diff fastq_paired_end_joiner.py @ 0:2793d1d765b9 draft

Imported from capsule None
author devteam
date Mon, 27 Jan 2014 09:25:44 -0500
parents
children 270a8ed8a300
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/fastq_paired_end_joiner.py	Mon Jan 27 09:25:44 2014 -0500
@@ -0,0 +1,38 @@
+#Dan Blankenberg
+import sys, os, shutil
+from galaxy_utils.sequence.fastq import fastqReader, fastqNamedReader, fastqWriter, fastqJoiner
+
+def main():
+    #Read command line arguments
+    input1_filename = sys.argv[1]
+    input1_type = sys.argv[2] or 'sanger'
+    input2_filename = sys.argv[3]
+    input2_type = sys.argv[4] or 'sanger'
+    output_filename = sys.argv[5]
+    
+    if input1_type != input2_type:
+        print "WARNING: You are trying to join files of two different types: %s and %s." % ( input1_type, input2_type )
+    
+    input2 = fastqNamedReader( open( input2_filename, 'rb' ), input2_type )
+    joiner = fastqJoiner( input1_type )
+    out = fastqWriter( open( output_filename, 'wb' ), format = input1_type )
+    
+    i = None
+    skip_count = 0
+    for i, fastq_read in enumerate( fastqReader( open( input1_filename, 'rb' ), format = input1_type ) ):
+        identifier = joiner.get_paired_identifier( fastq_read )
+        fastq_paired = input2.get( identifier )
+        if fastq_paired is None:
+            skip_count += 1
+        else:
+            out.write( joiner.join( fastq_read, fastq_paired ) )
+    out.close()
+    
+    if i is None:
+        print "Your file contains no valid FASTQ reads."
+    else:
+        print input2.has_data()
+        print 'Joined %s of %s read pairs (%.2f%%).' % ( i - skip_count + 1, i + 1, float( i - skip_count + 1 ) / float( i + 1 ) * 100.0 )
+
+if __name__ == "__main__":
+    main()