[Bioperl-guts-l] [14994] bioperl-live/trunk/Bio/SearchIO/fasta.pm: perltidy
Christopher John Fields
cjfields at dev.open-bio.org
Sun Nov 16 01:18:02 EST 2008
Revision: 14994
Author: cjfields
Date: 2008-11-16 01:18:02 -0500 (Sun, 16 Nov 2008)
Log Message:
-----------
perltidy
Modified Paths:
--------------
bioperl-live/trunk/Bio/SearchIO/fasta.pm
Modified: bioperl-live/trunk/Bio/SearchIO/fasta.pm
===================================================================
--- bioperl-live/trunk/Bio/SearchIO/fasta.pm 2008-11-16 06:14:41 UTC (rev 14993)
+++ bioperl-live/trunk/Bio/SearchIO/fasta.pm 2008-11-16 06:18:02 UTC (rev 14994)
@@ -19,9 +19,9 @@
# Do not use this object directly, use it through the SearchIO system
use Bio::SearchIO;
my $searchio = Bio::SearchIO->new(-format => 'fasta',
- -file => 'report.FASTA');
+ -file => 'report.FASTA');
while( my $result = $searchio->next_result ) {
- # ... do what you would normally doi with Bio::SearchIO.
+ # ... do what you would normally doi with Bio::SearchIO.
}
=head1 DESCRIPTION
@@ -66,10 +66,8 @@
=cut
-
# Let the code begin...
-
package Bio::SearchIO::fasta;
use vars qw(%MODEMAP %MAPPING $IDLENGTH);
use strict;
@@ -78,7 +76,8 @@
use Bio::Factory::ObjectFactory;
-BEGIN {
+BEGIN {
+
# Set IDLENGTH to a new value if you have
# compile FASTA with a different ID length
# (actually newest FASTA allows the setting of this
@@ -86,73 +85,72 @@
$IDLENGTH = 6;
# mapping of NCBI Blast terms to Bioperl hash keys
- %MODEMAP = ('FastaOutput' => 'result',
- 'Hit' => 'hit',
- 'Hsp' => 'hsp'
- );
+ %MODEMAP = (
+ 'FastaOutput' => 'result',
+ 'Hit' => 'hit',
+ 'Hsp' => 'hsp'
+ );
# This should really be done more intelligently, like with
# XSLT
- %MAPPING =
- (
- 'Hsp_bit-score' => 'HSP-bits',
- 'Hsp_score' => 'HSP-score',
- 'Hsp_sw-score' => 'HSP-swscore',
- 'Hsp_evalue' => 'HSP-evalue',
- 'Hsp_query-from'=> 'HSP-query_start',
- 'Hsp_query-to' => 'HSP-query_end',
- 'Hsp_hit-from' => 'HSP-hit_start',
- 'Hsp_hit-to' => 'HSP-hit_end',
- 'Hsp_positive' => 'HSP-conserved',
- 'Hsp_identity' => 'HSP-identical',
- 'Hsp_gaps' => 'HSP-hsp_gaps',
- 'Hsp_hitgaps' => 'HSP-hit_gaps',
- 'Hsp_querygaps' => 'HSP-query_gaps',
- 'Hsp_qseq' => 'HSP-query_seq',
- 'Hsp_hseq' => 'HSP-hit_seq',
- 'Hsp_midline' => 'HSP-homology_seq',
- 'Hsp_align-len' => 'HSP-hsp_length',
- 'Hsp_query-frame'=> 'HSP-query_frame',
- 'Hsp_hit-frame' => 'HSP-hit_frame',
+ %MAPPING = (
+ 'Hsp_bit-score' => 'HSP-bits',
+ 'Hsp_score' => 'HSP-score',
+ 'Hsp_sw-score' => 'HSP-swscore',
+ 'Hsp_evalue' => 'HSP-evalue',
+ 'Hsp_query-from' => 'HSP-query_start',
+ 'Hsp_query-to' => 'HSP-query_end',
+ 'Hsp_hit-from' => 'HSP-hit_start',
+ 'Hsp_hit-to' => 'HSP-hit_end',
+ 'Hsp_positive' => 'HSP-conserved',
+ 'Hsp_identity' => 'HSP-identical',
+ 'Hsp_gaps' => 'HSP-hsp_gaps',
+ 'Hsp_hitgaps' => 'HSP-hit_gaps',
+ 'Hsp_querygaps' => 'HSP-query_gaps',
+ 'Hsp_qseq' => 'HSP-query_seq',
+ 'Hsp_hseq' => 'HSP-hit_seq',
+ 'Hsp_midline' => 'HSP-homology_seq',
+ 'Hsp_align-len' => 'HSP-hsp_length',
+ 'Hsp_query-frame' => 'HSP-query_frame',
+ 'Hsp_hit-frame' => 'HSP-hit_frame',
- 'Hit_id' => 'HIT-name',
- 'Hit_len' => 'HIT-length',
- 'Hit_accession' => 'HIT-accession',
- 'Hit_def' => 'HIT-description',
- 'Hit_signif' => 'HIT-significance',
- 'Hit_score' => 'HIT-score',
+ 'Hit_id' => 'HIT-name',
+ 'Hit_len' => 'HIT-length',
+ 'Hit_accession' => 'HIT-accession',
+ 'Hit_def' => 'HIT-description',
+ 'Hit_signif' => 'HIT-significance',
+ 'Hit_score' => 'HIT-score',
- 'FastaOutput_program' => 'RESULT-algorithm_name',
- 'FastaOutput_version' => 'RESULT-algorithm_version',
- 'FastaOutput_query-def'=> 'RESULT-query_name',
- 'FastaOutput_querydesc'=> 'RESULT-query_description',
- 'FastaOutput_query-len'=> 'RESULT-query_length',
- 'FastaOutput_db' => 'RESULT-database_name',
- 'FastaOutput_db-len' => 'RESULT-database_entries',
- 'FastaOutput_db-let' => 'RESULT-database_letters',
+ 'FastaOutput_program' => 'RESULT-algorithm_name',
+ 'FastaOutput_version' => 'RESULT-algorithm_version',
+ 'FastaOutput_query-def' => 'RESULT-query_name',
+ 'FastaOutput_querydesc' => 'RESULT-query_description',
+ 'FastaOutput_query-len' => 'RESULT-query_length',
+ 'FastaOutput_db' => 'RESULT-database_name',
+ 'FastaOutput_db-len' => 'RESULT-database_entries',
+ 'FastaOutput_db-let' => 'RESULT-database_letters',
- 'Parameters_matrix' => { 'RESULT-parameters' => 'matrix'},
- 'Parameters_expect' => { 'RESULT-parameters' => 'expect'},
- 'Parameters_include' => { 'RESULT-parameters' => 'include'},
- 'Parameters_sc-match' => { 'RESULT-parameters' => 'match'},
- 'Parameters_sc-mismatch' => { 'RESULT-parameters' => 'mismatch'},
- 'Parameters_gap-open' => { 'RESULT-parameters' => 'gapopen'},
- 'Parameters_gap-ext' => { 'RESULT-parameters' => 'gapext'},
- 'Parameters_word-size' => { 'RESULT-parameters' => 'wordsize'},
- 'Parameters_ktup' => { 'RESULT-parameters' => 'ktup'},
- 'Parameters_filter' => {'RESULT-parameters' => 'filter'},
- 'Statistics_db-num' => { 'RESULT-statistics' => 'dbentries'},
- 'Statistics_db-len' => { 'RESULT-statistics' => 'dbletters'},
- 'Statistics_hsp-len' => { 'RESULT-statistics' => 'hsplength'},
- 'Statistics_eff-space' => { 'RESULT-statistics' => 'effectivespace'},
- 'Statistics_kappa' => { 'RESULT-statistics' => 'kappa' },
- 'Statistics_lambda' => { 'RESULT-statistics' => 'lambda' },
- 'Statistics_entropy' => { 'RESULT-statistics' => 'entropy'},
- );
+ 'Parameters_matrix' => { 'RESULT-parameters' => 'matrix' },
+ 'Parameters_expect' => { 'RESULT-parameters' => 'expect' },
+ 'Parameters_include' => { 'RESULT-parameters' => 'include' },
+ 'Parameters_sc-match' => { 'RESULT-parameters' => 'match' },
+ 'Parameters_sc-mismatch' => { 'RESULT-parameters' => 'mismatch' },
+ 'Parameters_gap-open' => { 'RESULT-parameters' => 'gapopen' },
+ 'Parameters_gap-ext' => { 'RESULT-parameters' => 'gapext' },
+ 'Parameters_word-size' => { 'RESULT-parameters' => 'wordsize' },
+ 'Parameters_ktup' => { 'RESULT-parameters' => 'ktup' },
+ 'Parameters_filter' => { 'RESULT-parameters' => 'filter' },
+ 'Statistics_db-num' => { 'RESULT-statistics' => 'dbentries' },
+ 'Statistics_db-len' => { 'RESULT-statistics' => 'dbletters' },
+ 'Statistics_hsp-len' => { 'RESULT-statistics' => 'hsplength' },
+ 'Statistics_eff-space' => { 'RESULT-statistics' => 'effectivespace' },
+ 'Statistics_kappa' => { 'RESULT-statistics' => 'kappa' },
+ 'Statistics_lambda' => { 'RESULT-statistics' => 'lambda' },
+ 'Statistics_entropy' => { 'RESULT-statistics' => 'entropy' },
+ );
}
-
use base qw(Bio::SearchIO);
=head2 new
@@ -170,16 +168,19 @@
=cut
sub _initialize {
- my($self, at args) = @_;
- $self->SUPER::_initialize(@args);
- return unless @args;
- my ($idlength) = $self->_rearrange([qw(IDLENGTH)], at args);
- $self->idlength($idlength || $IDLENGTH);
- $self->_eventHandler->register_factory('hsp',
- Bio::Factory::ObjectFactory->new(
- -type => 'Bio::Search::HSP::FastaHSP',
- -interface => 'Bio::Search::HSP::HSPI'));
- return 1;
+ my ( $self, @args ) = @_;
+ $self->SUPER::_initialize(@args);
+ return unless @args;
+ my ($idlength) = $self->_rearrange( [qw(IDLENGTH)], @args );
+ $self->idlength( $idlength || $IDLENGTH );
+ $self->_eventHandler->register_factory(
+ 'hsp',
+ Bio::Factory::ObjectFactory->new(
+ -type => 'Bio::Search::HSP::FastaHSP',
+ -interface => 'Bio::Search::HSP::HSPI'
+ )
+ );
+ return 1;
}
=head2 next_result
@@ -192,656 +193,1195 @@
=cut
-sub next_result{
- my ($self) = @_;
- local $/ = "\n";
- local $_;
+sub next_result {
+ my ($self) = @_;
+ local $/ = "\n";
+ local $_;
- my $data = '';
- my $seentop = 0;
- my $current_hsp;
- $self->start_document();
- my @hit_signifs;
- while( defined ($_ = $self->_readline )) {
- next if( ! $self->in_element('hsp') &&
- /^\s+$/); # skip empty lines
- if( m/(\S+)\s+searches\s+a\s+(protein\s+or\s+DNA\s+)?sequence/oxi ||
- /(\S+)\s+compares\s+a/ ||
- ( m/^\#\s+/ &&
- ($_ = $self->_readline) &&
- /(\S+)\s+searches\s+a\s+(protein\s+or\s+DNA\s+)?sequence/oxi ||
- /(\S+)\s+compares\s+a/
- )
- ) {
- if( $seentop ) {
- $self->_pushback($_);
- $self->end_element({ 'Name' => 'FastaOutput'});
- return $self->end_document();
- }
- $self->{'_reporttype'} = $1;
- $self->start_element({ 'Name' => 'FastaOutput' } );
- $self->{'_result_count'}++;
- $seentop = 1;
- $self->debug("reporttype is ".$self->{'_reporttype'}."\n") if $self->verbose > 0;
- $self->element({ 'Name' => 'FastaOutput_program',
- 'Data' => $self->{'_reporttype'}});
- $_ = $self->_readline();
- my ($version) = (/version\s+(\S+)/);
- $version = '' unless defined $version;
- $self->{'_version'} = $version;
- $self->element({ 'Name' => 'FastaOutput_version',
- 'Data' => $version});
+ my $data = '';
+ my $seentop = 0;
+ my $current_hsp;
+ $self->start_document();
+ my @hit_signifs;
+ while ( defined( $_ = $self->_readline ) ) {
+ next if ( !$self->in_element('hsp')
+ && /^\s+$/ ); # skip empty lines
+ if (
+ m/(\S+)\s+searches\s+a\s+(protein\s+or\s+DNA\s+)?sequence/oxi
+ || /(\S+)\s+compares\s+a/
+ || ( m/^\#\s+/
+ && ( $_ = $self->_readline )
+ && /(\S+)\s+searches\s+a\s+(protein\s+or\s+DNA\s+)?sequence/oxi
+ || /(\S+)\s+compares\s+a/ )
+ )
+ {
+ if ($seentop) {
@@ Diff output truncated at 10000 characters. @@
More information about the Bioperl-guts-l
mailing list