[Bioperl-guts-l] bioperl-live/Bio/ClusterIO dbsnp.pm,1.17,1.18

[Christopher John Fields]

Update of /home/repository/bioperl/bioperl-live/Bio/ClusterIO
In directory dev.open-bio.org:/tmp/cvs-serv16965

Modified Files:
	dbsnp.pm 
Log Message:
switch to XML::SAX (not heavily tested); this module needs to be heavily refactored per bug 2018

Index: dbsnp.pm
===================================================================
RCS file: /home/repository/bioperl/bioperl-live/Bio/ClusterIO/dbsnp.pm,v
retrieving revision 1.17
retrieving revision 1.18
diff -C2 -d -r1.17 -r1.18
*** dbsnp.pm	26 Sep 2006 22:03:05 -0000	1.17
--- dbsnp.pm	8 Mar 2007 18:40:06 -0000	1.18
***************
*** 55,69 ****
  use Bio::Root::Root;
  use Bio::Variation::SNP;
! use XML::Parser::PerlSAX;
! use XML::Handler::Subs;
  use Data::Dumper;
  use IO::File;
  
- use vars qw($DTD $DEBUG %MODEMAP %MAPPING);
- $DTD = 'ftp://ftp.ncbi.nih.gov/snp/specs/NSE.dtd';
  use base qw(Bio::ClusterIO);
  
! BEGIN {
!   %MAPPING = (
  #the ones commented out i haven't written methods for yet... -Allen
  			  'Rs_rsId'               => 'id',
--- 55,68 ----
  use Bio::Root::Root;
  use Bio::Variation::SNP;
! use XML::SAX;
  use Data::Dumper;
  use IO::File;
+ use Time::HiRes qw(tv_interval gettimeofday);
  
  use base qw(Bio::ClusterIO);
  
! our $DEBUG = 0;
! 
! our %MAPPING = (
  #the ones commented out i haven't written methods for yet... -Allen
  			  'Rs_rsId'               => 'id',
***************
*** 143,147 ****
  			  #there are lots more, but i don't need them at the moment... -Allen
  			  );
- }
  
  sub _initialize{
--- 142,145 ----
***************
*** 150,155 ****
     my ($usetempfile) = $self->_rearrange([qw(TEMPFILE)], at args);
     defined $usetempfile && $self->use_tempfile($usetempfile);
!    $self->{'_xmlparser'} = new XML::Parser::PerlSAX();
!    $DEBUG = 1 if( ! defined $DEBUG && $self->verbose > 0);
  }
  
--- 148,157 ----
     my ($usetempfile) = $self->_rearrange([qw(TEMPFILE)], at args);
     defined $usetempfile && $self->use_tempfile($usetempfile);
!     
!   # start up the parser factory
!   my $parserfactory = XML::SAX::ParserFactory->parser(
! 	  Handler => $self);
!   $self->{'_xmlparser'} = $parserfactory;
!   $DEBUG = 1 if( ! defined $DEBUG && $self->verbose > 0);
  }
  
***************
*** 168,171 ****
--- 170,177 ----
  #Adapted from Jason's blastxml.pm
  ###
+ 
+ # you shouldn't have to preparse this; the XML is well-formed and refers
+ # accurately to a remote DTD/schema
+ 
  sub next_cluster {
    my $self = shift;
***************
*** 181,187 ****
    while( defined( $_ = $self->_readline ) ){
  	#skip to beginning of refSNP entry
! 	if($_ !~ m!<Rs>! && $start){
  	  next;
! 	} elsif($_ =~ m!<Rs>! && $start){
  	  $start = 0;
  	} 
--- 187,193 ----
    while( defined( $_ = $self->_readline ) ){
  	#skip to beginning of refSNP entry
! 	if($_ !~ m{<Rs[^>]*>} && $start){
  	  next;
! 	} elsif($_ =~ m{<Rs[^>]*>} && $start){
  	  $start = 0;
  	} 
***************
*** 195,199 ****
  
  	#and stop at the end of the refSNP entry
! 	last if $_ =~ m!</Rs>!;
    }
  
--- 201,205 ----
  
  	#and stop at the end of the refSNP entry
! 	last if $_ =~ m{</Rs>};
    }
  
***************
*** 283,298 ****
    my ($self,$data) = @_;
    my $nm = $data->{'Name'};
!   my $at = $data->{'Attributes'};
! 
    if($nm eq 'Ss'){
  	$self->refsnp->add_subsnp;
  	return;
    }
    if(my $type = $MAPPING{$nm}){
  	if(ref $type eq 'HASH'){
  	  #okay, this is nasty.  what can you do?
  	  $self->{will_handle}   = (keys %$type)[0];
! 	  my $valkey             = (values %$type)[0];
! 	  $self->{last_data}     = $at->{$valkey};
  	} else {
  	  $self->{will_handle} = $type;
--- 289,306 ----
    my ($self,$data) = @_;
    my $nm = $data->{'Name'};
!   my $at = $data->{'Attributes'}->{'{}value'};
!   
!   #$self->debug(Dumper($at)) if $nm = ;
!   
    if($nm eq 'Ss'){
  	$self->refsnp->add_subsnp;
  	return;
    }
+   
    if(my $type = $MAPPING{$nm}){
  	if(ref $type eq 'HASH'){
  	  #okay, this is nasty.  what can you do?
  	  $self->{will_handle}   = (keys %$type)[0];
! 	  $self->{last_data}     = $at->{Value};
  	} else {
  	  $self->{will_handle} = $type;