def login(self):
url = data_formats.urls['msigdb']['login1']
self.pre_session = dataio.curl(
url, init_url=url, silent=False, cache=False, init_headers=True)
url = data_formats.urls['msigdb']['login2']
post = {'j_username': self.user, 'j_password': '******'}
self.session = dataio.curl(
url,
init_url=url,
post=post,
req_headers=self.pre_session,
silent=False,
cache=False,
init_headers=True)
def load_set(self, setname, map_ids=True):
url = data_formats.urls['msigdb']['one_set'] % setname
data = dataio.curl(url, req_headers=self.session, silent=True)
data = data.split('\n')
self.info[setname] = data[1][2:]
self.write_set((j for j in (i.strip() for i in data[2:])
if len(j) > 0), setname, 'symbol', map_ids)
def load_collection(self,
collname,
id_type='entrez',
map_ids=True,
cachedir='cache'):
if os.path.exists(os.path.join(cachedir, 'gsea-%s.pickle' % collname)):
self.load([collname])
return None
url = self.collections[collname]['urls'][id_type]
data = dataio.curl(
url,
req_headers=self.session,
silent=False,
cache=False,
write_cache=True)
data = data.split('\n')
names = []
prg = progress.Progress(len(data), 'Loading gene sets', 1)
for line in (l.split('\t') for l in data if len(l) > 0):
prg.step()
setname = line[0].strip()
self.write_set(line[2:], setname, id_type, map_ids)
self.get_desc(setname)
names.append(setname)
prg.terminate()
self.groups[collname] = set(names)
self.save([collname], cachedir=cachedir)
def list_collections(self):
renm = re.compile(r'(.+)\([^0-9]*([0-9]*)[^0-9]*\)')
url = data_formats.urls['msigdb']['coll']
html = dataio.curl(url, req_headers=self.session, silent=False)
soup = bs4.BeautifulSoup(html, 'lxml')
for col in soup.find('table', class_='lists1').find_all('tr'):
lname, num = renm.findall(col.find('th').text.replace('\n', ''))[0]
sname = col.find('a').attrs['name']
urls = dict(
[(d.attrs['href'].split('.')[-2],
data_formats.urls['msigdb']['url_stem'] % d.attrs['href'])
for d in col.find_all('a')[-3:]])
self.collections[sname] = {
'name': lname,
'count': int(num),
'urls': urls
}
def get_desc(self, setname):
url = data_formats.urls['msigdb']['one_set'] % setname
txt = dataio.curl(url, req_headers=self.session, silent=True)
self.info[setname] = txt.split('\n')[1][2:]