPOS ABNORMAL); + + my $buff = shift || return; + my @arr = unpack("vvvvvvVVVv", $buff); + + print STDERR "unpack_cnt: ",join(" ",@arr),"\n" if ($self->{'debug'}); + + my $IDTYPE = shift @arr; + foreach (@flds) { + $self->{cnt}->{$IDTYPE}->{$_} = abs(shift @arr); + } +} + 1; =head1 BUGS