Mercurial > repos > sanbi-uwc > data_manager_rnastar_index_builder

comparison data_manager/rnastar_index_builder.py @ 12:a42dbab7917c draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/pvanheus/data_manager_rnastar_index_builder commit 599e0c3b37869cc016156f7058e16d6c268bc703
author sanbi-uwc
date Mon, 15 Feb 2016 06:55:55 -0500
parents 6096741e8d83
children e905fcc6fd15
comparison
equal deleted inserted replaced
11:6096741e8d83 12:a42dbab7917c
7 from os import environ, mkdir, makedirs 7 from os import environ, mkdir, makedirs
8 from os.path import isdir, exists 8 from os.path import isdir, exists
9 import shlex 9 import shlex
10 import sys 10 import sys
11 11
12 def get_id_name( params, dbkey, fasta_description=None): 12
13 #TODO: ensure sequence_id is unique and does not already appear in location file 13 def get_id_name(params, dbkey, fasta_description=None):
14 # TODO: ensure sequence_id is unique and does not already appear in location file
14 sequence_id = params['param_dict']['sequence_id'] 15 sequence_id = params['param_dict']['sequence_id']
15 if not sequence_id: 16 if not sequence_id:
16 sequence_id = dbkey 17 sequence_id = dbkey
17 18
18 sequence_name = params['param_dict']['sequence_name'] 19 sequence_name = params['param_dict']['sequence_name']
20 sequence_name = fasta_description 21 sequence_name = fasta_description
21 if not sequence_name: 22 if not sequence_name:
22 sequence_name = dbkey 23 sequence_name = dbkey
23 return sequence_id, sequence_name 24 return sequence_id, sequence_name
24 25
26
25 def make_rnastar_index(output_directory, fasta_filename): 27 def make_rnastar_index(output_directory, fasta_filename):
26 #STAR
27 # --runMode genomeGenerate
28 # --genomeDir tempstargenomedir
29 # --genomeFastaFiles $input1
30 # --runThreadsN \${GALAXY_SLOTS:-1}
31 # --genomeChrBinNbits $advanced_options.chr_bin_nbits
32
33 if exists(output_directory) and not isdir(output_directory): 28 if exists(output_directory) and not isdir(output_directory):
34 print("Output directory path already exists but is not a directory: {}".format(output_directory), file=sys.stderr) 29 print("Output directory path already exists but is not a directory: {}".format(output_directory),
30 file=sys.stderr)
35 elif not exists(output_directory): 31 elif not exists(output_directory):
36 mkdir(output_directory) 32 mkdir(output_directory)
37 33
38 if 'GALAXY_SLOTS' in environ: 34 if 'GALAXY_SLOTS' in environ:
39 nslots = environ['GALAXY_SLOTS'] 35 nslots = environ['GALAXY_SLOTS']
40 else: 36 else:
41 nslots = 1 37 nslots = 1
42 38
43 # cmdline_str = 'STAR --runMode genomeGenerate --genomeDir {} --genomeFastaFiles {} --runThreadsN {}'.format(output_directory, 39 cmdline_str = 'STAR --runMode genomeGenerate --genomeDir {} --genomeFastaFiles {} --runThreadsN {}'.format(
44 # fasta_filename, 40 output_directory,
45 # nslots) 41 fasta_filename,
46 # cmdline = shlex.split(cmdline_str) 42 nslots)
47 cmdline = ('touch', '{}/foo'.format(output_directory)) 43 cmdline = shlex.split(cmdline_str)
44 # cmdline = ('touch', '{}/foo'.format(output_directory))
48 try: 45 try:
49 check_call(cmdline) 46 check_call(cmdline)
50 except CalledProcessError: 47 except CalledProcessError:
51 print("Error building RNA STAR index", file=sys.stderr) 48 print("Error building RNA STAR index", file=sys.stderr)
52 return(output_directory) 49 return (output_directory)
53 50
54 parser = argparse.ArgumentParser(description="Generate RNA STAR genome index and JSON describing this")
55 parser.add_argument('output_filename')
56 parser.add_argument('--fasta_filename')
57 parser.add_argument('--fasta_dbkey')
58 parser.add_argument('--fasta_description', default=None)
59 parser.add_argument('--data_table_name', default='rnastar_index')
60 args = parser.parse_args()
61 51
62 filename = args.output_filename 52 def main():
53 parser = argparse.ArgumentParser(description="Generate RNA STAR genome index and JSON describing this")
54 parser.add_argument('output_filename')
55 parser.add_argument('--fasta_filename')
56 parser.add_argument('--fasta_dbkey')
57 parser.add_argument('--fasta_description', default=None)
58 parser.add_argument('--data_table_name', default='rnastar_index')
59 args = parser.parse_args()
63 60
64 params = load(open(filename, 'rb')) 61 filename = args.output_filename
65 output_directory = params[ 'output_data' ][0]['extra_files_path']
66 makedirs( output_directory )
67 data_manager_dict = {}
68 62
69 make_rnastar_index(output_directory, args.fasta_filename) 63 params = load(open(filename, 'rb'))
70 (sequence_id, sequence_name) = get_id_name(params, args.fasta_dbkey, args.fasta_description) 64 output_directory = params['output_data'][0]['extra_files_path']
71 data_table_entry = dict(value=sequence_id, dbkey=args.fasta_dbkey, name=sequence_name, path=output_directory) 65 makedirs(output_directory)
72 66
73 output_datatable_dict = dict(data_tables={args.data_table_name : [data_table_entry]}) 67 make_rnastar_index(output_directory, args.fasta_filename)
74 open( filename, 'wb' ).write( dumps( output_datatable_dict ) ) 68 (sequence_id, sequence_name) = get_id_name(params, args.fasta_dbkey, args.fasta_description)
69 data_table_entry = dict(value=sequence_id, dbkey=args.fasta_dbkey, name=sequence_name, path=output_directory)
70
71 output_datatable_dict = dict(data_tables={args.data_table_name: [data_table_entry]})
72 open(filename, 'wb').write(dumps(output_datatable_dict))
73
74 if __name__ == "__main__": main()