Mercurial > repos > greg > data_manager_malt_index_builder

comparison data_manager/malt_index_builder.py @ 0:d69ebf52c233 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author greg
date Tue, 12 Oct 2021 14:15:35 +0000
parents
children 787f1ca9045a
comparison
equal deleted inserted replaced
-1:000000000000 0:d69ebf52c233
1 #!/usr/bin/env python
2
3 import json
4 import optparse
5 import os
6 import subprocess
7 import sys
8
9 DEFAULT_DATA_TABLE_NAME = "malt_indices"
10
11
12 def get_id_name(params, dbkey, fasta_description=None):
13 sequence_id = params['param_dict']['sequence_id']
14 if not sequence_id:
15 sequence_id = dbkey
16
17 sequence_name = params['param_dict']['sequence_name']
18 if not sequence_name:
19 sequence_name = fasta_description
20 if not sequence_name:
21 sequence_name = dbkey
22 return sequence_id, sequence_name
23
24
25 def build_malt_index(data_manager_dict, fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, sequence_type, shapes, max_hits_per_seed, protein_reduct, data_table_name=DEFAULT_DATA_TABLE_NAME):
26 fasta_base_name = os.path.split(fasta_filename)[-1]
27 sym_linked_fasta_filename = os.path.join(target_directory, fasta_base_name)
28 os.symlink(fasta_filename, sym_linked_fasta_filename)
29 args = ['malt-build', '--input', sym_linked_fasta_filename, '--sequenceType', sequence_type, '--index', 'index']
30 threads = os.environ.get('GALAXY_SLOTS')
31 if threads:
32 args.extend(['--threads', threads])
33 if shapes is not None:
34 args.extend('--shapes', shapes)
35 if max_hits_per_seed is not None:
36 args.extend('--maxHitsPerSeed', max_hits_per_seed)
37 if protein_reduct is not None:
38 args.extend('--proteinReduct', protein_reduct)
39 proc = subprocess.Popen(args=args, shell=False, cwd=target_directory)
40 return_code = proc.wait()
41 if return_code:
42 sys.exit('Error building index, return_code: %d' % return_code)
43 data_table_entry = dict(value=sequence_id, dbkey=dbkey, name=sequence_name, path=fasta_base_name)
44 _add_data_table_entry(data_manager_dict, data_table_name, data_table_entry)
45
46
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)
51 return data_manager_dict
52
53
54 def main():
55 parser = optparse.OptionParser()
56 parser.add_option('-f', '--fasta_filename', dest='fasta_filename', action='store', type="string", help='fasta filename')
57 parser.add_option('-d', '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", 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('-e', '--sequence_type', dest='sequence_type', action='store', type="string", help='DNA or Protein sequences')
60 parser.add_option('-p', '--shapes', dest='shapes', action='store', type="string", default=None, help='Comma-separated list of seed shapes')
61 parser.add_option('-m', '--max_hits_per_seed', dest='max_hits_per_seed', action='store', type="int", default=None, help='Maximum number of hits per seed')
62 parser.add_option('-r', '--protein_reduct', dest='protein_reduct', action='store', type="string", default=None, help='Name or definition of protein alphabet reduction')
63 (options, args) = parser.parse_args()
64
65 filename = args[0]
66
67 with open(filename) as fh:
68 params = json.load(fh)
69 target_directory = params['output_data'][0]['extra_files_path']
70 os.mkdir(target_directory)
71 data_manager_dict = {}
72
73 dbkey = options.fasta_dbkey
74
75 if dbkey in [None, '', '?']:
76 raise Exception('"%s" is not a valid dbkey. You must specify a valid dbkey.' % (dbkey))
77
78 sequence_id, sequence_name = get_id_name(params, dbkey=dbkey, fasta_description=options.fasta_description)
79
80 # Build the index.
81 build_malt_index(data_manager_dict, options.fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name, options.sequence_type, options.shapes, options.max_hits_per_seed, options.protein_reduct, data_table_name=DEFAULT_DATA_TABLE_NAME)
82
83 # Save info to json file.
84 with open(filename, 'w') as fh:
85 json.dump(data_manager_dict, fh, sort_keys=True)
86
87
88 if __name__ == "__main__":
89 main()