async def _fetch_from_pubmed(self, doi=None, pmid=None, title=None):
result = {'doi': None, 'pmid': None, 'title': None, 'authors': None,
'journal': None, 'date': None}
if pmid is None:
url = URL('https://pubmed.ncbi.nlm.nih.gov/')
if doi is not None:
query = doi
elif title is not None:
query = title
else:
return result
encoded_query = 'term=' + urllib.parse.quote(f'"{query}"')
url._val = urllib.parse.SplitResult(url._val.scheme,
url._val.netloc,
url._val.path,
encoded_query,
url._val.fragment)
else:
url = URL('https://pubmed.ncbi.nlm.nih.gov/' + pmid)
retry = True
while retry:
async with self.session.get(url, ssl=sslcontext) as response:
if response.status == 200:
retry = False
soup = BeautifulSoup(await response.read(), 'lxml')
else:
print(response.status)
d_soup = soup.find('span', {'class': 'doi'})
if d_soup is None:
if title is None:
return result
# This means there was more than one search result,
# so we compare the titles of all of the search results
# to our title and pick the closest one that is more
# than 90% similar
d_soup = soup.find_all('a', {'class': 'labs-docsum-title'})
max_score = 0
max_link = ''
for d in d_soup:
try:
score = SequenceMatcher(None, title,
d.text.strip()).ratio()
except AttributeError:
return result
if score > max_score:
max_score = score
max_link = d['href']
if max_score > 0.9:
async with self.session.get('https://pubmed.ncbi.nlm.nih.gov' +
max_link,
ssl=sslcontext) as response:
soup = BeautifulSoup(await response.read(),
'lxml')
d_soup = soup.find('span', {'class': 'doi'})
else:
return result
try:
result['doi'] = d_soup.find('a').text.strip()
except AttributeError:
return result
if doi is not None:
assert doi.lower() == result['doi'].lower()
result['pmid'] = soup.find('strong',
{'class': 'current-id'}).text.strip()
if pmid is not None:
assert pmid == result['pmid']
cite = soup.find('span', {'class': 'cit'}).text
date = cite[:cite.find(';')]
result['date'] = self._parse_date(date).isoformat()
a_list = soup.find('div', {'class': 'authors-list'})
if a_list is not None:
authors = a_list.find_all('a', {'class': 'full-name'})
else:
authors = []
result['authors'] = []
for a in authors:
result['authors'].append(a.text.strip())
result['title'] = soup.find('h1',
{'class': 'heading-title'}).text.strip()
result['journal'] = soup.find(
'button', {'id': 'full-view-journal-trigger'})['title'].strip()
return result
