Mercurial > repos > earlhaminst > ensembl_get_sequences

annotate get_feature_info.py @ 4:3b686142e9c2 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
author earlhaminst
date Fri, 13 Apr 2018 09:43:51 -0400
parents 4b7261f484bb
children 0fa1d1cc417d
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
1
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
1 # A simple tool to connect to the Ensembl server and retrieve feature
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
2 # information using the Ensembl REST API.
2
4b7261f484bb planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 099d38157cec200f0a343579ca9babcd8acb266f
earlhaminst
parents: 1
diff changeset
3 from __future__ import print_function
4b7261f484bb planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 099d38157cec200f0a343579ca9babcd8acb266f
earlhaminst
parents: 1
diff changeset
4
1
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
5 import json
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
6 import optparse
4
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
7 from itertools import islice
1
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
8
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
9 import requests
2
4b7261f484bb planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 099d38157cec200f0a343579ca9babcd8acb266f
earlhaminst
parents: 1
diff changeset
10 from six.moves.urllib.parse import urljoin
1
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
11
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
12 parser = optparse.OptionParser()
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
13 parser.add_option('-i', '--input', help='List of Ensembl IDs')
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
14 parser.add_option('-e', '--expand', type='choice', choices=['0', '1'],
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
15 default='0',
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
16 help='Expands the search to include any connected features. e.g. If the object is a gene, its transcripts, translations and exons will be returned as well.')
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
17
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
18 parser.add_option('-s', '--species', type='choice',
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
19 choices=['ensembl', 'ensemblgenomes'], default='ensembl',
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
20 help='Specify the genome databases for vertebrates and other eukaryotic species')
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
21
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
22 parser.add_option('-f', '--format', type='choice',
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
23 choices=['full', 'condensed'], default='full',
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
24 help='Specify the formats to emit from this endpoint')
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
25 options, args = parser.parse_args()
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
26 if options.input is None:
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
27 raise Exception('-i option must be specified')
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
28
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
29
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
30 server = 'http://rest.%s.org' % options.species
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
31 ext = 'lookup/id'
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
32
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
33 headers = {'Content-Type': 'application/json', 'Accept': 'application/json'}
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
34 params = dict((k, getattr(options, k)) for k in ['format', 'expand'])
4
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
35
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
36 first = True
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
37
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
38 print('{')
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
39
1
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
40 with open(options.input) as f:
4
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
41 while True:
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
42 ids = [line.strip() for line in islice(f, 50)]
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
43 if not ids:
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
44 break
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
45 if not first:
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
46 print(",")
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
47 data = {'ids': ids}
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
48 r = requests.post(urljoin(server, ext), params=params, headers=headers,
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
49 data=json.dumps(data))
1
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
50
4
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
51 if not r.ok:
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
52 r.raise_for_status()
1
e5dd4bd78bbc planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit aaf8d501c3a92ed415fdf9293a65468c72aae984-dirty
earlhaminst
parents:
diff changeset
53
4
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
54 print(r.text[1:-1])
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
55
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
56 first = False
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
57
3b686142e9c2 planemo upload for repository https://github.com/TGAC/earlham-galaxytools/tree/master/tools/Ensembl-REST commit 95bab1105cf8a7b07c668f08f712399e8775a4ae
earlhaminst
parents: 2
diff changeset
58 print('}')