package IsisDB; use strict; use Carp; use Data::Dumper; BEGIN { use Exporter (); use vars qw ($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS); $VERSION = 0.03; @ISA = qw (Exporter); #Give a hoot don't pollute, do not export more than needed by default @EXPORT = qw (); @EXPORT_OK = qw (); %EXPORT_TAGS = (); } =head1 NAME IsisDB - Read CDS/ISIS database =head1 SYNOPSIS use IsisDB my $isis = new IsisDB( isisdb => './cds/cds', ); =head1 DESCRIPTION This module will read CDS/ISIS databases and create hash values out of it. It can be used as perl-only alternative to OpenIsis module. =head1 METHODS =cut # my $ORDN; # Nodes Order # my $ORDF; # Leafs Order # my $N; # Number of Memory buffers for nodes # my $K; # Number of buffers for first level index # my $LIV; # Current number of Index Levels # my $POSRX; # Pointer to Root Record in N0x # my $NMAXPOS; # Next Available position in N0x # my $FMAXPOS; # Next available position in L0x # my $ABNORMAL; # Formal BTree normality indicator # # some binary reads # sub Read32 { my $self = shift; my $f = shift || die "Read32 needs file handle"; read($$f,$b,4) || die "can't read 4 bytes from $$f from position ".tell($f); return unpack("l",$b); } =head2 new Open CDS/ISIS database my $isis = new IsisDB( isisdb => './cds/cds', read_fdt => 1, debug => 1, include_deleted => 1, ); Options are described below: =over 5 =item isisdb Prefix path to CDS/ISIS. It should contain full or relative path to database and common prefix of C<.FDT>, C<.MST>, C<.CNT>, C<.XRF> and C<.MST> files. =item read_fdt Boolean flag to specify if field definition table should be read. It's off by default. =item debug Dump a C of debugging output. =item include_deleted Don't skip logically deleted records. =back It will also set C<$isis-E{'maxmfn'}> which is maximum MFN stored in database. =cut sub new { my $class = shift; my $self = {}; bless($self, $class); croak "new needs database name (isisdb) as argument!" unless ({@_}->{isisdb}); foreach my $v (qw{isisdb debug include_deleted}) { $self->{$v} = {@_}->{$v}; } # if you want to read .FDT file use read_fdt argument when creating class! if ({@_}->{read_fdt} && -e $self->{isisdb}.".FDT") { # read the $db.FDT file for tags my $fieldzone=0; open(fileFDT, $self->{isisdb}.".FDT") || croak "can't read '$self->{isisdb}.FDT': $!"; while () { chomp; if ($fieldzone) { my $name=substr($_,0,30); my $tag=substr($_,50,3); $name =~ s/\s+$//; $tag =~ s/\s+$//; $self->{'TagName'}->{$tag}=$name; } if (/^\*\*\*/) { $fieldzone=1; } } close(fileFDT); } # Get the Maximum MFN from $db.MST open(fileMST,$self->{isisdb}.".MST") || croak "can't read '$self->{isisdb}.MST': $!"; # MST format: (* = 32 bit signed) # CTLMFN* always 0 # NXTMFN* MFN to be assigned to the next record created # 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(fileMST,4,0); $self->{'NXTMFN'}=$self->Read32(\*fileMST) || carp "NXTNFN is zero"; # save maximum MFN $self->{'maxmfn'} = $self->{'NXTMFN'} - 1; close(fileMST); # Get the index information from $db.CNT open(fileCNT, $self->{isisdb}.".CNT") || croak "can't read '$self->{isisdb}.CNT': $!"; # There is two 26 Bytes fixed lenght records # 0: IDTYPE BTree type 16 # 2: ORDN Nodes Order 16 # 4: ORDF Leafs Order 16 # 6: N Number of Memory buffers for nodes 16 # 8: K Number of buffers for first level index 16 # 10: LIV Current number of Index Levels 16 # 12: POSRX* Pointer to Root Record in N0x 32 # 16: NMAXPOS* Next Available position in N0x 32 # 20: FMAXPOS* Next available position in L0x 32 # 24: ABNORMAL Formal BTree normality indicator 16 # length: 26 bytes 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("ssssssllls", $buff); print "unpack_cnt: ",join(" ",@arr),"\n" if ($self->{'debug'}); my $IDTYPE = shift @arr; foreach (@flds) { $self->{$IDTYPE}->{$_} = abs(shift @arr); } } my $buff; read(fileCNT, $buff, 26); $self->unpack_cnt($buff); read(fileCNT, $buff, 26); $self->unpack_cnt($buff); close(fileCNT); print Dumper($self) if ($self->{debug}); # open files for later open($self->{'fileXRF'}, $self->{isisdb}.".XRF") || croak "can't open '$self->{isisdb}.XRF': $!"; open($self->{'fileMST'}, $self->{isisdb}.".MST") || croak "can't open '$self->{isisdb}.MST': $!"; $self ? return $self : return undef; } =head2 fetch Read record with selected MFN my $rec = $isis->fetch(55); Returns hash with keys which are field names and values are unpacked values for that field. =cut sub fetch { my $self = shift; my $mfn = shift || croak "fetch needs MFN as argument!"; print "fetch: $mfn\n" if ($self->{debug}); # XXX check this? my $mfnpos=($mfn+int(($mfn-1)/127))*4; print "seeking to $mfnpos in file '$self->{isisdb}.XRF'\n" if ($self->{debug}); seek($self->{'fileXRF'},$mfnpos,0); # read XRFMFB abd XRFMFP my $pointer=$self->Read32(\*{$self->{'fileXRF'}}); my $XRFMFB = int($pointer/2048); my $XRFMFP = $pointer - ($XRFMFB*2048); print "XRFMFB: $XRFMFB XRFMFP: $XRFMFP\n" if ($self->{debug}); # XXX fix this to be more readable!! # e.g. (XRFMFB - 1) * 512 + XRFMFP my $offset = $pointer; my $offset2=int($offset/2048)-1; my $offset22=int($offset/4096); my $offset3=$offset-($offset22*4096); if ($offset3>512) { $offset3=$offset3-2048; } my $offset4=($offset2*512)+$offset3; print "$offset - $offset2 - $offset3 - $offset4\n" if ($self->{debug}); # Get Record Information seek($self->{'fileMST'},$offset4,0); my $value=$self->Read32(\*{$self->{'fileMST'}}); if ($value!=$mfn) { print ("Error: The MFN:".$mfn." is not found in MST(".$value.")"); return -1; # XXX deleted record? } # $MFRL=$self->Read16($fileMST); # $MFBWB=$self->Read32($fileMST); # $MFBWP=$self->Read16($fileMST); # $BASE=$self->Read16($fileMST); # $NVF=$self->Read16($fileMST); # $STATUS=$self->Read16($fileMST); my $buff; read($self->{'fileMST'}, $buff, 14); my ($MFRL,$MFBWB,$MFBWP,$BASE,$NVF,$STATUS) = unpack("slssss", $buff); print "MFRL: $MFRL MFBWB: $MFBWB MFBWP: $MFBWP BASE: $BASE NVF: $NVF STATUS: $STATUS\n" if ($self->{debug}); # delete old record delete $self->{record}; if (! $self->{'include_deleted'} && $MFRL < 0) { print "## logically deleted record $mfn, skipping...\n" if ($self->{debug}); return; } # Get Directory Format my @FieldPOS; my @FieldLEN; my @FieldTAG; read($self->{'fileMST'}, $buff, 6 * $NVF); my $fld_len = 0; for (my $i = 0 ; $i < $NVF ; $i++) { # $TAG=$self->Read16($fileMST); # $POS=$self->Read16($fileMST); # $LEN=$self->Read16($fileMST); my ($TAG,$POS,$LEN) = unpack("sss", substr($buff,$i * 6, 6)); print "TAG: $TAG POS: $POS LEN: $LEN\n" if ($self->{debug}); # The TAG does not exists in .FDT so we set it to 0. # # XXX This is removed from perl version; .FDT file is updated manually, so # you will often have fields in .MST file which aren't in .FDT. On the other # hand, IsisMarc doesn't use .FDT files at all! #if (! $self->{TagName}->{$TAG}) { # $TAG=0; #} push @FieldTAG,$TAG; push @FieldPOS,$POS; push @FieldLEN,$LEN; $fld_len += $LEN; } # Get Variable Fields read($self->{'fileMST'},$buff,$fld_len); for (my $i = 0 ; $i < $NVF ; $i++) { push @{$self->{record}->{$FieldTAG[$i]}}, substr($buff,$FieldPOS[$i],$FieldLEN[$i]); } close(fileMST); print Dumper($self) if ($self->{debug}); return $self->{'record'}; } =head2 to_ascii Dump ascii output of selected MFN print $isis->to_ascii(55); =cut sub to_ascii { my $self = shift; my $mfn = shift || croak "need MFN"; my $rec = $self->fetch($mfn); my $out = "0\t$mfn"; foreach my $f (sort keys %{$rec}) { $out .= "\n$f\t".join("\n$f\t",@{$self->{record}->{$f}}); } $out .= "\n"; return $out; } # # XXX porting from php left-over: # # do I *REALLY* need those methods, or should I use # $self->{something} directly? # # Probably direct usage is better! # sub TagName { my $self = shift; return $self->{TagName}; } sub NextMFN { my $self = shift; return $self->{NXTMFN}; } 1; =head1 BUGS This module has been very lightly tested. Use with caution and report bugs. =head1 AUTHOR Dobrica Pavlinusic CPAN ID: DPAVLIN dpavlin@rot13.org http://www.rot13.org/~dpavlin/ This module is based heavily on code from LIBISIS.PHP - Library to read ISIS files V0.1.1 written in php and (c) 2000 Franck Martin - released under LGPL. =head1 COPYRIGHT This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself. The full text of the license can be found in the LICENSE file included with this module. =head1 SEE ALSO L, perl(1).