annotate tools/fastq/fastq_filter.py @ 1:cdcb0ce84a1b

Uploaded
author xuebing
date Fri, 09 Mar 2012 19:45:15 -0500
parents 9071e359b9a3
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
1 #Dan Blankenberg
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
2 import sys, os, shutil
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
3 from galaxy_utils.sequence.fastq import fastqReader, fastqWriter
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
4
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
5 def main():
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
6 #Read command line arguments
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
7 input_filename = sys.argv[1]
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
8 script_filename = sys.argv[2]
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
9 output_filename = sys.argv[3]
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
10 additional_files_path = sys.argv[4]
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
11 input_type = sys.argv[5] or 'sanger'
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
12
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
13 #Save script file for debuging/verification info later
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
14 os.mkdir( additional_files_path )
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
15 shutil.copy( script_filename, os.path.join( additional_files_path, 'debug.txt' ) )
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
16
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
17 out = fastqWriter( open( output_filename, 'wb' ), format = input_type )
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
18
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
19 i = None
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
20 reads_kept = 0
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
21 for i, fastq_read in enumerate( fastqReader( open( input_filename ), format = input_type ) ):
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
22 local = {'fastq_read':fastq_read, 'ret_val':False}
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
23 execfile( script_filename, {}, local )
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
24 if local['ret_val']:
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
25 out.write( fastq_read )
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
26 reads_kept += 1
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
27 out.close()
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
28 if i is None:
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
29 print "Your file contains no valid fastq reads."
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
30 else:
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
31 print 'Kept %s of %s reads (%.2f%%).' % ( reads_kept, i + 1, float( reads_kept ) / float( i + 1 ) * 100.0 )
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
32
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
33 if __name__ == "__main__":
9071e359b9a3 Uploaded
xuebing
parents:
diff changeset
34 main()