Mercurial > repos > iuc > data_manager_homer_preparse

comparison data_manager/homer_genome_preparse.py @ 0:e6a2110ac3b9 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_homer_preparse commit e49d856e0181edc6120220a1b819cba2466a4289"
author iuc
date Sun, 08 Aug 2021 11:01:47 +0000
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:e6a2110ac3b9
1 #!/usr/bin/env python
2 # Dan Blankenberg for bowtie2
3 # Modified by Lucille Delisle for homer
4 from __future__ import print_function
5
6 import json
7 import optparse
8 import os
9 import subprocess
10 import sys
11
12 DEFAULT_DATA_TABLE_NAME = "homer_preparse"
13
14
15 def get_id_name(params, dbkey, fasta_description=None):
16 # TODO: ensure sequence_id is unique and does not already appear in location file
17 sequence_id = params['param_dict']['sequence_id']
18 if not sequence_id:
19 sequence_id = dbkey
20
21 sequence_name = params['param_dict']['sequence_name']
22 if not sequence_name:
23 sequence_name = fasta_description
24 if not sequence_name:
25 sequence_name = dbkey
26 return sequence_id, sequence_name
27
28
29 def homer_preparse(data_manager_dict, fasta_filename, params, target_directory, dbkey, sequence_id,
30 sequence_name, size, mask, version,
31 data_table_name=DEFAULT_DATA_TABLE_NAME):
32 args = ['preparseGenome.pl', fasta_filename, '-size', str(size), '-preparsedDir', target_directory]
33 if mask:
34 args.append('-mask')
35 proc = subprocess.Popen(args=args, shell=False, cwd=target_directory)
36 return_code = proc.wait()
37 if return_code:
38 print("Error preparsing genome.", file=sys.stderr)
39 sys.exit(return_code)
40 mask_suffix = 'r' if mask else ''
41 mask_suffix_name = ' masked' if mask else ''
42 data_table_entry = dict(value=sequence_id + mask_suffix + '_' + str(size), dbkey=dbkey,
43 mask=str(mask), size=str(size), name=sequence_name + mask_suffix_name + ' (' + str(size) + 'bp)',
44 path=sequence_id + mask_suffix + '_' + str(size),
45 path_fasta=fasta_filename,
46 version=version)
47 _add_data_table_entry(data_manager_dict, data_table_name, data_table_entry)
48
49
50 def _add_data_table_entry(data_manager_dict, data_table_name, data_table_entry):
51 data_manager_dict['data_tables'] = data_manager_dict.get('data_tables', {})
52 data_manager_dict['data_tables'][data_table_name] = data_manager_dict['data_tables'].get(data_table_name, [])
53 data_manager_dict['data_tables'][data_table_name].append(data_table_entry)
54 return data_manager_dict
55
56
57 def main():
58 parser = optparse.OptionParser()
59 parser.add_option('-f', '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename')
60 parser.add_option('-d', '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey')
61 parser.add_option('-t', '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description')
62 parser.add_option('-s', '--size', dest='size', action='store', type="int", default=200, help='fragment size')
63 parser.add_option('-m', '--mask', dest='mask', action='store_true', default=False, help='mask the lower case bases (repeats)')
64 parser.add_option('-n', '--data_table_name', dest='data_table_name', action='store', type="string", default=None, help='data_table_name')
65 parser.add_option('--index_version', dest='index_version', action='store', type="string", default=None, help='index version')
66 (options, args) = parser.parse_args()
67
68 filename = args[0]
69
70 with open(filename) as fh:
71 params = json.load(fh)
72 target_directory = params['output_data'][0]['extra_files_path']
73 os.mkdir(target_directory)
74 data_manager_dict = {}
75
76 dbkey = options.fasta_dbkey
77
78 if dbkey in [None, '', '?']:
79 raise Exception('"%s" is not a valid dbkey. You must specify a valid dbkey.' % (dbkey))
80
81 sequence_id, sequence_name = get_id_name(params, dbkey=dbkey, fasta_description=options.fasta_description)
82
83 # preparse the genome
84 homer_preparse(data_manager_dict, options.fasta_filename, params, target_directory, dbkey, sequence_id,
85 sequence_name, options.size, options.mask, options.index_version,
86 data_table_name=options.data_table_name or DEFAULT_DATA_TABLE_NAME)
87
88 # save info to json file
89 with open(filename, 'w') as fh:
90 json.dump(data_manager_dict, fh, sort_keys=True)
91
92
93 if __name__ == "__main__":
94 main()