Mercurial > repos > earlhaminst > ensembl_get_sequences

comparison get_genetree.py @ 7:c79ce2342f1e draft default tip

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 8f8808de862973aedbf87abd4dfa9d2dc7219322
author earlhaminst
date Mon, 17 Feb 2025 14:49:24 +0000
parents 7af66c2b3831
children
comparison
equal deleted inserted replaced
6:7af66c2b3831 7:c79ce2342f1e
1 # A simple tool to connect to the Ensembl server and retrieve genetree using 1 # A simple tool to connect to the Ensembl server and retrieve genetree using
2 # the Ensembl REST API. 2 # the Ensembl REST API.
3 from __future__ import print_function
4
5 import optparse 3 import optparse
4 from urllib.parse import urljoin
6 5
7 import requests 6 import requests
8 from six.moves.urllib.parse import urljoin
9 7
10 parser = optparse.OptionParser() 8 parser = optparse.OptionParser()
11 parser.add_option('--id_type', type='choice', default='gene_id', 9 parser.add_option(
12 choices=['gene_id', 'gene_tree_id'], help='Input type') 10 "--id_type",
13 parser.add_option('-i', '--input', help='Ensembl ID') 11 type="choice",
14 parser.add_option('--format', type='choice', 12 default="gene_id",
15 choices=['json', 'orthoxml', 'phyloxml', 'nh'], 13 choices=["gene_id", "gene_tree_id"],
16 default='json', help='Output format') 14 help="Input type",
17 parser.add_option('-s', '--sequence', type='choice', 15 )
18 choices=['protein', 'cdna', 'none'], default='protein', 16 parser.add_option("--species", help="Species name/alias")
19 help='The type of sequence to bring back. Setting it to none results in no sequence being returned') 17 parser.add_option("-i", "--input", help="Ensembl ID")
18 parser.add_option(
19 "--format",
20 type="choice",
21 choices=["json", "orthoxml", "phyloxml", "nh"],
22 default="json",
23 help="Output format",
24 )
25 parser.add_option(
26 "-s",
27 "--sequence",
28 type="choice",
29 choices=["protein", "cdna", "none"],
30 default="protein",
31 help="The type of sequence to bring back. Setting it to none results in no sequence being returned",
32 )
20 33
21 parser.add_option('-a', '--aligned', type='choice', choices=['0', '1'], 34 parser.add_option(
22 default='0', help='Return the aligned string if true. Otherwise, return the original sequence (no insertions)') 35 "-a",
23 parser.add_option('-c', '--cigar_line', type='choice', choices=['0', '1'], 36 "--aligned",
24 default='0', 37 type="choice",
25 help='Return the aligned sequence encoded in CIGAR format') 38 choices=["0", "1"],
26 parser.add_option('--nh_format', type='choice', 39 default="0",
27 choices=['full', 'display_label_composite', 'simple', 'species', 'species_short_name', 'ncbi_taxon', 'ncbi_name', 'njtree', 'phylip'], 40 help="Return the aligned string if true. Otherwise, return the original sequence (no insertions)",
28 default='simple', 41 )
29 help='The format of a NH (New Hampshire) request') 42 parser.add_option(
43 "-c",
44 "--cigar_line",
45 type="choice",
46 choices=["0", "1"],
47 default="0",
48 help="Return the aligned sequence encoded in CIGAR format",
49 )
50 parser.add_option(
51 "--nh_format",
52 type="choice",
53 choices=[
54 "full",
55 "display_label_composite",
56 "simple",
57 "species",
58 "species_short_name",
59 "ncbi_taxon",
60 "ncbi_name",
61 "njtree",
62 "phylip",
63 ],
64 default="simple",
65 help="The format of a NH (New Hampshire) request",
66 )
30 options, args = parser.parse_args() 67 options, args = parser.parse_args()
31 if options.input is None: 68 if options.input is None:
32 raise Exception('-i option must be specified') 69 raise Exception("-i option must be specified")
33 70
34 server = 'https://rest.ensembl.org' 71 server = "https://rest.ensembl.org"
35 72
36 if options.id_type == 'gene_id': 73 if options.id_type == "gene_id":
37 ext = 'genetree/member/id' 74 ext = f"genetree/member/id/{options.species}/{options.input}"
38 elif options.id_type == 'gene_tree_id': 75 elif options.id_type == "gene_tree_id":
39 ext = 'genetree/id' 76 ext = f"genetree/id/{options.input}"
40 77
41 if options.format == 'json': 78 if options.format == "json":
42 content_type = 'application/json' 79 content_type = "application/json"
43 elif options.format == 'orthoxml': 80 elif options.format == "orthoxml":
44 content_type = 'text/x-orthoxml+xml' 81 content_type = "text/x-orthoxml+xml"
45 elif options.format == 'phyloxml': 82 elif options.format == "phyloxml":
46 content_type = 'text/x-phyloxml+xml' 83 content_type = "text/x-phyloxml+xml"
47 elif options.format == 'nh': 84 elif options.format == "nh":
48 content_type = 'text/x-nh' 85 content_type = "text/x-nh"
49 headers = {'Content-Type': content_type} 86 headers = {"Content-Type": content_type}
50 params = dict((k, getattr(options, k)) for k in ['sequence', 'aligned', 'cigar_line', 'nh_format']) 87 params = {
51 r = requests.get(urljoin(server, '/'.join([ext, options.input])), params=params, headers=headers) 88 k: getattr(options, k) for k in ("sequence", "aligned", "cigar_line", "nh_format")
89 }
90 r = requests.get(urljoin(server, ext), params=params, headers=headers)
52 91
53 if not r.ok: 92 if not r.ok:
54 r.raise_for_status() 93 r.raise_for_status()
55 94
56 print(r.text) 95 print(r.text)