Mercurial > repos > mcharles > rapsosnp

changeset 9:0e7c6fe60646 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author mcharles
date Mon, 20 Oct 2014 05:58:31 -0400
parents d857538d9fea
children 0a6c1cfe4dc8
files rapsodyn/SplitFileUpTo10x.pl rapsodyn/SplitFileUpTo10x.xml rapsodyn/extractseq.pl
diffstat 3 files changed, 239 insertions(+), 1 deletions(-) [+]
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/rapsodyn/SplitFileUpTo10x.pl	Mon Oct 20 05:58:31 2014 -0400
@@ -0,0 +1,110 @@
+#!/usr/bin/perl
+#V1.0.0
+use strict;
+use warnings;
+
+my $in = $ARGV[0];
+my $NB_LINE_GROUP = $ARGV[1];
+my $NB_FILE = $ARGV[2];
+my $out1 = $ARGV[3];
+my $out2 = $ARGV[4];
+my $out3 = $ARGV[5];
+my $out4 = $ARGV[6];
+my $out5 = $ARGV[7];
+my $out6 = $ARGV[8];
+my $out7 = $ARGV[9];
+my $out8 = $ARGV[10];
+my $out9 = $ARGV[11];
+my $out10 = $ARGV[12];
+
+if (($NB_FILE<2)||($NB_FILE>10)){
+	print STDERR "File must be split in 2-10\n";
+	exit(0);
+}
+my @in;
+my $current_aggregate;
+open(IN, $in) or die ("Can't open $in\n");
+while (my $line =<IN>){
+	$current_aggregate = $line;
+	my $nb_line=1;
+	if ($NB_LINE_GROUP>1){
+		while (my $add_line = <IN>){
+			$current_aggregate .= $add_line;
+			$nb_line++;
+			if ($nb_line == $NB_LINE_GROUP){
+				last;
+			}
+		}
+	}	
+	push (@in,$current_aggregate);
+}
+close (IN);
+
+open (OUT1,">$out1") or die ("Can't open $out1");
+open (OUT2,">$out2") or die ("Can't open $out2");
+open (OUT3,">$out3") or die ("Can't open $out3");
+open (OUT4,">$out4") or die ("Can't open $out4");
+open (OUT5,">$out5") or die ("Can't open $out5");
+open (OUT6,">$out6") or die ("Can't open $out6");
+open (OUT7,">$out7") or die ("Can't open $out7");
+open (OUT8,">$out8") or die ("Can't open $out8");
+open (OUT9,">$out9") or die ("Can't open $out9");
+open (OUT10,">$out10") or die ("Can't open $out10");
+
+
+for (my $i=0;$i<=$#in;$i++){
+	if ($i <= $#in * 1 / $NB_FILE){
+		print OUT1 $in[$i];
+	}
+	elsif ($i <= $#in * 2 / $NB_FILE){
+		print OUT2 $in[$i];
+	}
+	elsif (($NB_FILE>=3)&&($i <= $#in * 3 / $NB_FILE)){
+		print OUT3 $in[$i];
+	}
+	elsif (($NB_FILE>=4)&&($i <= $#in * 4 / $NB_FILE)){
+		print OUT4 $in[$i];
+	}
+	elsif (($NB_FILE>=5)&&($i <= $#in * 5 / $NB_FILE)){
+		print OUT5 $in[$i];
+	}
+	elsif (($NB_FILE>=6)&&($i <= $#in * 6 / $NB_FILE)){
+		print OUT6 $in[$i];
+	}
+	elsif (($NB_FILE>=7)&&($i <= $#in * 7 / $NB_FILE)){
+		print OUT7 $in[$i];
+	}
+	elsif (($NB_FILE>=8)&&($i <= $#in * 8 / $NB_FILE)){
+		print OUT8 $in[$i];
+	}
+	elsif (($NB_FILE>=9)&&($i <= $#in * 9 / $NB_FILE)){
+		print OUT9 $in[$i];
+	}
+	elsif (($NB_FILE>=10)&&($i <= $#in * 10 / $NB_FILE)){
+		print OUT10 $in[$i];
+	}
+	else {
+
+	}
+
+
+}
+
+
+close (OUT1);
+close (OUT2);
+close (OUT3);
+close (OUT4);
+close (OUT5);
+close (OUT6);
+close (OUT7);
+close (OUT8);
+close (OUT9);
+close (OUT10);
+
+
+
+
+
+
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/rapsodyn/SplitFileUpTo10x.xml	Mon Oct 20 05:58:31 2014 -0400
@@ -0,0 +1,128 @@
+<tool id="SplitFileUpTo10x" name="SplitFileUpTo10x" version="1.00">
+<description>Split file in 2-10</description>
+<command interpreter="perl">
+    SplitFileUpTo10x.pl $input_file $line_number $file_number $output_file1 $output_file2 $output_file3 $output_file4 $output_file5 $output_file6 $output_file7 $output_file8 $output_file9 $output_file10
+</command>
+<inputs>
+<param name="out_format" type="select" label="Format">
+	<option value="pileup">pileup</option>
+	<option value="fastq">fastq</option>
+	<option value="fastqsanger">fastqsanger</option>
+	<option value="sam">sam</option>
+	<option value="fasta">fasta</option>
+	<option value="tabular">tabular</option>
+</param>
+<param name="input_file"  type="data" format="txt,tabular,astq,fastqsanger,fastqillumina,pileup,sam,fasta" label="Select a suitable input file from your history"/>
+<param name="line_number" type="integer" value="4" label="Number of line read at a time before splitting"/>
+<param name="file_number" type="integer" value="4" label="Number of split files"/>
+</inputs>
+<outputs>
+<data name="output_file1" format="txt" label="PART1 on ${on_string}">
+	<change_format>
+		<when input="out_format" value="pileup" format="pileup" />
+		<when input="out_format" value="fastq" format="fastq" />
+		<when input="out_format" value="sam" format="sam" />
+		<when input="out_format" value="fasta" format="fasta" />
+		<when input="out_format" value="tabular" format="tabular" />
+		<when input="out_format" value="fastqsanger" format="fastqsanger" />
+	</change_format>
+</data>
+<data name="output_file2" format="txt" label="PART2 on ${on_string}">
+	<change_format>
+		<when input="out_format" value="pileup" format="pileup" />
+		<when input="out_format" value="fastq" format="fastq" />
+		<when input="out_format" value="sam" format="sam" />
+		<when input="out_format" value="fasta" format="fasta" />
+		<when input="out_format" value="tabular" format="tabular" />
+		<when input="out_format" value="fastqsanger" format="fastqsanger" />
+	</change_format>
+</data>
+<data name="output_file3" format="txt" label="PART3 on ${on_string}">
+	<change_format>
+		<when input="out_format" value="pileup" format="pileup" />
+		<when input="out_format" value="fastq" format="fastq" />
+		<when input="out_format" value="sam" format="sam" />
+		<when input="out_format" value="fasta" format="fasta" />
+		<when input="out_format" value="tabular" format="tabular" />
+		<when input="out_format" value="fastqsanger" format="fastqsanger" />
+	</change_format>
+</data>
+<data name="output_file4" format="txt" label="PART4 on ${on_string}">
+	<change_format>
+		<when input="out_format" value="pileup" format="pileup" />
+		<when input="out_format" value="fastq" format="fastq" />
+		<when input="out_format" value="sam" format="sam" />
+		<when input="out_format" value="fasta" format="fasta" />
+		<when input="out_format" value="tabular" format="tabular" />
+		<when input="out_format" value="fastqsanger" format="fastqsanger" />
+	</change_format>
+</data>
+<data name="output_file5" format="txt" label="PART5 on ${on_string}">
+	<change_format>
+		<when input="out_format" value="pileup" format="pileup" />
+		<when input="out_format" value="fastq" format="fastq" />
+		<when input="out_format" value="sam" format="sam" />
+		<when input="out_format" value="fasta" format="fasta" />
+		<when input="out_format" value="tabular" format="tabular" />
+		<when input="out_format" value="fastqsanger" format="fastqsanger" />
+	</change_format>
+</data>
+<data name="output_file6" format="txt" label="PART6 on ${on_string}">
+	<change_format>
+		<when input="out_format" value="pileup" format="pileup" />
+		<when input="out_format" value="fastq" format="fastq" />
+		<when input="out_format" value="sam" format="sam" />
+		<when input="out_format" value="fasta" format="fasta" />
+		<when input="out_format" value="tabular" format="tabular" />
+		<when input="out_format" value="fastqsanger" format="fastqsanger" />
+	</change_format>
+</data>
+<data name="output_file7" format="txt" label="PART7 on ${on_string}">
+	<change_format>
+		<when input="out_format" value="pileup" format="pileup" />
+		<when input="out_format" value="fastq" format="fastq" />
+		<when input="out_format" value="sam" format="sam" />
+		<when input="out_format" value="fasta" format="fasta" />
+		<when input="out_format" value="tabular" format="tabular" />
+		<when input="out_format" value="fastqsanger" format="fastqsanger" />
+	</change_format>
+</data>
+<data name="output_file8" format="txt" label="PART8 on ${on_string}">
+	<change_format>
+		<when input="out_format" value="pileup" format="pileup" />
+		<when input="out_format" value="fastq" format="fastq" />
+		<when input="out_format" value="sam" format="sam" />
+		<when input="out_format" value="fasta" format="fasta" />
+		<when input="out_format" value="tabular" format="tabular" />
+		<when input="out_format" value="fastqsanger" format="fastqsanger" />
+	</change_format>
+</data>
+<data name="output_file9" format="txt" label="PART9 on ${on_string}">
+	<change_format>
+		<when input="out_format" value="pileup" format="pileup" />
+		<when input="out_format" value="fastq" format="fastq" />
+		<when input="out_format" value="sam" format="sam" />
+		<when input="out_format" value="fasta" format="fasta" />
+		<when input="out_format" value="tabular" format="tabular" />
+		<when input="out_format" value="fastqsanger" format="fastqsanger" />
+	</change_format>
+</data>
+<data name="output_file10" format="txt" label="PART10 on ${on_string}">
+	<change_format>
+		<when input="out_format" value="pileup" format="pileup" />
+		<when input="out_format" value="fastq" format="fastq" />
+		<when input="out_format" value="sam" format="sam" />
+		<when input="out_format" value="fasta" format="fasta" />
+		<when input="out_format" value="tabular" format="tabular" />
+		<when input="out_format" value="fastqsanger" format="fastqsanger" />
+	</change_format>
+</data>
+
+</outputs>
+
+<help>
+
+
+
+</help>
+</tool>
--- a/rapsodyn/extractseq.pl	Fri Oct 10 07:05:36 2014 -0400
+++ b/rapsodyn/extractseq.pl	Mon Oct 20 05:58:31 2014 -0400
@@ -89,7 +89,7 @@
 		}
 		$variant{"desc"}=$descriptor;
 		
-		print ">",$variant{"ref"},"_",$descriptor,"\n",$variant{"SEQ"},"\n";
+		print ">",$variant{"ref"},"_",$descriptor,"_","\n",$variant{"SEQ"},"\n";