def new_entity(self, uri, label=None, now=None):
if not now:
# Get a UTC timestamp in ISO 8601 format
now = datetime.utcnow().isoformat()
uri = unicode(uri)
if uri.startswith('http://'):
resource_uri = URIRef(uri)
else:
resource_uri = LU[get_qname(uri)]
ascii_label = label.encode('ascii', 'ignore') if label else 'undefined'
resource_entity_uri = LU[urllib.quote_plus(u"entity/{}_{}".format(
ascii_label, now),
safe="/")]
self.graph.add((resource_uri, RDFS['label'], Literal(label)))
self.graph.add((resource_entity_uri, RDF.type, PROV['Entity']))
if label:
self.graph.add(
(resource_entity_uri, RDFS['label'], Literal(label)))
else:
self.graph.add((resource_entity_uri, RDFS['label'], Literal(uri)))
self.graph.add(
(resource_entity_uri, PROV['wasDerivedFrom'], resource_uri))
self.graph.add((resource_entity_uri, PROV['wasGeneratedAt'],
Literal(now, datatype=XSD['dateTime'])))
return resource_entity_uri
def new_entity(self, uri, label=None, now=None):
if not now:
# Get a UTC timestamp in ISO 8601 format
now = datetime.utcnow().isoformat()
uri = unicode(uri)
if uri.startswith('http://') :
resource_uri = URIRef(uri)
else :
resource_uri = LU[get_qname(uri)]
ascii_label = label.encode('ascii', 'ignore') if label else 'undefined'
resource_entity_uri = LU[urllib.quote_plus(u"entity/{}_{}".format(ascii_label, now), safe="/")]
self.graph.add((resource_uri, RDFS['label'], Literal(label)))
self.graph.add((resource_entity_uri, RDF.type, PROV['Entity']))
if label:
self.graph.add((resource_entity_uri, RDFS['label'], Literal(label)))
else :
self.graph.add((resource_entity_uri, RDFS['label'], Literal(uri)))
self.graph.add((resource_entity_uri, PROV['wasDerivedFrom'], resource_uri))
self.graph.add((resource_entity_uri, PROV['wasGeneratedAt'], Literal(now, datatype=XSD['dateTime'])))
return resource_entity_uri
def get_rdf(nanopub_id, article, urls, provenance_trail):
"""Takes everything we know about the article specified in article_id, and builds a simple RDF graph.
We only consider the URLs of checkboxes that were selected by the user.
Returns the RDF graph as a ConjunctiveGraph"""
article_id = article['article_id']
i = article
article_id_qname = get_qname(article_id)
nanopub_id_qname = get_qname(nanopub_id)
nano = LU["nanopublication/{}".format(article_id_qname)]
assertion = LU["assertion/{}".format(article_id_qname)]
provenance = LU["provenance/{}".format(article_id_qname)]
# We don't provide any 'supporting' information for the nanopublication
store = plugin.get('IOMemory', Store)()
np_graph = Graph(store, identifier=nano)
a_graph = Graph(store, identifier=assertion)
p_graph = Graph(store, identifier=provenance)
np_graph = associate_namespaces(np_graph)
a_graph = associate_namespaces(a_graph)
p_graph = associate_namespaces(p_graph)
p_graph += trail_to_prov(provenance_trail)
# A bit annoying, but we need both the DOI and the Owner before we can start
if 'doi' in i:
# If the article has a DOI assigned, we will use it in our RDF rendering
doi = i['doi']
article_uri = URIRef(doi)
else :
# If it doesn't, we'll gues the DOI URI of our own making
doi = "http://dx.doi.org/10.6084/m9.figshare.{}".format(article_id)
article_uri = URIRef(doi)
a_graph.add((article_uri,OWL.sameAs,LU[article_id_qname]))
a_graph.add((article_uri,LUV['doi'],URIRef(doi)))
# print "Processing owner..."
if 'owner' in i:
owner = i['owner']
o_id = owner['id']
o_label = owner['full_name']
o_qname = get_qname(o_id)
owner_uri = LU[o_qname]
a_graph.add((article_uri,LUV['owner'],owner_uri))
a_graph.add((LU[o_qname],FOAF['name'],Literal(o_label)))
a_graph.add((LU[o_qname],LUV['id'],Literal(o_id)))
a_graph.add((LU[o_qname],RDF.type,LUV['Owner']))
else :
owner_uri = None
# Add the stuff necessary to define the nanopublication
np_graph.add((nano, RDF.type, NANOPUB['Nanopublication']))
np_graph.add((nano, NANOPUB['hasAssertion'], assertion))
np_graph.add((nano, NANOPUB['hasProvenance'], provenance))
nanopub_doi = "http://dx.doi.org/10.6084/m9.figshare.{}".format(nanopub_id)
nanopub_uri = URIRef(nanopub_doi)
np_graph.add((nanopub_uri, OWL.sameAs, LU[nanopub_id_qname]))
np_graph.add((nano, RDFS['seeAlso'], nanopub_uri))
now = datetime.now()
nowstr = datetime.now().strftime("%Y%m%dT%H%M%S%z")
# Add the necessary provenance information
user_qname = re.sub(' ','_', g.user.nickname)
user_uri = LU['person/{}'.format(user_qname)]
activity_uri = LU['linkitup_{}'.format(nowstr)]
p_graph.add((nano, RDF.type, PROV['Entity']))
p_graph.add((nano, DCTERMS['license'], URIRef('http://creativecommons.org/publicdomain/zero/1.0/')))
p_graph.add((nano, PROV['wasGeneratedBy'], activity_uri))
p_graph.add((nano, PROV['wasGeneratedAt'], Literal(now)))
p_graph.add((nano, PROV['wasAttributedTo'], user_uri))
p_graph.add((activity_uri, RDF.type, PROV['Activity']))
p_graph.add((activity_uri, PROV['used'], article_uri))
p_graph.add((activity_uri, PROV['generated'], nano))
p_graph.add((activity_uri, PROV['endedAtTime'], Literal(now)))
p_graph.add((activity_uri, PROV['wasStartedBy'], user_uri))
p_graph.add((activity_uri, PROV['wasInfluencedBy'], URIRef('http://linkitup.data2semantics.org')))
p_graph.add((user_uri, RDF.type, PROV['Person']))
p_graph.add((user_uri, RDF.type, PROV['Agent']))
p_graph.add((user_uri, RDF.type, FOAF['Person']))
p_graph.add((user_uri, FOAF['name'], Literal(g.user.nickname)))
p_graph.add((URIRef('http://linkitup.data2semantics.org'), RDF.type, PROV['SoftwareAgent']))
p_graph.add((URIRef('http://linkitup.data2semantics.org'), RDF.type, PROV['Agent']))
p_graph.add((URIRef('http://linkitup.data2semantics.org'), FOAF['name'], Literal('Linkitup')))
p_graph.add((article_uri, RDF.type, PROV['Entity']))
if owner_uri :
p_graph.add((article_uri, PROV['wasAttributedTo'], owner_uri))
# Add the stuff necessary to define the Open Annotation annotation
p_graph.add((nano, RDF.type, OA['Annotation']))
p_graph.add((nano, OA['hasBody'], assertion))
p_graph.add((nano, OA['hasTarget'], article_uri))
p_graph.add((nano, OA['wasAnnotatedBy'], user_uri))
p_graph.add((nano, OA['wasAnnotatedAt'], Literal(now)))
a_graph.add((article_uri,LUV['article_id'],Literal(article_id)))
# print "Processing defined type"
dt = i['defined_type']
o_qname = get_qname(quote(dt))
a_graph.add((article_uri,LUV['defined_type'],LU[o_qname]))
a_graph.add((LU[o_qname],SKOS.prefLabel,Literal(dt)))
a_graph.add((LU[o_qname],RDF.type,LUV['DefinedType']))
# print "Processing published date"
date = i['published_date']
pydate = datetime.strptime(date,'%H:%M, %b %d, %Y')
a_graph.add((article_uri,LUV['published_date'],Literal(pydate)))
# print "Processing description"
description = i['description']
a_graph.add((article_uri,SKOS.description, Literal(description)))
if len(i['authors']) > 0 :
# print "Processing authors..."
author_count = 0
seq = BNode()
a_graph.add((article_uri,LUV['authors'],seq))
a_graph.add((seq,RDF.type,RDF.Seq))
for author in i['authors'] :
a_id = author['id']
a_label = author['full_name'].strip()
a_first = author['first_name'].strip()
a_last = author['last_name'].strip()
a_qname = get_qname(a_id)
author_count = author_count + 1
member = URIRef('http://www.w3.org/1999/02/22-rdf-syntax-ns#_{}'.format(author_count))
a_graph.add((seq,member,LU[a_qname]))
a_graph.add((LU[a_qname],FOAF['name'],Literal(a_label)))
a_graph.add((LU[a_qname],FOAF['firstName'],Literal(a_first)))
a_graph.add((LU[a_qname],FOAF['lastName'],Literal(a_last)))
a_graph.add((LU[a_qname],LUV['id'],Literal(a_id)))
a_graph.add((LU[a_qname],RDF.type,LUV['Author']))
a_graph.add((LU[a_qname],RDF.type,FOAF['Person']))
# print "Processing tags..."
for tag in i['tags'] :
# print tag
t_id = tag['id']
t_label = tag['name']
t_qname = get_qname(t_id)
a_graph.add((article_uri,LUV['tag'],LU[t_qname]))
a_graph.add((LU[t_qname],SKOS.prefLabel,Literal(t_label)))
a_graph.add((LU[t_qname],LUV['id'],Literal(t_id)))
a_graph.add((LU[t_qname],RDF.type,LUV['Tag']))
# print "Processing links..."
for link in i['links'] :
# print link
l_id = link['id']
l_value = link['link']
l_qname = get_qname(l_id)
a_graph.add((article_uri,LUV['link'],LU[l_qname]))
a_graph.add((LU[l_qname],LUV['id'],Literal(l_id)))
a_graph.add((LU[l_qname],RDFS.seeAlso,URIRef(l_value)))
a_graph.add((LU[l_qname],FOAF['page'],URIRef(l_value)))
a_graph.add((LU[l_qname],RDF.type,LUV['Link']))
# print "Checking if link matches a Wikipedia/DBPedia page..."
if l_value.startswith('http://en.wikipedia.org/wiki/') :
l_match = re.sub('http://en.wikipedia.org/wiki/','http://dbpedia.org/resource/',l_value)
a_graph.add((LU[l_qname],SKOS.exactMatch,URIRef(l_match)))
# print "Processing files..."
for f in i['files'] :
# print f
f_id = f['id']
f_value = f['name']
f_mime = f['mime_type']
f_size = f['size']
f_qname = get_qname(f_id)
a_graph.add((article_uri,LUV['file'],LU[f_qname]))
a_graph.add((LU[f_qname],LUV['id'],Literal(f_id)))
a_graph.add((LU[f_qname],RDFS.label,Literal(f_value)))
a_graph.add((LU[f_qname],LUV['mime_type'],Literal(f_mime)))
a_graph.add((LU[f_qname],LUV['size'],Literal(f_size)))
a_graph.add((LU[f_qname],RDF.type,LUV['File']))
# print "Processing categories..."
for cat in i['categories'] :
# print cat
c_id = cat['id']
c_value = cat['name']
c_qname = get_qname(c_id)
a_graph.add((article_uri,LUV['category'],LU[c_qname]))
a_graph.add((LU[c_qname],LUV['id'],Literal(c_id)))
a_graph.add((LU[c_qname],RDFS.label,Literal(c_value)))
a_graph.add((LU[c_qname],RDF.type,LUV['Category']))
for k,u in urls.items() :
original_qname = get_qname(u['original'])
uri = u['uri']
if u['type'] == 'mapping':
a_graph.add((LU[original_qname],SKOS.exactMatch,URIRef(uri) ))
elif u['type'] == 'reference':
a_graph.add((LU[original_qname],DCTERMS['references'],URIRef(uri) ))
elif u['type'] == 'link' :
a_graph.add((LU[original_qname],SKOS.related, URIRef(uri)))
else :
a_graph.add((LU[original_qname],SKOS.related, URIRef(uri)))
graph = ConjunctiveGraph(store)
# out = ""
# for s, p, o, gr in graph.quads((None, None, None)) :
# out += "{} > {} {} {}\n".format(gr.identifier, s, p, o)
#
# app.logger.debug(out)
return graph
def get_rdf(nanopub_id, article, urls, provenance_trail):
"""Takes everything we know about the article specified in article_id, and builds a simple RDF graph.
We only consider the URLs of checkboxes that were selected by the user.
Returns the RDF graph as a ConjunctiveGraph"""
article_id = article['article_id']
i = article
article_id_qname = get_qname(article_id)
nanopub_id_qname = get_qname(nanopub_id)
nano = LU["nanopublication/{}".format(article_id_qname)]
assertion = LU["assertion/{}".format(article_id_qname)]
provenance = LU["provenance/{}".format(article_id_qname)]
# We don't provide any 'supporting' information for the nanopublication
store = plugin.get('IOMemory', Store)()
np_graph = Graph(store, identifier=nano)
a_graph = Graph(store, identifier=assertion)
p_graph = Graph(store, identifier=provenance)
np_graph = associate_namespaces(np_graph)
a_graph = associate_namespaces(a_graph)
p_graph = associate_namespaces(p_graph)
p_graph += trail_to_prov(provenance_trail)
# A bit annoying, but we need both the DOI and the Owner before we can start
if 'doi' in i:
# If the article has a DOI assigned, we will use it in our RDF rendering
doi = i['doi']
article_uri = URIRef(doi)
else :
# If it doesn't, we'll gues the DOI URI of our own making
doi = "http://dx.doi.org/10.6084/m9.figshare.{}".format(article_id)
article_uri = URIRef(doi)
a_graph.add((article_uri,OWL.sameAs,LU[article_id_qname]))
a_graph.add((article_uri,LUV['doi'],URIRef(doi)))
# print "Processing owner..."
if 'owner' in i:
owner = i['owner']
o_id = owner['id']
o_label = owner['full_name']
o_qname = get_qname(o_id)
owner_uri = LU[o_qname]
a_graph.add((article_uri,LUV['owner'],owner_uri))
a_graph.add((LU[o_qname],FOAF['name'],Literal(o_label)))
a_graph.add((LU[o_qname],LUV['id'],Literal(o_id)))
a_graph.add((LU[o_qname],RDF.type,LUV['Owner']))
else :
owner_uri = None
# Add the stuff necessary to define the nanopublication
np_graph.add((nano, RDF.type, NANOPUB['Nanopublication']))
np_graph.add((nano, NANOPUB['hasAssertion'], assertion))
np_graph.add((nano, NANOPUB['hasProvenance'], provenance))
nanopub_doi = "http://dx.doi.org/10.6084/m9.figshare.{}".format(nanopub_id)
nanopub_uri = URIRef(nanopub_doi)
np_graph.add((nanopub_uri, OWL.sameAs, LU[nanopub_id_qname]))
np_graph.add((nano, RDFS['seeAlso'], nanopub_uri))
now = datetime.now()
nowstr = datetime.now().strftime("%Y%m%dT%H%M%S%z")
# Add the necessary provenance information
user_qname = re.sub(' ','_', g.user.nickname)
user_uri = LU['person/{}'.format(user_qname)]
activity_uri = LU['linkitup_{}'.format(nowstr)]
p_graph.add((nano, RDF.type, PROV['Entity']))
p_graph.add((nano, DCTERMS['license'], URIRef('http://creativecommons.org/publicdomain/zero/1.0/')))
p_graph.add((nano, PROV['wasGeneratedBy'], activity_uri))
p_graph.add((nano, PROV['wasGeneratedAt'], Literal(now)))
p_graph.add((nano, PROV['wasAttributedTo'], user_uri))
p_graph.add((activity_uri, RDF.type, PROV['Activity']))
p_graph.add((activity_uri, PROV['used'], article_uri))
p_graph.add((activity_uri, PROV['generated'], nano))
p_graph.add((activity_uri, PROV['endedAtTime'], Literal(now)))
p_graph.add((activity_uri, PROV['wasStartedBy'], user_uri))
p_graph.add((activity_uri, PROV['wasInfluencedBy'], URIRef('http://linkitup.data2semantics.org')))
p_graph.add((user_uri, RDF.type, PROV['Person']))
p_graph.add((user_uri, RDF.type, PROV['Agent']))
p_graph.add((user_uri, RDF.type, FOAF['Person']))
p_graph.add((user_uri, FOAF['name'], Literal(g.user.nickname)))
p_graph.add((URIRef('http://linkitup.data2semantics.org'), RDF.type, PROV['SoftwareAgent']))
p_graph.add((URIRef('http://linkitup.data2semantics.org'), RDF.type, PROV['Agent']))
p_graph.add((URIRef('http://linkitup.data2semantics.org'), FOAF['name'], Literal('Linkitup')))
p_graph.add((article_uri, RDF.type, PROV['Entity']))
if owner_uri :
p_graph.add((article_uri, PROV['wasAttributedTo'], owner_uri))
# Add the stuff necessary to define the Open Annotation annotation
p_graph.add((nano, RDF.type, OA['Annotation']))
p_graph.add((nano, OA['hasBody'], assertion))
p_graph.add((nano, OA['hasTarget'], article_uri))
p_graph.add((nano, OA['wasAnnotatedBy'], user_uri))
p_graph.add((nano, OA['wasAnnotatedAt'], Literal(now)))
a_graph.add((article_uri,LUV['article_id'],Literal(article_id)))
# print "Processing defined type"
dt = i['defined_type']
o_qname = get_qname(quote(dt))
a_graph.add((article_uri,LUV['defined_type'],LU[o_qname]))
a_graph.add((LU[o_qname],SKOS.prefLabel,Literal(dt)))
a_graph.add((LU[o_qname],RDF.type,LUV['DefinedType']))
# print "Processing published date"
date = i['published_date']
pydate = datetime.strptime(date,'%H:%M, %b %d, %Y')
a_graph.add((article_uri,LUV['published_date'],Literal(pydate)))
# print "Processing description"
description = i['description']
a_graph.add((article_uri,SKOS.description, Literal(description)))
if len(i['authors']) > 0 :
# print "Processing authors..."
author_count = 0
seq = BNode()
a_graph.add((article_uri,LUV['authors'],seq))
a_graph.add((seq,RDF.type,RDF.Seq))
for author in i['authors'] :
a_id = author['id']
a_label = author['full_name'].strip()
a_first = author['first_name'].strip()
a_last = author['last_name'].strip()
a_qname = get_qname(a_id)
author_count = author_count + 1
member = URIRef('http://www.w3.org/1999/02/22-rdf-syntax-ns#_{}'.format(author_count))
a_graph.add((seq,member,LU[a_qname]))
a_graph.add((LU[a_qname],FOAF['name'],Literal(a_label)))
a_graph.add((LU[a_qname],FOAF['firstName'],Literal(a_first)))
a_graph.add((LU[a_qname],FOAF['lastName'],Literal(a_last)))
a_graph.add((LU[a_qname],LUV['id'],Literal(a_id)))
a_graph.add((LU[a_qname],RDF.type,LUV['Author']))
a_graph.add((LU[a_qname],RDF.type,FOAF['Person']))
# print "Processing tags..."
for tag in i['tags'] :
# print tag
t_id = tag['id']
t_label = tag['name']
t_qname = get_qname(t_id)
a_graph.add((article_uri,LUV['tag'],LU[t_qname]))
a_graph.add((LU[t_qname],SKOS.prefLabel,Literal(t_label)))
a_graph.add((LU[t_qname],LUV['id'],Literal(t_id)))
a_graph.add((LU[t_qname],RDF.type,LUV['Tag']))
# print "Processing links..."
for link in i['links'] :
# print link
l_id = link['id']
l_value = link['link']
l_qname = get_qname(l_id)
a_graph.add((article_uri,LUV['link'],LU[l_qname]))
a_graph.add((LU[l_qname],LUV['id'],Literal(l_id)))
a_graph.add((LU[l_qname],RDFS.seeAlso,URIRef(l_value)))
a_graph.add((LU[l_qname],FOAF['page'],URIRef(l_value)))
a_graph.add((LU[l_qname],RDF.type,LUV['Link']))
# print "Checking if link matches a Wikipedia/DBPedia page..."
if l_value.startswith('http://en.wikipedia.org/wiki/') :
l_match = re.sub('http://en.wikipedia.org/wiki/','http://dbpedia.org/resource/',l_value)
a_graph.add((LU[l_qname],SKOS.exactMatch,URIRef(l_match)))
# print "Processing files..."
for f in i['files'] :
# print f
f_id = f['id']
f_value = f['name']
f_mime = f['mime_type']
f_size = f['size']
f_qname = get_qname(f_id)
a_graph.add((article_uri,LUV['file'],LU[f_qname]))
a_graph.add((LU[f_qname],LUV['id'],Literal(f_id)))
a_graph.add((LU[f_qname],RDFS.label,Literal(f_value)))
a_graph.add((LU[f_qname],LUV['mime_type'],Literal(f_mime)))
a_graph.add((LU[f_qname],LUV['size'],Literal(f_size)))
a_graph.add((LU[f_qname],RDF.type,LUV['File']))
# print "Processing categories..."
for cat in i['categories'] :
# print cat
c_id = cat['id']
c_value = cat['name']
c_qname = get_qname(c_id)
a_graph.add((article_uri,LUV['category'],LU[c_qname]))
a_graph.add((LU[c_qname],LUV['id'],Literal(c_id)))
a_graph.add((LU[c_qname],RDFS.label,Literal(c_value)))
a_graph.add((LU[c_qname],RDF.type,LUV['Category']))
for k,u in urls.items() :
original_qname = get_qname(u['original'])
uri = u['uri']
if u['type'] == 'mapping':
a_graph.add((LU[original_qname],SKOS.exactMatch,URIRef(uri) ))
elif u['type'] == 'reference':
a_graph.add((LU[original_qname],DCTERMS['references'],URIRef(uri) ))
elif u['type'] == 'link' :
a_graph.add((LU[original_qname],SKOS.related, URIRef(uri)))
else :
a_graph.add((LU[original_qname],SKOS.related, URIRef(uri)))
graph = ConjunctiveGraph(store)
# out = ""
# for s, p, o, gr in graph.quads((None, None, None)) :
# out += "{} > {} {} {}\n".format(gr.identifier, s, p, o)
#
# app.logger.debug(out)
return graph