comparison aresite2/aresite2.py @ 3:755662977150 draft

Uploaded repo
author jfallmann
date Wed, 01 Feb 2017 09:59:29 -0500
parents
children
comparison
equal deleted inserted replaced
2:9b1beb18b477 3:755662977150
1 # A simple tool to connect to the AREsite server and retrieve feature
2 # information using the AREsite REST Interface.
3 # Parts of this code are from https://toolshed.g2.bx.psu.edu/repos/earlhaminst/ensembl_get_feature_info
4 from __future__ import print_function
5
6 import json
7 import optparse
8 import sys
9 import urllib
10 import urllib2
11 import time
12 import requests
13 from six.moves.urllib.parse import urljoin
14
15 usage = "usage: %prog [options] arg1 arg2"
16 parser = optparse.OptionParser(usage=usage)
17 parser.add_option('-g', '--gene', help='Gene ID to search for')
18 parser.add_option('-m', '--motif', help='Motif to look for', default='ATTTA', type=str)
19 parser.add_option('-s', '--species', type='choice',
20 choices=['Homo_sapiens', 'Mus_musculus', 'Danio_rerio', 'Drosophila_melanogaster', 'Caenorhabditis_elegans'], default='Homo_sapiens',
21 help='Specify the species to investigate')
22 options, args = parser.parse_args()
23
24 if options.gene is None:
25 raise Exception('- Specify the gene you want to look for!')
26
27 if "," in options.motif :
28 raise Exception('- Please only search for single motifs at once')
29
30 class AREsiteRestClient(object):
31 def __init__(self, server='http://rna.tbi.univie.ac.at/AREsite2/api/', reqs_per_sec=1):
32 self.server = server
33 self.reqs_per_sec = reqs_per_sec
34 self.req_count = 0
35 self.last_req = 0
36
37 def perform_rest_action(self, endpoint, hdrs=None, params=None):
38 if hdrs is None:
39 hdrs = {}
40
41 if 'Content-Type' not in hdrs:
42 hdrs['Content-Type'] = 'application/json'
43
44 if params:
45 endpoint += '?' + urllib.urlencode(params)
46
47 data = None
48
49 # check if we need to rate limit ourselves
50 if self.req_count >= self.reqs_per_sec:
51 delta = time.time() - self.last_req
52 if delta < 1:
53 time.sleep(1 - delta)
54 self.last_req = time.time()
55 self.req_count = 0
56
57 try:
58 request = urllib2.Request(self.server + endpoint, headers=hdrs)
59 response = urllib2.urlopen(request)
60 content = response.read()
61 if content:
62 data = json.loads(content)
63 self.req_count += 1
64
65 except urllib2.HTTPError, e:
66 # check if we are being rate limited by the server
67 if e.code == 429:
68 if 'Retry-After' in e.headers:
69 retry = e.headers['Retry-After']
70 time.sleep(float(retry))
71 self.perform_rest_action(endpoint, hdrs, params)
72 else:
73 sys.stderr.write('Request failed for {0}: Status code: {1.code} Reason: {1.reason}\n'.format(endpoint, e))
74
75 return data
76
77 def get_motifs(self, species, gene, motifs):
78 query = str('?query={0}&species={1}&list={2}'.format(gene, species, motifs))
79 if query:
80 aresite = self.perform_rest_action(
81 query
82 )
83 return aresite
84 return None
85
86 def run(species, gene, motifs):
87 client = AREsiteRestClient()
88 aresite = client.get_motifs(species, gene, motifs)
89 if aresite:
90
91 mots = aresite["exact_motifs"]
92 starts = aresite["motif_starts"]
93 ends = aresite["motif_ends"]
94 chrs = aresite["chromosomes"]
95 strands = aresite["strands"]
96 transcripts = aresite["transcripts"]
97 genes = aresite["genes"]
98 evh = aresite["hur_evidence"]
99 evt = aresite["ttp_evidence"]
100 eva = aresite["auf_evidence"]
101 anno = aresite["annotation"]
102
103 aresite = zip(chrs,starts,ends,mots,anno,strands,genes,transcripts,evh,evt,eva)
104
105 def getKey(item):
106 return item[1]
107
108 aresite = sorted(aresite, key=getKey)
109
110 # outfile = 'AREsite2_Rest_{0}_{1}_{2}.bed'.format(motifs,gene,species)
111 # f = open(outfile, 'w')
112
113 for i in range(len(aresite)):
114 # f.write("\t".join(aresite[i])+"\n")
115 print ("\t".join(aresite[i])+"\n")
116
117
118 if __name__ == '__main__':
119 run(options.species, options.gene, options.motif)