Mercurial > repos > devteam > data_manager_twobit_builder
comparison data_manager/twobit_builder.py @ 2:c9896bb1f03c draft
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_twobit_builder commit 8652f36a3a3838dca989426961561e81432acf4f
author | iuc |
---|---|
date | Tue, 04 Apr 2017 17:54:54 -0400 |
parents | 26e3ae5f74ed |
children | ced9f4a48806 |
comparison
equal
deleted
inserted
replaced
1:298d15849a14 | 2:c9896bb1f03c |
---|---|
1 #!/usr/bin/env python | 1 #!/usr/bin/env python |
2 #Dan Blankenberg | 2 # Dan Blankenberg |
3 from __future__ import print_function | |
3 | 4 |
4 import sys, os, tempfile, optparse, uuid, subprocess | 5 import optparse |
6 import os | |
7 import subprocess | |
8 import sys | |
9 import tempfile | |
10 from json import dumps, loads | |
5 | 11 |
6 from json import loads, dumps | 12 CHUNK_SIZE = 2**20 # 1mb |
7 | 13 |
8 | 14 |
9 CHUNK_SIZE = 2**20 #1mb | |
10 | |
11 def get_id_name( params, dbkey, fasta_description=None): | 15 def get_id_name( params, dbkey, fasta_description=None): |
12 #TODO: ensure sequence_id is unique and does not already appear in location file | 16 # TODO: ensure sequence_id is unique and does not already appear in location file |
13 sequence_id = params['param_dict']['sequence_id'] | 17 sequence_id = params['param_dict']['sequence_id'] |
14 if not sequence_id: | 18 if not sequence_id: |
15 sequence_id = dbkey #uuid.uuid4() generate and use an uuid | 19 sequence_id = dbkey # uuid.uuid4() generate and use an uuid |
16 | 20 |
17 sequence_name = params['param_dict']['sequence_name'] | 21 sequence_name = params['param_dict']['sequence_name'] |
18 if not sequence_name: | 22 if not sequence_name: |
19 sequence_name = fasta_description | 23 sequence_name = fasta_description |
20 if not sequence_name: | 24 if not sequence_name: |
21 sequence_name = dbkey | 25 sequence_name = dbkey |
22 return sequence_id, sequence_name | 26 return sequence_id, sequence_name |
23 | 27 |
28 | |
24 def build_twobit( data_manager_dict, fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name ): | 29 def build_twobit( data_manager_dict, fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name ): |
25 twobit_base_name = "%s.2bit" % ( sequence_id ) | 30 twobit_base_name = "%s.2bit" % ( sequence_id ) |
26 twobit_filename = os.path.join( target_directory, twobit_base_name ) | 31 twobit_filename = os.path.join( target_directory, twobit_base_name ) |
27 | 32 |
28 args = [ 'faToTwoBit', fasta_filename, twobit_filename ] | 33 args = [ 'faToTwoBit', fasta_filename, twobit_filename ] |
29 tmp_stderr = tempfile.NamedTemporaryFile( prefix = "tmp-data-manager-twobit-builder-stderr" ) | 34 tmp_stderr = tempfile.NamedTemporaryFile( prefix="tmp-data-manager-twobit-builder-stderr" ) |
30 proc = subprocess.Popen( args=args, shell=False, cwd=target_directory, stderr=tmp_stderr.fileno() ) | 35 proc = subprocess.Popen( args=args, shell=False, cwd=target_directory, stderr=tmp_stderr.fileno() ) |
31 return_code = proc.wait() | 36 return_code = proc.wait() |
32 if return_code: | 37 if return_code: |
33 tmp_stderr.flush() | 38 tmp_stderr.flush() |
34 tmp_stderr.seek(0) | 39 tmp_stderr.seek(0) |
35 print >> sys.stderr, "Error building index:" | 40 print("Error building index:", file=sys.stderr) |
36 while True: | 41 while True: |
37 chunk = tmp_stderr.read( CHUNK_SIZE ) | 42 chunk = tmp_stderr.read( CHUNK_SIZE ) |
38 if not chunk: | 43 if not chunk: |
39 break | 44 break |
40 sys.stderr.write( chunk ) | 45 sys.stderr.write( chunk ) |
41 sys.exit( return_code ) | 46 sys.exit( return_code ) |
42 tmp_stderr.close() | 47 tmp_stderr.close() |
43 #lastz_seqs | 48 # lastz_seqs |
44 data_table_entry = dict( value=sequence_id, name=sequence_name, path=twobit_base_name ) | 49 data_table_entry = dict( value=sequence_id, name=sequence_name, path=twobit_base_name ) |
45 | 50 |
46 _add_data_table_entry( data_manager_dict, "lastz_seqs", data_table_entry ) | 51 _add_data_table_entry( data_manager_dict, "lastz_seqs", data_table_entry ) |
47 #twobit.loc | 52 # twobit.loc |
48 data_table_entry = dict( value=sequence_id, path=twobit_base_name ) | 53 data_table_entry = dict( value=sequence_id, path=twobit_base_name ) |
49 | 54 |
50 _add_data_table_entry( data_manager_dict, "twobit", data_table_entry ) | 55 _add_data_table_entry( data_manager_dict, "twobit", data_table_entry ) |
51 #alignseq | 56 # alignseq |
52 data_table_entry = dict( type="seq", value=sequence_id, path=twobit_base_name ) | 57 data_table_entry = dict( type="seq", value=sequence_id, path=twobit_base_name ) |
53 | 58 |
54 _add_data_table_entry( data_manager_dict, "alignseq_seq", data_table_entry ) | 59 _add_data_table_entry( data_manager_dict, "alignseq_seq", data_table_entry ) |
60 | |
55 | 61 |
56 def _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ): | 62 def _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ): |
57 data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} ) | 63 data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} ) |
58 data_manager_dict['data_tables'][ data_table_name ] = data_manager_dict['data_tables'].get( data_table_name, [] ) | 64 data_manager_dict['data_tables'][ data_table_name ] = data_manager_dict['data_tables'].get( data_table_name, [] ) |
59 data_manager_dict['data_tables'][ data_table_name ].append( data_table_entry ) | 65 data_manager_dict['data_tables'][ data_table_name ].append( data_table_entry ) |
60 return data_manager_dict | 66 return data_manager_dict |
61 | 67 |
68 | |
62 def main(): | 69 def main(): |
63 #Parse Command Line | |
64 parser = optparse.OptionParser() | 70 parser = optparse.OptionParser() |
65 parser.add_option( '-f', '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename' ) | 71 parser.add_option( '-f', '--fasta_filename', dest='fasta_filename', action='store', type="string", default=None, help='fasta_filename' ) |
66 parser.add_option( '-d', '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey' ) | 72 parser.add_option( '-d', '--fasta_dbkey', dest='fasta_dbkey', action='store', type="string", default=None, help='fasta_dbkey' ) |
67 parser.add_option( '-t', '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description' ) | 73 parser.add_option( '-t', '--fasta_description', dest='fasta_description', action='store', type="string", default=None, help='fasta_description' ) |
68 (options, args) = parser.parse_args() | 74 (options, args) = parser.parse_args() |
69 | 75 |
70 filename = args[0] | 76 filename = args[0] |
71 | 77 |
72 params = loads( open( filename ).read() ) | 78 params = loads( open( filename ).read() ) |
73 | 79 |
74 target_directory = params[ 'output_data' ][0]['extra_files_path'] | 80 target_directory = params[ 'output_data' ][0]['extra_files_path'] |
75 os.mkdir( target_directory ) | 81 os.mkdir( target_directory ) |
76 data_manager_dict = {} | 82 data_manager_dict = {} |
77 | 83 |
78 dbkey = options.fasta_dbkey | 84 dbkey = options.fasta_dbkey |
79 | 85 |
80 if dbkey in [ None, '', '?' ]: | 86 if dbkey in [ None, '', '?' ]: |
81 raise Exception( '"%s" is not a valid dbkey. You must specify a valid dbkey.' % ( dbkey ) ) | 87 raise Exception( '"%s" is not a valid dbkey. You must specify a valid dbkey.' % ( dbkey ) ) |
82 | 88 |
83 sequence_id, sequence_name = get_id_name( params, dbkey=dbkey, fasta_description=options.fasta_description ) | 89 sequence_id, sequence_name = get_id_name( params, dbkey=dbkey, fasta_description=options.fasta_description ) |
84 | 90 |
85 #build the index | 91 # build the index |
86 build_twobit( data_manager_dict, options.fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name ) | 92 build_twobit( data_manager_dict, options.fasta_filename, params, target_directory, dbkey, sequence_id, sequence_name ) |
87 | 93 |
88 #save info to json file | 94 # save info to json file |
89 open( filename, 'wb' ).write( dumps( data_manager_dict ) ) | 95 open( filename, 'wb' ).write( dumps( data_manager_dict ) ) |
90 | |
91 if __name__ == "__main__": main() | |
92 | 96 |
97 | |
98 if __name__ == "__main__": | |
99 main() |