Bio::SearchIO blastxml
SummaryIncluded librariesPackage variablesSynopsisDescriptionGeneral documentationMethods
Toolbar
WebCvs
Summary
Bio::SearchIO::blastxml - A SearchIO implementation of NCBI Blast XML parsing.
Package variables
Privates (from "my" definitions)
%VALID_TYPE = ( 'BLAST' => 'Bio::SearchIO::XML::BlastHandler', 'PSIBLAST' => 'Bio::SearchIO::XML::PsiBlastHandler', 'PSI-BLAST' => 'Bio::SearchIO::XML::PsiBlastHandler' )
Included modules
Bio::Root::Root
Bio::SearchIO::IteratedSearchResultEventBuilder
Bio::SearchIO::XML::BlastHandler
IO::File
XML::SAX
Inherit
Bio::SearchIO
Synopsis
    use Bio::SearchIO;
    my $searchin = Bio::SearchIO->new(-format => 'blastxml',
				     -file   => 't/data/plague_yeast.bls.xml');
                     
    while( my $result = $searchin->next_result ) {
        ....
    }

    # one can also request that the parser NOT keep the XML data in memory
    # by using the tempfile initialization flag.
    
    $searchin = Bio::SearchIO->new(-tempfile => 1,
				     -format => 'blastxml',
				     -file   => 't/data/plague_yeast.bls.xml');
                     
    while( my $result = $searchin->next_result ) {
       ....
    }
    
    # PSI-BLAST parsing (default is normal BLAST)
    $searchin = Bio::SearchIO->new(
                     -format => 'blastxml',
                     -blasttype => 'psiblast',
				     -file   => 't/data/plague_yeast.bls.xml');
                     
    while( my $result = $searchin->next_result ) {
       ....
    }
Description
This object implements a NCBI Blast XML parser. It requires XML::SAX; it is
also recommended (for faster parsing) that XML::SAX::ExpatXS or XML::LibXML
be installed. Either 'XML::SAX::ExpatXS' or 'XML::LibXML::SAX::Parser' should
be set as the default parser in ParserDetails.ini. This file is located in the
SAX subdirectory of XML in your local perl library (normally in the 'site'
directory).
Two different XML handlers currently exist to deal with logical differences
between how normal BLAST reports and PSI-BLAST reports are logically parsed into
BioPerl objects; this is explicitly settable using the -blasttype parameter.
The default is for parsing a normal BLAST report ('blast'), but if one is
expecting PSI-BLAST report parsing, -blasttype must be set explicitly to
'psiblast'. This is due to a lack of any information in the XML output which
tells the parser the report is derived from a PSI-BLAST run vs. a normal BLAST
run.
There is one additional initialization flag from the SearchIO defaults. That is
the -tempfile flag. If specified as true, then the parser will write out each
report to a temporary filehandle rather than holding the entire report as a
string in memory. The reason this is done in the first place is NCBI reports
have an uncessary <?xml version="1.0"?> at the beginning of each report
and RPS-BLAST reports have an additional unnecessary RPS-BLAST tag at the top of
each report. So we currently have implemented the work around by preparsing the
file (yes it makes the process slower, but it works). We are open to suggestions
on how to optimize this in the future.
Methods
_initializeDescriptionCode
next_resultDescriptionCode
result_countDescriptionCode
use_tempfileDescriptionCode
blasttypeDescriptionCode
saxparser
No description
Code
_chunk_normalblast
No description
Code
_chunk_psiblast
No description
Code
Methods description
_initializecode    nextTop
 Title   : _initialize
 Usage   : private
 Function: Initializes the object - this is chained through new in SearchIO
next_resultcodeprevnextTop
 Title   : next_result
 Usage   : my $hit = $searchio->next_result;
 Function: Returns the next Result from a search
 Returns : Bio::Search::Result::ResultI object
 Args    : none
result_countcodeprevnextTop
 Title   : result_count
 Usage   : $num = $stream->result_count;
 Function: Gets the number of Blast results that have been successfully parsed
           at the point of the method call.  This is not the total # of results
           in the file.
 Returns : integer
 Args    : none
 Throws  : none
use_tempfilecodeprevnextTop
 Title   : use_tempfile
 Usage   : $obj->use_tempfile($newval)
 Function: Get/Set boolean flag on whether or not use a tempfile
 Example : 
 Returns : value of use_tempfile
 Args    : newvalue (optional)
blasttypecodeprevnextTop
 Title   : blasttype
 Usage   : $obj->blasttype($newtype)
 Function: Get/Set BLAST report type.
 Returns : BLAST report type
 Args    : case-insensitive string of types BLAST or PSIBLAST (default: BLAST)
 Note    : this is used to determine how reports are 'chunked' (in cases
           where multiple queries are submitted) and which XML handler
           to use when parsing the report(s)
Methods code
_initializedescriptionprevnextTop
sub _initialize {
    my ($self,@args) = @_;   
    $self->SUPER::_initialize(@args);
    my ($usetempfile, $blasttype,$xmlcompact) = $self->_rearrange([qw(
                                            TEMPFILE
                                            BLASTTYPE
                                            XMLCOMPACT)],@args);
    $blasttype ||= 'BLAST';
    $self->{_xml_compact} = $xmlcompact || 0;
    $self->blasttype(uc $blasttype);
    defined $usetempfile && $self->use_tempfile($usetempfile);
    $self->{_result_count} = 0;
    eval {  require Time::HiRes };	
    if( $@ ) { $DEBUG = 0; }
    $DEBUG = 1 if( ! defined $DEBUG && ($self->verbose > 0));
}
next_resultdescriptionprevnextTop
sub next_result {
    my ($self) = @_;
    
    my $result;

    my ($tfh);
    
    # XMLCOMPACT
    # WU-BLAST has an XML_COMPACT option which needs to be preprocessed before
    # passing on to the parser.  
    if ($self->{_xml_compact}) {
        $self->debug("XMLCOMPACT mode\n");
        my ($tfh2, $filename) = IO::File->new_tmpfile or $self->throw("Unable to open temp file: $!");
        $tfh2->autoflush(1);
        my $fh = $self->_fh;
        while (my $line = <$fh>) {
            $line =~ s/></>\n</g;
            print $tfh2 $line;
        }
        seek($tfh2,0,0);
        close $fh;
        # redirect self's IO to use new tempfile
        $self->_fh($tfh2);
    }
    
    if( $self->use_tempfile ) {
        $tfh = IO::File->new_tmpfile or $self->throw("Unable to open temp file: $!");	
        $tfh->autoflush(1);
    }
    
    my $okaytoprocess = ($self->blasttype =~ /PSI/) ? $self->_chunk_psiblast($tfh) :
        $self->_chunk_normalblast($tfh);
    
    return unless( $okaytoprocess);
    
    my %parser_args;
    if( defined $tfh ) {
	seek($tfh,0,0);
	%parser_args = ('Source' => { 'ByteStream' => $tfh });
    } else {
	%parser_args = ('Source' => { 'String' => $self->{'_blastdata'} });
    }

    my $starttime;
    if(  $DEBUG ) {  $starttime = [ Time::HiRes::gettimeofday() ]; }
    
    eval {
	$result = $self->{'_xmlparser'}->parse(%parser_args);
    };
    
    if( $@ ) {
	$self->warn("error in parsing a report:\n $@");
	$result = undef;
    }
    if( $DEBUG ) {
	$self->debug( sprintf("parsing took %f seconds\n", Time::HiRes::tv_interval($starttime)));
    }
    # parsing magic here - but we call event handlers rather than 
    # instantiating things 
    if (defined $result) {
        # result count is handled here, as the BLASTXML reports are
        # broken up into smaller easier to digest bits
        $self->{_result_count}++;
        return $result;
    } else {
        return;
    }
}
result_countdescriptionprevnextTop
sub result_count {
    my $self = shift;
    return $self->{_result_count};
}
use_tempfiledescriptionprevnextTop
sub use_tempfile {
   my ($self,$value) = @_;
   if( defined $value) {
      $self->{'_use_tempfile'} = $value;
    }
    return $self->{'_use_tempfile'};
}
blasttypedescriptionprevnextTop
sub blasttype {
    my ($self,$value) = @_;
    if ($value) {
        $self->throw("$value is not a supported BLAST type") unless exists $VALID_TYPE{$value};
        my $ok;
        eval {
            $ok = $self->_load_module($VALID_TYPE{$value});
        };
        if ($@) {
            print STDERR <<END;
$self: data module $VALID_TYPE{$value} cannot be found
Exception $@
For more information about the Bio::SearchIO::blastxml system please see the Bio::SearchIO::blastxml. 
END
            return unless $ok;
        }
        # BlastHandler does the heavy lifting
        my $xmlhandler = $VALID_TYPE{$value}->new(-verbose => $self->verbose);
        
        # The XML handler does the heavy work, passes data to object handler
        if ($value =~ /^PSI/) {
            my $handler = Bio::SearchIO::IteratedSearchResultEventBuilder->new();
            $self->{'_handler'} = $handler; # cache
        }
        $xmlhandler->eventHandler($self->_eventHandler());

        # start up the parser factory
        my $parserfactory = XML::SAX::ParserFactory->parser(
            Handler => $xmlhandler);
        $self->{'_xmlparser'} = $parserfactory;
        $self->saxparser(ref($parserfactory));
        
        $self->{'_blasttype'} = $value;
    }
    return $self->{'_blasttype'};
}
saxparserdescriptionprevnextTop
sub saxparser {
    my $self = shift;
    return ref($self->{'_xmlparser'});
}
_chunk_normalblastdescriptionprevnextTop
sub _chunk_normalblast {
    my ($self, $tfh) = @_;
    
    local $/ = "\n";
    local $_;
    $self->{'_blastdata'} = '';
    
    my ($sawxmlheader, $okaytoprocess);
    
    my $mode = 'header';

    my $tail = << 'XML_END';
  </BlastOutput_iterations>
</BlastOutput>
XML_END

    # no buffering needed (famous last words...)
    my $fh = $self->_fh;
    
    #chop up XML into edible bits for the parser
    while( defined( my $line = <$fh>) ) {
        next if $line =~ m{^\s*</BlastOutput_iterations>}xmso || $line =~ m{^</BlastOutput>}xmso;
        if( $line =~ m{^RPS-BLAST}i ) {
            $self->{'_type'} = 'RPS-BLAST';
            next;
        } elsif ($line =~ m{^<\?xml\sversion="1.0"}xms) {# <?xml version="1.0"?> & <?xml version="1.0" encoding="UTF-8"?>
            delete $self->{'_header'} if exists $self->{'_header'};
            $sawxmlheader++;
            $mode = 'header';
        } elsif ($line =~ m{^\s*<Iteration>}xmso) {
            if (!$sawxmlheader) {
                if (defined $tfh) {
                    print $tfh $self->{'_header'}
                } else {
                    $self->{'_blastdata'} .= $self->{'_header'};
                }
            }
            $mode = 'iteration';
        } elsif ($line =~ m{^\s*</Iteration>}xmso) {
            if (defined $tfh) {
                print $tfh $line.$tail;
            } else {
                $self->{'_blastdata'} .= $line.$tail;
            }
            $okaytoprocess++;
            last;
        }
        if (defined $tfh) {
            print $tfh $line;
        } else {
            $self->{'_blastdata'} .= $line;
        }
        $self->{"_$mode"} .= $line if $mode eq 'header';
    }
    return $okaytoprocess;
}
_chunk_psiblastdescriptionprevnextTop
sub _chunk_psiblast {
    my ($self, $tfh) = @_;
    
    local $/ = "\n";
    local $_;
    $self->{'_blastdata'} = '';
    
    my ($sawxmlheader, $okaytoprocess);

    # no buffering needed (famous last words...)
    my $fh = $self->_fh;
    
    #chop up XML into edible bits for the parser
    while( defined( my $line = <$fh>) ) {
        if (defined $tfh) {
            print $tfh $line;
        } else {
            $self->{'_blastdata'} .= $line;
        }
        #$self->{"_$mode"} .= $line;
        if ($line =~ m{^</BlastOutput>}xmso) {
            $okaytoprocess++;
            last;
        }
    }
    #$self->debug($self->{'_blastdata'}."\n");
    return $okaytoprocess;
}

1;
}
General documentation
DEPENDENCIESTop
In addition to parts of the Bio:: hierarchy, this module uses:
 XML::SAX
It is also recommended that XML::SAX::ExpatXS be installed and made the default
XML::SAX parser using , along with the Expat library () for faster parsing.
XML::SAX::Expat is not recommended; XML::SAX::ExpatXS is considered the current
replacement for XML::SAX:Expat and is actively being considered to replace
XML::SAX::Expat. XML::SAX::Expat will work, but only if you have local copies of
the NCBI BLAST DTDs. This is due to issues with NCBI's BLAST XML format. The
DTDs and the web address to obtain them are:
  NCBI_BlastOutput.dtd	    
  NCBI_BlastOutput.mod.dtd

  http://www.ncbi.nlm.nih.gov/data_specs/dtd/
FEEDBACKTop
Mailing ListsTop
User feedback is an integral part of the evolution of this and other
Bioperl modules. Send your comments and suggestions preferably to
the Bioperl mailing list. Your participation is much appreciated.
  bioperl-l@bioperl.org                  - General discussion
  http://bioperl.org/wiki/Mailing_lists  - About the mailing lists
Support Top
Please direct usage questions or support issues to the mailing list:
bioperl-l@bioperl.org
rather than to the module maintainer directly. Many experienced and
reponsive experts will be able look at the problem and quickly
address it. Please include a thorough description of the problem
with code and data examples if at all possible.
Reporting BugsTop
Report bugs to the Bioperl bug tracking system to help us keep track
of the bugs and their resolution. Bug reports can be submitted via the
web:
  https://redmine.open-bio.org/projects/bioperl/
AUTHOR - Jason StajichTop
Email jason-at-bioperl.org
APPENDIXTop
The rest of the documentation details each of the object methods.
Internal methods are usually preceded with a _
newTop
 Title   : new
 Usage   : my $searchio = Bio::SearchIO->new(-format => 'blastxml',
					    -file   => 'filename',
					    -tempfile => 1);
 Function: Initializes the object - this is chained through new in SearchIO
 Returns : Bio::SearchIO::blastxml object
 Args    : One additional argument from the format and file/fh parameters.
           -tempfile    => boolean.  Defaults to false.  Write out XML data
                           to a temporary filehandle to send to PerlSAX parser.