Mercurial > repos > sanbi-uwc > novo_align

annotate novo_align.py @ 27:df170f4d2eea draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/SANBI-SA/tools-sanbi-uwc/tree/master/tools/novo_align commit 018e4e2f405e9e55f3b7c3ba92b29f957e468f2b
author sanbi-uwc
date Mon, 07 May 2018 09:42:08 -0400
parents 482aef16b6f5
children 395aac51afa6
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
22
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
1 #!/usr/bin/env python
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
2 from __future__ import print_function
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
3 import argparse
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
4 import shlex
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
5 import os
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
6 import sys
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
7 import logging
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
8 log = logging.getLogger( __name__ )
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
9
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
10
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
11 def novo_align(output_filename, index_filename, fwd_file, rev_file ):
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
12 param = r'@RG\tID:RG\tSM:$i\tPL:ILLUMINA'
27
df170f4d2eea planemo upload for repository https://github.com/SANBI-SA/tools-sanbi-uwc/tree/master/tools/novo_align commit 018e4e2f405e9e55f3b7c3ba92b29f957e468f2b
sanbi-uwc
parents: 22
diff changeset
13 cmdline_str = "novoalign -c 8 -k -d {} -f {} {} -i PE 250, 100 -o SAM '{}' | samtools sort - | samtools view -b - > {}".format(
22
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
14 index_filename,
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
15 fwd_file,
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
16 rev_file,
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
17 param,
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
18 output_filename)
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
19 try:
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
20 os.system(cmdline_str)
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
21 except:
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
22 print("Error running the nova-align", file=sys.stderr)
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
23
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
24
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
25 def newSplit(value):
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
26 lex = shlex.shlex(value)
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
27 lex.quotes = '"'
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
28 lex.whitespace_split = True
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
29 lex.commenters = ''
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
30 return list(lex)
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
31
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
32
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
33 def main():
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
34 parser = argparse.ArgumentParser(description="Generate a BAM file from the Novo Align tool")
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
35 parser.add_argument('output_filename')
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
36 parser.add_argument('--index_filename')
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
37 parser.add_argument('--forward_filename')
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
38 parser.add_argument('--reverse_filename')
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
39 args = parser.parse_args()
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
40
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
41 # a dirty way of referencing the file
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
42 index_file_path = args.index_filename # removed old logic here: + "/" + args.index_filename.split("/")[-1]
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
43 novo_align(args.output_filename, index_file_path, args.forward_filename, args.reverse_filename)
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
44
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
45
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
46 if __name__ == "__main__":
482aef16b6f5 planemo upload for repository https://github.com/zipho/novo_align commit 3aa343058f631ac4117c9bc86989dc449f84135f
sanbi-uwc
parents:
diff changeset
47 main()