def test_related_titles(self):
# if two titles, one with a type and the other with an id, first should
# be part of the second
# sample from hughes-edwards1145
two_titles = '''<c01 level="file" xmlns="%s">
<did>
<container type="box">1</container>
<container type="folder">6</container>
<unittitle><title type="poetry" render="doublequote">Mother to Son</title>,
poem in
<title source="oclc" authfilenumber="10870853">The People's Voice</title>,
May 9, 1942
</unittitle>
</did>
</c01>''' % EAD_NAMESPACE
g = self._render_item_to_rdf(two_titles)
# there should be a manuscript in the output
ms_triples = list(g.triples((None, rdflib.RDF.type, self.BIBO.Manuscript)))
self.assert_(ms_triples, 'RDFa output should include an item with type bibo:Manuscript')
# first element of the first triple should be the article node
ms_node = ms_triples[0][0]
self.assertEqual(u'Mother to Son',
unicode(g.value(ms_node, self.DC.title)),
'poem title should be set as dc:title')
self.assertEqual(u'poetry',
unicode(g.value(ms_node, self.SCHEMA_ORG.genre)),
'genre should be set as poetry')
book_uriref = rdflib.URIRef(title_rdf_identifier('oclc', '10870853'))
self.assertTrue((ms_node, self.DC.isPartOf, book_uriref) in g,
'poem should be part of document with an id')
self.assertTrue((book_uriref, rdflib.RDF.type, self.BIBO.Document),
'document with id should be a bibo:Document')
self.assert_(g.triples((None, self.SCHEMA_ORG.mentions, book_uriref)),
'document with id should be mentioned by context (i.e. collection)')
def test_extra_titles(self):
# more than two titles in a unittitle
# sample from hughes-edwards1145
two_titles = '''<c01 level="file" xmlns="%s">
<did>
<container type="box">1</container>
<container type="folder">7</container>
<unittitle><title type="essays" render="doublequote">The Need for Heroes</title>,
essay in <title source="ISSN" authfilenumber=" 2169-1010">The Crisis</title>,
June 1941 [includes pages in which poems
<title type="poetry" render="doublequote">The Negro Speaks of Rivers</title>
and <title type="poetry" render="doublequote">NAACP</title> appear]
</unittitle>
</did>
</c01>''' % EAD_NAMESPACE
g = self._render_item_to_rdf(two_titles)
# first two titles should work roughly as related title does above
# there should be *three* manuscripts in the output
ms_triples = list(g.triples((None, rdflib.RDF.type, self.BIBO.Manuscript)))
self.assertEqual(3, len(ms_triples),
'RDFa output should include three items with type bibo:Manuscript')
# since rdf is unsorted, we have to find them by title
ms_by_title = {}
for ms in ms_triples:
s, p, o = ms
if unicode(g.value(s, self.DC.title)) == unicode('The Need for Heroes'):
ms_by_title['Need'] = s
if unicode(g.value(s, self.DC.title)) == unicode('The Negro Speaks of Rivers'):
ms_by_title['Negro'] = s
if unicode(g.value(s, self.DC.title)) == unicode('NAACP'):
ms_by_title['NAACP'] = s
self.assertEqual(u'essays',
unicode(g.value(ms_by_title['Need'], self.SCHEMA_ORG.genre)),
'essay title genre should be set as essays')
self.assertEqual(u'poetry',
unicode(g.value(ms_by_title['Negro'], self.SCHEMA_ORG.genre)),
'poem title genre should be set as poetry')
self.assertEqual(u'poetry',
unicode(g.value(ms_by_title['NAACP'], self.SCHEMA_ORG.genre)),
'poem title genre should be set as poetry')
book_uriref = rdflib.URIRef(title_rdf_identifier('issn', '2169-1010'))
# things that should be consistent for each of the typed titles
for ms_node in ms_by_title.itervalues():
# each item should be part of the book with title
self.assertTrue((ms_node, self.DC.isPartOf, book_uriref) in g,
'item should be part of document with an id')
# each item should be mentioned in context (i.e. collection)
self.assert_(g.triples((None, self.SCHEMA_ORG.mentions, ms_node)),
'item should be mentioned by context (i.e. collection)')
self.assertTrue((book_uriref, rdflib.RDF.type, self.BIBO.Periodical),
'document with issn should be a bibo:Periodical')
self.assert_(g.triples((None, self.SCHEMA_ORG.mentions, book_uriref)),
'document with id should be mentioned by context (i.e. collection)')
def format_title(node, default_rel):
"display a title node as semantic information"
title_type = node.get("type", None)
title_source = node.get("source", None)
title_authfileno = node.get("authfilenumber", None)
# lower-case title source for consistency of checking (isbn/issn/oclc)
if title_source is not None:
title_source = title_source.strip().lower()
# strip space from authfileno, to avoid any issues due to human error
if title_authfileno is not None:
title_authfileno = title_authfileno.strip()
start, end = "", ""
# special case we can't do anything with
# if a title is inside the bioghist, we can assume it was created by the
# originator, *however* there is no inverse relationship to specify
# a title was created by a person. We also can't assume
# any relation to the collection. So, skip these titles for now.
if node.xpath("ancestor::e:bioghist", **eadns):
return (start, end)
# similar special case: if a title is inside a series scopecontent note
# which is related to a series unititle person (see note on series_section_rdfa),
# do not generate any RDFa for that title
if node.xpath("ancestor::e:scopecontent/preceding-sibling::e:did/e:unittitle[e:corpname or e:persname]", **eadns):
return (start, end)
# for now, ignore titles in correspondence series
# (getting associated with the person inappropriately)
if default_rel == "schema:knows arch:correspondedWith":
return start, end
# if isbn # or issn is available, include it
meta_tags = ""
if title_source in ["isbn", "issn"] and title_authfileno is not None:
meta_tags += '<meta property="schema:%s" content="%s"/>' % (title_source, title_authfileno)
# title attribute carries genre information
if title_type is not None:
meta_tags += '<meta property="schema:genre" content="%s"/>' % title_type
# generate URI/URN for item when possible
# TODO: abstract into reusable function
resource_id = None
if title_authfileno is not None:
resource_id = title_rdf_identifier(title_source, title_authfileno)
# resource attribute for inclusion
resource = ' resource="%s"' % resource_id if resource_id else ""
# if title is inside the scopecontent, it needs to be wrapped as a document
# just use the generic "mentions" relation
if node.xpath("ancestor::e:scopecontent", **eadns):
# mark as a generic document or periodical and include whatever meta tags are available
itemtype = "bibo:Periodical" if title_source == "issn" else "bibo:Document"
start = '<span rel="schema:mentions" typeof="%s"%s><span property="dc:title">' % (itemtype, resource)
end = "</span>%s</span>" % meta_tags
# if default rel is set to mention, assume we are patching in extra titles
# after file item unittitle context
elif default_rel == "schema:mentions":
if title_source == "issn":
itemtype = "bibo:Periodical"
elif title_source in ["isbn", "oclc"]:
itemtype = "bibo:Document"
else:
# technically we should be checking if this is a printed materials series...
itemtype = "bibo:Manuscript"
# if this title has a type but no authfilenumber and there is a
# sibling title with an id, relate them
if (
title_type is not None
and title_authfileno is None
and node.xpath("parent::e:unittitle/e:title[@authfilenumber]", **eadns)
):
rel_id = None
rel_authfileno = node.xpath("normalize-space(parent::e:unittitle/e:title/@authfilenumber)", **eadns).strip()
rel_source = node.xpath("normalize-space(parent::e:unittitle/e:title/@source)", **eadns).lower()
rel_id = title_rdf_identifier(rel_source, rel_authfileno)
if rel_id is not None:
meta_tags += '<span property="dcterms:isPartOf" resource="%s"/>' % rel_id
start = '<span rel="schema:mentions" typeof="%s"%s><span property="dc:title">' % (itemtype, resource)
end = "</span>%s</span>" % meta_tags
# Otherwise, only add semantic information if there is a title type OR
# if title occurs in a file-level unittitle.
# (in that case, we assume it is title of the item in the container)
elif node.xpath('parent::e:unittitle and ancestor::e:*[@level="file"]', **eadns) or title_type is not None:
start, end = '<span property="dc:title">', "</span>%s" % meta_tags
# include meta tags after the title, since it should be in the
# context of the item, which is the whole unitittle
# check for special case: multiple titles with an author
# (persname tagged with a role, i.e. this is a Belfast Group sheet),
multiple_with_author = (
title_type is None
and node.xpath("count(parent::e:unittitle/e:title)", **eadns) > 1
and node.xpath("preceding-sibling::e:persname[@role]", **eadns)
)
# special case: no good way to relate more than two titles in a unittitle,
# so just skip them when generating rdfa
if node.xpath("count(preceding-sibling::e:title)", **eadns) >= 2 and not multiple_with_author:
start, end = "", ""
# if ISSN with preceding title, assume article in a periodical
elif title_source == "issn" and node.xpath("count(preceding-sibling::e:title)", **eadns) == 1:
# adapted from schema.org article example: http://schema.org/Article
start = '<span property="dcterms:isPartOf" typeof="bibo:Periodical"%s><span property="dc:title">' % resource
# include any meta tags (genre, issn) inside the periodical entity
end = "</span>%s</span>" % meta_tags
# otherwise, if current title has an id or no type and follows a title with a type,
# assume generic part/whole relationship
elif (title_authfileno is not None or title_type is None) and node.xpath(
"count(./preceding-sibling::e:title[@type])", **eadns
) == 1:
start = '<span property="dcterms:isPartOf" typeof="bibo:Document"%s><span property="dc:title">' % resource
# include any meta tags (e.g. isbn) inside the document entity
end = "</span>%s</span>" % meta_tags
# if no type and there are multiple titles, AND there is a persname
# tagged with a role (i.e. this is a Belfast Group sheet),
# then use RDFa list notation to generate a sequence
elif (
title_type is None
and node.xpath("count(parent::e:unittitle/e:title)", **eadns) > 1
and node.xpath("preceding-sibling::e:persname[@role]", **eadns)
):
start = '<span inlist="inlist" property="dc:title">'
return (start, end)