|
0
|
1 #!/usr/bin/env python
|
|
|
2 # Dan Blankenberg
|
|
|
3 from __future__ import print_function
|
|
|
4
|
|
|
5 import optparse
|
|
|
6 import os
|
|
|
7 import subprocess
|
|
|
8 import sys
|
|
|
9 from json import dumps, loads
|
|
|
10
|
|
|
11 CHUNK_SIZE = 2**20
|
|
|
12 TWO_GB = 2**30 * 2
|
|
|
13 DEFAULT_DATA_TABLE_NAME = "mummer4_indexes"
|
|
|
14
|
|
|
15
|
|
1
|
16 def get_id_name(params, dbkey, fasta_description=None):
|
|
0
|
17 # TODO: ensure sequence_id is unique and does not already appear in location file
|
|
|
18 sequence_id = params['param_dict']['sequence_id']
|
|
|
19 if not sequence_id:
|
|
|
20 sequence_id = dbkey
|
|
|
21
|
|
|
22 sequence_name = params['param_dict']['sequence_name']
|
|
|
23 if not sequence_name:
|
|
|
24 sequence_name = fasta_description
|
|
|
25 if not sequence_name:
|
|
|
26 sequence_name = dbkey
|
|
|
27 return sequence_id, sequence_name
|
|
|
28
|
|
|
29
|
|
1
|
30 def build_mummer4_index(data_manager_dict, fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, data_table_name=DEFAULT_DATA_TABLE_NAME):
|
|
0
|
31 # TODO: allow multiple FASTA input files
|
|
1
|
32 fasta_base_name = os.path.split(fasta_filename)[-1]
|
|
|
33 sym_linked_fasta_filename = os.path.join(target_directory, fasta_base_name)
|
|
|
34 os.symlink(fasta_filename, sym_linked_fasta_filename)
|
|
0
|
35
|
|
1
|
36 args = ['mummer', '-save', dbkey, sym_linked_fasta_filename, sym_linked_fasta_filename]
|
|
|
37 args.append(sym_linked_fasta_filename)
|
|
|
38 proc = subprocess.Popen(args=args, shell=False, cwd=target_directory)
|
|
0
|
39 return_code = proc.wait()
|
|
|
40 if return_code:
|
|
|
41 print("Error building index.", file=sys.stderr)
|
|
1
|
42 sys.exit(return_code)
|
|
|
43 data_table_entry = dict(value=sequence_id, dbkey=dbkey, name=sequence_name, path=dbkey)
|
|
|
44 _add_data_table_entry(data_manager_dict, data_table_name, data_table_entry)
|
|
0
|
45
|
|
|
46
|
|
1
|
47 def _add_data_table_entry(data_manager_dict, data_table_name, data_table_entry):
|
|
|
48 data_manager_dict['data_tables'] = data_manager_dict.get('data_tables', {})
|
|
|
49 data_manager_dict['data_tables'][data_table_name] = data_manager_dict['data_tables'].get(data_table_name, [])
|
|
|
50 data_manager_dict['data_tables'][data_table_name].append(data_table_entry)
|
|
0
|
51 return data_manager_dict
|
|
|
52
|
|
|
53
|
|
|
54 def main():
|
|
|
55 parser = optparse.OptionParser()
|
|
1
|
56 parser.add_option('-f', '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename')
|
|
|
57 parser.add_option('-d', '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey')
|
|
|
58 parser.add_option('-t', '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description')
|
|
|
59 parser.add_option('-n', '--data_table_name', dest='data_table_name', action='store', type="string", default=None, help='data_table_name')
|
|
0
|
60 (options, args) = parser.parse_args()
|
|
|
61
|
|
|
62 filename = args[0]
|
|
|
63
|
|
1
|
64 params = loads(open(filename).read())
|
|
|
65 target_directory = params['output_data'][0]['extra_files_path']
|
|
|
66 os.mkdir(target_directory)
|
|
0
|
67 data_manager_dict = {}
|
|
|
68
|
|
|
69 dbkey = options.fasta_dbkey
|
|
|
70
|
|
1
|
71 if dbkey in [None, '', '?']:
|
|
|
72 raise Exception('"%s" is not a valid dbkey. You must specify a valid dbkey.' % (dbkey))
|
|
0
|
73
|
|
1
|
74 sequence_id, sequence_name = get_id_name(params, dbkey=dbkey, fasta_description=options.fasta_description)
|
|
0
|
75
|
|
|
76 # build the index
|
|
1
|
77 build_mummer4_index(data_manager_dict, options.fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, data_table_name=options.data_table_name or DEFAULT_DATA_TABLE_NAME)
|
|
0
|
78
|
|
|
79 # save info to json file
|
|
1
|
80 open(filename, 'wb').write(dumps(data_manager_dict))
|
|
0
|
81
|
|
|
82
|
|
|
83 if __name__ == "__main__":
|
|
|
84 main()
|
