def resolve(self, citations, document = None):
citation = {}
if not utopia.citation.has_link(citations, {'mime': 'application/pdf'}, {'whence': 'pmc'}):
# Try to resolve the PMC ID from either the DOI or the PubMed ID
pmcid = utopia.citation.pick_from(citations, 'identifiers/pmc', default=None)
if pmcid is None:
doi = utopia.citation.pick_from(citations, 'identifiers/doi', default=None, record_in=citation)
pmid = utopia.citation.pick_from(citations, 'identifiers/pubmed', default=None, record_in=citation)
if doi is not None and pmcid is None:
pmcid = utopia.tools.pmc.identify(doi, 'doi')
if pmid is not None and pmcid is None:
pmcid = utopia.tools.pmc.identify(pmid, 'pmid')
# Generate PMC link to PDF
if pmcid is not None:
pdf_url = 'http://www.ncbi.nlm.nih.gov/pmc/articles/{0}/pdf/'.format(pmcid)
citation.update({
'links': [{
'url': pdf_url,
'mime': 'application/pdf',
'type': 'article',
'title': 'Download article from PubMed Central',
}],
'identifiers': {'pmc': pmcid}
})
return citation
def resolve(self, citations, document = None):
# If an ArXiv ID is present, look it up
citation = {}
arxiv_id = utopia.citation.pick_from(citations, 'identifiers[arxiv]', None, record_in=citation)
if arxiv_id is not None:
citation.update(utopia.tools.arxiv.resolve(arxiv_id))
return citation
def resolve(self, citations, document=None):
# If a DOI is present, look it up
citation = {}
doi = utopia.citation.pick_from(citations,
'identifiers[doi]',
default=None,
record_in=citation)
if doi is not None:
citation.update(utopia.tools.crossref.resolve(doi))
return citation
def resolve(self, citations, document=None):
# If a PubMed ID is present, look it up
citation = {}
pmid = utopia.citation.pick_from(citations,
'identifiers[pubmed]',
default=None,
record_in=citation)
title = utopia.citation.pick_from(citations,
'title:pubmed',
default=None)
if title is None and pmid is not None:
citation.update(utopia.tools.pubmed.resolve(pmid))
return citation
def resolve(self, citations, document=None):
citation = {}
pubmed_id = utopia.citation.pick_from(citations,
'identifiers[pubmed]',
None,
record_in=citation)
if pubmed_id is None:
doi = utopia.citation.pick_from(citations,
'identifiers[doi]',
None,
record_in=citation)
if doi is not None:
pubmed_id = utopia.tools.pubmed.identify(doi, 'doi')
if pubmed_id is not None:
citation['identifiers'] = {'pubmed': pubmed_id}
if pubmed_id is None:
title = utopia.citation.pick_from(citations,
'title',
None,
record_in=citation)
if title is not None:
title = title.strip(' .')
pubmed_results = utopia.tools.pubmed.search(title)
pubmed_title = pubmed_results.get('title', '').strip(' .')
if len(pubmed_title) > 0:
matched = False
pubmed_pmid = pubmed_results.get('identifiers',
{}).get('pubmed')
if re.sub(r'[^\w]+', ' ',
title).strip().lower() == re.sub(
r'[^\w]+', ' ', pubmed_title).strip(
).lower(): # Fuzzy match
matched = True
elif document is not None:
# Accept the pubmed title over the scraped title, if present in the document
matches = document.findInContext(
'', pubmed_title, '') # Fuzzy match
if len(matches) > 0:
matched = True
pubmed_title = matches[0].text()
if matched:
citation.update(pubmed_results)
citation['title'] = pubmed_title
return citation
def resolve(self, citations, document=None):
# Multiple responses leads to a no-op
for citation in citations:
if utopia.citation.pick(citation,
'provenance/whence',
default=None) == 'cermine':
# Bail if cermine results are already present
return None
# Get all the citations that don't look structured
structure_keys = set(['title', 'authors', 'year'])
citation = {}
unstructured = utopia.citation.pick_from(citations,
'unstructured',
default=None,
record_in=citation)
if unstructured is not None and len(
structure_keys & set(unstructured.citation.keys())) == 0:
structured = utopia.tools.cited.parse(unstructured)
if len(structured) > 0:
citation.update(structured[0])
return citation
def resolve(self, citations, document=None):
citation = {}
doi = utopia.citation.pick_from(citations,
'identifiers[doi]',
default=None,
record_in=citation)
title = utopia.citation.pick_from(citations,
'title',
default=None,
record_in=citation)
if doi is not None or title is not None:
if doi is None:
xref_results = utopia.tools.crossref.search(title)
if len(xref_results) == 1:
best = xref_results[0]
xref_title = best.get('title', '').strip(' .')
if len(xref_title) > 0:
matched = False
if document is not None and spineapi is not None:
# Accept the crossref title if present in the document (do magic dash pattern thing)
xref_title = re.sub(
ur'[^-\u002D\u007E\u00AD\u058A\u05BE\u1400\u1806\u2010-\u2015\u2053\u207B\u208B\u2212\u2E17\u2E3A\u2E3B\u301C\u3030\u30A0\uFE31\uFE32\uFE58\uFE63\uFF0D]+',
lambda x: re.escape(x.group(0)), xref_title)
xref_title = re.sub(
ur'[\u002D\u007E\u00AD\u058A\u05BE\u1400\u1806\u2010-\u2015\u2053\u207B\u208B\u2212\u2E17\u2E3A\u2E3B\u301C\u3030\u30A0\uFE31\uFE32\uFE58\uFE63\uFF0D-]+',
lambda x: r'\p{{Pd}}{{{0}}}'.format(
len(x.group(0))), xref_title)
matches = document.search(
xref_title,
spineapi.RegExp + spineapi.IgnoreCase)
matched = (len(matches) > 0)
else:
matched = (xref_title.lower() == title)
if matched:
citation.update(best)
doi = citation.get('identifiers', {}).get('doi')
if doi is not None and doi.startswith(
'http://dx.doi.org/'):
doi = doi[18:]
citation['identifiers']['doi'] = doi
if doi is not None:
if None not in (document, title):
# What is this DOI's article's title according to crossref?
try:
xref_results = utopia.tools.crossref.resolve(doi)
xref_title = xref_results.get('title', '')
if len(xref_title) > 0:
print 'crossref: resolved title:', xref_title.encode(
'utf8')
if re.sub(r'[^\w]+', ' ', title).strip() == re.sub(
r'[^\w]+', ' ',
xref_title).strip(): # Fuzzy match
print 'crossref: titles match precisely'
citation.update(xref_results)
else:
# Accept the crossref title over the scraped title, if present in the document
matches = document.findInContext(
'', xref_title, '') # Fuzzy match
if len(matches) > 0:
citation.update(xref_results)
print 'crossref: overriding scraped title with crossref title'
else:
print 'crossref: ignoring resolved citations'
# FIXME should we discard the DOI at this point?
except Exception as e:
import traceback
traceback.print_exc()
return citation