Mercurial > repos > devteam > data_manager_fetch_ncbi_taxonomy

annotate data_manager/data_manager.py @ 4:cef5c909ccb8 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_fetch_ncbi_taxonomy/ commit 79946a95eb9a4c98fde04fa2cb3bc6a31a28e407"
author iuc
date Tue, 07 Jul 2020 00:05:59 +0000
parents f7b443a73374
children eaca3e270bf6
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
1 import argparse
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
2 import datetime
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
3 import json
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
4 import os
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
5 import shutil
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
6 import tarfile
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
7 import zipfile
4
cef5c909ccb8 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 79946a95eb9a4c98fde04fa2cb3bc6a31a28e407"
iuc
parents: 3
diff changeset
8 try:
cef5c909ccb8 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 79946a95eb9a4c98fde04fa2cb3bc6a31a28e407"
iuc
parents: 3
diff changeset
9 # For Python 3.0 and later
cef5c909ccb8 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 79946a95eb9a4c98fde04fa2cb3bc6a31a28e407"
iuc
parents: 3
diff changeset
10 from urllib.request import Request, urlopen
cef5c909ccb8 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 79946a95eb9a4c98fde04fa2cb3bc6a31a28e407"
iuc
parents: 3
diff changeset
11 except ImportError:
cef5c909ccb8 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 79946a95eb9a4c98fde04fa2cb3bc6a31a28e407"
iuc
parents: 3
diff changeset
12 # Fall back to Python 2 imports
cef5c909ccb8 "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 79946a95eb9a4c98fde04fa2cb3bc6a31a28e407"
iuc
parents: 3
diff changeset
13 from urllib2 import Request, urlopen
0
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
14
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
15
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
16 def url_download(url, workdir):
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
17 file_path = os.path.join(workdir, 'download.dat')
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
18 if not os.path.exists(workdir):
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
19 os.makedirs(workdir)
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
20 src = None
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
21 dst = None
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
22 try:
1
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
23 req = Request(url)
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
24 src = urlopen(req)
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
25 with open(file_path, 'wb') as dst:
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
26 while True:
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
27 chunk = src.read(2**10)
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
28 if chunk:
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
29 dst.write(chunk)
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
30 else:
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
31 break
0
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
32 finally:
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
33 if src:
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
34 src.close()
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
35 if tarfile.is_tarfile(file_path):
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
36 fh = tarfile.open(file_path, 'r:*')
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
37 elif zipfile.is_zipfile(file_path):
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
38 fh = zipfile.ZipFile(file_path, 'r')
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
39 else:
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
40 return
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
41 fh.extractall(workdir)
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
42 os.remove(file_path)
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
43
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
44
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
45 def main(args):
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
46 workdir = os.path.join(os.getcwd(), 'taxonomy')
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
47 url_download(args.url, workdir)
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
48 data_manager_entry = {}
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
49 data_manager_entry['value'] = args.name.lower()
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
50 data_manager_entry['name'] = args.name
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
51 data_manager_entry['path'] = '.'
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
52 data_manager_json = dict(data_tables=dict(ncbi_taxonomy=data_manager_entry))
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
53 params = json.loads(open(args.output).read())
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
54 target_directory = params['output_data'][0]['extra_files_path']
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
55 os.mkdir(target_directory)
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
56 output_path = os.path.abspath(os.path.join(os.getcwd(), 'taxonomy'))
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
57 for filename in os.listdir(workdir):
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
58 shutil.move(os.path.join(output_path, filename), target_directory)
2
3d628f1e1bef "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 04825a097d217d30deb5fcb6906fe718023e25b6"
iuc
parents: 1
diff changeset
59 with open(args.output, 'w') as out:
3d628f1e1bef "planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 04825a097d217d30deb5fcb6906fe718023e25b6"
iuc
parents: 1
diff changeset
60 out.write(json.dumps(data_manager_json, sort_keys=True))
0
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
61
1
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
62
0
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
63 if __name__ == '__main__':
1
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
64 parser = argparse.ArgumentParser(description='Create data manager json.')
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
65 parser.add_argument('--out', dest='output', action='store', help='JSON filename')
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
66 parser.add_argument('--name', dest='name', action='store', default=str(datetime.date.today()), help='Data table entry unique ID')
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
67 parser.add_argument('--url', dest='url', action='store', default='ftp://ftp.ncbi.nih.gov/pub/taxonomy/taxdump.tar.gz', help='Download URL')
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
68 args = parser.parse_args()
8dd4076813ef planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 8652f36a3a3838dca989426961561e81432acf4f
iuc
parents: 0
diff changeset
69
0
078e803565cf planemo upload for repository https://github.com/galaxyproject/tools-devteam/blob/master/data_managers/data_manager_fetch_ncbi_taxonomy/ commit 86cf90107482cab1cb47fc0d42d6705f8077daa7
devteam
parents:
diff changeset
70 main(args)