Mercurial > repos > brenninc > data_manager_for_directory_data

comparison data_manager/directory_data_manager.py @ 0:43083927dfc5 draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author brenninc
date Sat, 07 May 2016 16:59:47 -0400
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:43083927dfc5
1 #!/usr/bin/env python
2
3 import gzip
4 import json
5 import optparse # using optparse as hydra still python 2.6
6 import os.path
7 import shutil
8
9 def _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry ):
10 data_manager_dict['data_tables'] = data_manager_dict.get( 'data_tables', {} )
11 data_manager_dict['data_tables'][ data_table_name ] = data_manager_dict['data_tables'].get( data_table_name, [] )
12 data_manager_dict['data_tables'][ data_table_name ].append( data_table_entry )
13 return data_manager_dict
14
15
16 def get_param(name, params, default=None, check_tab=True):
17 value = params.get(name)
18 print name, value
19 return check_param(name, value, default=default, check_tab=check_tab)
20
21
22 def check_param(name, value, default=None, check_tab=True):
23 if value in [ None, '', '?' ]:
24 if default:
25 print "Using {0} for {1} as no value provided".format( default, name )
26 value = default
27 else:
28 raise Exception( '{0} is not a valid {1}. You must specify a valid {1}.'.format( value, name ) )
29 if check_tab and "\t" in value:
30 raise Exception( '{0} is not a valid {1}. It may not contain a tab because these are used as seperators by galaxy .'.format( value, name ) )
31 return value
32
33 def check_extension(extension):
34 extension = extension.strip()
35 if extension[0] == ".":
36 extension = extension[1:]
37 return extension
38
39
40 def check_path(path, original_extension):
41 files = os.listdir(path)
42 check = "." + original_extension
43 for a_file in files:
44 if a_file.endswith(check):
45 return True
46 raise Exception( 'path {0} does not contain any files ending with {1}'.format( path, check ) )
47
48
49 def main():
50
51 #Parse Command Line
52 parser = optparse.OptionParser()
53 parser.add_option( '--data_table_name', action='store', type="string", default=None, help='path' )
54 parser.add_option( '--json_output_file', action='store', type="string", default=None, help='path' )
55 (options, args) = parser.parse_args()
56
57 data_table_name = check_param("data_table_name", options.data_table_name)
58 json_output_file = check_param("json_output_file", options.json_output_file, check_tab=False)
59
60 param_dict = json.loads( open( json_output_file ).read() )
61 params = param_dict.get("param_dict")
62 print "input params:"
63 print params
64
65 data_table_entry = {}
66 data_table_entry["original_extension"] = check_extension(get_param("original_extension", params))
67 data_table_entry["galaxy_extension"] = check_extension(get_param("galaxy_extension", params))
68 data_table_entry["decompress"] = get_param("decompress", params)
69 if not (data_table_entry["decompress"] in ["No","Yes"]):
70 raise Exception( "Only legal values for dcompress are No and Yes." )
71 data_table_entry["path"] = get_param("path", params)
72 check_path(data_table_entry["path"], data_table_entry["original_extension"])
73
74 basename = os.path.basename(data_table_entry["path"])
75 filename = os.path.splitext(basename)[0]
76 data_table_entry["name"] = get_param("name", params, default=filename)
77 data_table_entry["value"] = get_param("value", params, default=data_table_entry["name"])
78 data_table_entry["dbkey"] = get_param("dbkey", params, default=data_table_entry["value"])
79
80 data_manager_dict = {}
81 _add_data_table_entry( data_manager_dict, data_table_name, data_table_entry )
82
83 print "output:"
84 print data_manager_dict
85 # save info to json file
86 with open( json_output_file, 'wb' ) as output_file:
87 output_file.write( json.dumps( data_manager_dict ) )
88 output_file.write( "\n" )
89
90
91 if __name__ == "__main__":
92 main()