def report(query, label, refresh=True):
"""Print out an org-mode report for search results.
Parameters
----------
query : str
The search query based on which results the report should be
generated.
label : str
The label used in the document title ("Report for ...").
refresh : bool (optional, default=True)
Whether to refresh a cached file containing results of a previous
query or not.
"""
# Header
print('*** Report for {}\n'.format(label))
print('#+attr_latex: :placement [H] :center nil')
# Perform query
s = ScopusSearch(query, refresh=refresh)
journal_res = [p for p in s.results if p.aggregationType == "Journal"]
# Parse results
doc_types = Counter([p.aggregationType for p in s.results])
paper_cites = {(p.title, p.doi): int(p.citedby_count) for p in journal_res}
Ncites = sum(paper_cites.values())
papers = len(journal_res)
author_count = [len(p.authid.split(";")) for p in journal_res]
au_counts = defaultdict(lambda: 0)
j_counts = defaultdict(lambda: 0)
for p in journal_res:
for auth in zip(p.authname.split(";"), p.authid.split(";")):
key = (auth[0], auth[1])
au_counts[key] += 1
jkey = (p.publicationName, p.source_id, p.issn)
j_counts[jkey] += 1
# Document information
print('#+caption: Types of documents found for {}.'.format(label))
print('| Document type | count |\n|-')
for key, value in doc_types.items():
print('| {} | {} |'.format(key, value))
print('\n\n{} articles ({} citations) '
'found by {} authors'.format(papers, Ncites, len(au_counts)))
# Author counts {(name, scopus-id): count}
auth_url = "[[https://www.scopus.com/authid/detail.uri?authorId={}][{}]]"
view = [(auth_url.format(k[1], k[0]), v, k[1])
for k, v in au_counts.items()]
view.sort(reverse=True, key=itemgetter(1))
print('\n#+attr_latex: :placement [H] :center nil')
print('#+caption: Author publication counts for {0}.'.format(label))
print('| name | count | categories |\n|-')
for name, count, identifier in view[:20]:
cats = ', '.join([
'{} ({})'.format(cat[0], cat[1])
for cat in get_subject_docs(identifier, refresh)[0:3]
])
print('| {} | {} | {} |'.format(name, count, cats))
# Journal information
jour_url = '[[https://www.scopus.com/source/sourceInfo.url?sourceId={}][{}]]'
jview = [(jour_url.format(k[1], k[0][0:50]), k[1], k[2], v)
for k, v in j_counts.items()]
jview.sort(reverse=True, key=itemgetter(3))
print('\n\n#+attr_latex: :placement [H] :center nil')
print('#+caption: Journal publication counts for {}.'.format(label))
print('| Journal | count |\n|-')
for journal, sid, issn, count in jview[0:12]:
print('| {} | {} |'.format(journal, count))
# Top cited papers
pview = [('[[{}][{}]]'.format(k[1], k[0][0:60]), int(v))
for k, v in paper_cites.items()]
pview.sort(reverse=True, key=itemgetter(1))
h_index = hindex([p[1] for p in pview])
print('\n\n#+attr_latex: :placement [H] :center nil')
print('#+caption: Top cited publication'
' counts for {}. h-index = {}.'.format(label, h_index))
print('| title | cite count |\n|-')
for title, count in pview[0:10]:
print('| {} | {} |'.format(title, count))
# Plot authors per publication
plt.figure()
plt.hist(author_count, 20)
plt.xlabel('# authors')
plt.ylabel('frequency')
plt.savefig('{}-nauthors-per-publication.png'.format(label))
# Bibliography
print('\n\n#+caption: Number of authors '
'on each publication for {}.'.format(label))
print('[[./{}-nauthors-per-publication.png]]'.format(label))
print('''**** Bibliography :noexport:
:PROPERTIES:
:VISIBILITY: folded
:END:''')
for i, p in enumerate(journal_res):
abstract = AbstractRetrieval(p.eid)
print('{}. {}\n'.format(i + 1, abstract))
def calculate_h_index(df_authors, db_path, has_key=False):
# Collect three stats: (i) author name and his/her h-index, (ii) citation list of each pmid, and (iii) author pmids
author_2_hindex = dict()
author_2_hindex_return = dict()
pmid_2_cite = dict()
author_2_pmids = dict()
h_index_db = dict()
with open(db_path) as h_index_db_file:
h_index_db = json.load(h_index_db_file)
for name in tqdm(df_authors):
if name[0] == ' ':
author = name[1:] # + ' ' + surname
else:
author = name # + ' ' + surname
author_pmids = []
# First check if the name already exists in our db, if 0, then try again
call_api_flag = True
if author in h_index_db['h_indices']:
h_index = h_index_db['h_indices'][author]
# print(f"{author} found in current database ({db_path}) with a value of {h_index}")
if h_index != 0:
# print(f"{author} found in current database ({db_path}) with a value of {h_index}")
author_2_hindex_return[author] = h_index
call_api_flag = False
else:
# print(f"{author} found in current database ({db_path}) with a value of {h_index} ... retrying API call")
# BELOW IS JUST FOR DEMO!!!
#TODO
author_2_hindex_return[author] = h_index
call_api_flag = False
else:
print(f"{author} not found in current database ({db_path})")
# This ensures that we are not checking short and very common names which takes forever to collect information
if len(author) > 5 and call_api_flag:
if has_key:
page = pull_url(author, Entrez.email, Entrez.api_key)
else:
page = pull_url(author, Entrez.email)
soup = BeautifulSoup(page.content, 'xml')
ids = soup.find_all('Id', {})
for id_ in ids:
author_pmids.append(id_.get_text())
author_2_pmids[author] = author_pmids
citations = []
retrieved = get_links_ids([
int(pmid) for pmid in author_pmids
if int(pmid) not in pmid_2_cite.keys()
])
for pmid in retrieved.keys():
link_list = []
if pmid in retrieved:
link_list = retrieved[pmid]
pmid_2_cite[pmid] = link_list
citations.append(len(link_list))
author_2_hindex[author] = int(hindex(citations))
author_2_hindex_return[author] = author_2_hindex[author]
elif call_api_flag:
author_2_hindex[author] = -1
author_2_hindex_return[author] = -1
h_index_db['h_indices'].update(author_2_hindex)
h_index_db['pmids'].update(author_2_pmids)
h_index_db['citations'].update(pmid_2_cite)
#TODO: make changes to path to overwrite current file
with open(
os.path.join(os.path.dirname(__file__),
f'./data/author_h_indexes.json'), 'w') as output:
output.write(
json.dumps({
'h_indices': h_index_db['h_indices'],
'pmids': h_index_db['pmids'],
'citations': h_index_db['citations']
}))
return author_2_hindex_return
def test_hindex_with_nan(self):
citations = [6, 10, 5, 46, np.nan, 2]
received = hindex(citations)
expected = 4
self.assertEqual(received, expected)
def test_hindex_with_only_nan(self):
citations = [np.nan, np.nan]
received = hindex(citations, ignore_nan=False)
self.assertTrue(np.isnan(received))
def test_hindex(self):
citations = [6, 10, 5, 46, 0, 2]
received = hindex(citations)
expected = 4
self.assertEqual(received, expected)
def calculateHIndex(self):
authors_with_hIndex = 0
paperCitations = []
for eachCitation in self.citationArticles:
paperCitations.append(eachCitation["Citations"])
return hindex(paperCitations)
def test_hindex_with_only_nan(self):
citations = [np.nan, np.nan]
received = hindex(citations, ignore_nan=False)
self.assertTrue(np.isnan(received))
def test_hindex_with_nan(self):
citations = [6, 10, 5, 46, np.nan, 2]
received = hindex(citations)
expected = 4
self.assertEqual(received, expected)
def test_hindex(self):
citations = [6, 10, 5, 46, 0, 2]
received = hindex(citations)
expected = 4
self.assertEqual(received, expected)