def test_all_authors_get_viaf_lookup(self):
# TODO: The code this calls could be refactored quite a bit --
# we don't really need to test all of process_item() here.
# But ATM it does seem to be our only test of process_item().
oclc = MockOCLCLinkedDataAPI()
viaf = MockVIAFClient()
provider = LinkedDataCoverageProvider(
self._db, api=oclc, viaf_api=viaf
)
# Here's a placeholder that will be filled in with information from
# OCLC Linked Data.
edition = self._edition()
for i in edition.contributions:
self._db.delete(i)
self._db.commit()
identifier = edition.primary_identifier
# OCLC Linked Data is going to mention two authors -- one with
# a sort name + VIAF, and one with a VIAF but no sort name.
contributor1 = ContributorData(viaf="1")
contributor2 = ContributorData(viaf="2", sort_name="Jordan, Robert")
idata = IdentifierData(type=identifier.type,
identifier=identifier.identifier)
metadata = Metadata(
DataSource.OCLC_LINKED_DATA,
contributors=[contributor1, contributor2],
primary_identifier=idata,
title=u"foo"
)
oclc.queue_info_for(metadata)
# Our OCLC Linked Data client is going to try to fill in the
# data, asking VIAF about the contributors.
lookup1 = (ContributorData(
viaf="1", display_name="Display Name",
family_name="Family", sort_name="Name, Sort",
wikipedia_name="Wikipedia_Name"), None, None)
lookup2 = (ContributorData(
viaf="2", wikipedia_name="Robert_Jordan_(Author)",
biography="That guy."), None, None)
viaf.queue_lookup(lookup1, lookup2)
provider.process_item(identifier)
# Both authors have had their information updated with the
# VIAF results.
filled_in = sorted(
[(x.sort_name, x.display_name, x.viaf, x.wikipedia_name, x.biography)
for x in edition.contributors]
)
eq_(
[(u'Jordan, Robert', None, u'2', u'Robert_Jordan_(Author)', u'That guy.'),
(u'Name, Sort', u'Display Name', u'1', u'Wikipedia_Name', None)],
filled_in
)
class RedoOCLC(Explain):
def __init__(self):
self.coverage = LinkedDataCoverageProvider(self._db)
@property
def oclcld(self):
return DataSource.lookup(self._db, DataSource.OCLC_LINKED_DATA)
def run(self):
id_type, identifier = sys.argv[1:]
identifier, ignore = Identifier.for_foreign_id(
self._db, id_type, identifier
)
self.fix_identifier(identifier)
def fix_identifier(self, primary_identifier):
equivalent_ids = primary_identifier.equivalent_identifier_ids(
levels=6, threshold=0)
return self.fix_identifier_with_equivalents(primary_identifier, equivalent_ids)
def fix_identifier_with_equivalents(self, primary_identifier, equivalent_ids):
for edition in primary_identifier.primarily_identifies:
print "BEFORE"
self.explain(self._db, edition)
print "-" * 80
t1 = self._db.begin_nested()
equivalencies = self._db.query(Equivalency).filter(
Equivalency.data_source == self.oclcld).filter(
Equivalency.input_id.in_(equivalent_ids)
)
print "DELETING %d" % equivalencies.count()
for e in equivalencies:
if e.strength == 0:
print "DELETING %r" % e
self._db.delete(e)
t1.commit()
self.coverage.process_item(primary_identifier)
equivalent_ids = primary_identifier.equivalent_identifier_ids(
levels=6, threshold=0)
equivalencies = self._db.query(Equivalency).filter(
Equivalency.data_source == self.oclcld).filter(
Equivalency.input_id.in_(equivalent_ids),
)
for edition in primary_identifier.primarily_identifies:
if edition.work:
edition.work.calculate_presentation()
self.explain(self._db, edition)
print "I WOULD NOW EXPECT EVERYTHING TO BE FINE."
class RedoOCLC(Explain):
def __init__(self):
self.coverage = LinkedDataCoverageProvider(self._db)
@property
def oclcld(self):
return DataSource.lookup(self._db, DataSource.OCLC_LINKED_DATA)
def run(self):
id_type, identifier = sys.argv[1:]
identifier, ignore = Identifier.for_foreign_id(self._db, id_type,
identifier)
self.fix_identifier(identifier)
def fix_identifier(self, primary_identifier):
equivalent_ids = primary_identifier.equivalent_identifier_ids(
levels=6, threshold=0)
return self.fix_identifier_with_equivalents(primary_identifier,
equivalent_ids)
def fix_identifier_with_equivalents(self, primary_identifier,
equivalent_ids):
for edition in primary_identifier.primarily_identifies:
print "BEFORE"
self.explain(self._db, edition)
print "-" * 80
t1 = self._db.begin_nested()
equivalencies = self._db.query(Equivalency).filter(
Equivalency.data_source == self.oclcld).filter(
Equivalency.input_id.in_(equivalent_ids))
print "DELETING %d" % equivalencies.count()
for e in equivalencies:
if e.strength == 0:
print "DELETING %r" % e
self._db.delete(e)
t1.commit()
self.coverage.process_item(primary_identifier)
for edition in primary_identifier.primarily_identifies:
if edition.work:
edition.work.calculate_presentation()
self.explain(self._db, edition)
print "I WOULD NOW EXPECT EVERYTHING TO BE FINE."
def test_process_item_exception(self):
class DoomedOCLCLinkedData(OCLCLinkedData):
def info_for(self, identifier):
raise IOError("Exception!")
provider = LinkedDataCoverageProvider(self._db, api=DoomedOCLCLinkedData(self._db))
edition = self._edition()
identifier = edition.primary_identifier
result = provider.process_item(identifier)
assert isinstance(result, CoverageFailure)
assert "Exception!" in result.exception
def test_process_item_exception_missing_isbn(self):
class DoomedOCLCLinkedData(OCLCLinkedData):
def info_for(self, identifier):
raise IOError("Tried, but couldn't find location")
provider = LinkedDataCoverageProvider(
self._db, api=DoomedOCLCLinkedData(self._db)
)
edition = self._edition()
identifier = edition.primary_identifier
result = provider.process_item(identifier)
assert isinstance(result, CoverageFailure)
assert "OCLC doesn't know about this ISBN" in result.exception
def test_process_item_exception_missing_isbn(self):
class DoomedOCLCLinkedData(OCLCLinkedData):
def info_for(self, identifier):
raise IOError("Tried, but couldn't find location")
provider = LinkedDataCoverageProvider(self._db,
api=DoomedOCLCLinkedData(
self._db))
edition = self._edition()
identifier = edition.primary_identifier
result = provider.process_item(identifier)
assert isinstance(result, CoverageFailure)
assert "OCLC doesn't know about this ISBN" in result.exception
def test_process_item_exception(self):
class DoomedOCLCLinkedData(OCLCLinkedData):
def info_for(self, identifier):
raise IOError("Exception!")
provider = LinkedDataCoverageProvider(self._db,
api=DoomedOCLCLinkedData(
self._db))
edition = self._edition()
identifier = edition.primary_identifier
result = provider.process_item(identifier)
assert isinstance(result, CoverageFailure)
assert "Exception!" in result.exception
def test_viaf_authors_get_viaf_lookup(self):
# TODO: The code this calls could be refactored quite a bit --
# we don't really need to test all of process_item() here.
# But ATM it does seem to be our only test of process_item().
oclc = MockOCLCLinkedDataAPI()
viaf = MockVIAFClient()
provider = LinkedDataCoverageProvider(self._db,
api=oclc,
viaf_api=viaf)
# Here's a placeholder that will be filled in with information from
# OCLC Linked Data.
edition = self._edition()
for i in edition.contributions:
self._db.delete(i)
self._db.commit()
identifier = edition.primary_identifier
# OCLC Linked Data is going to mention two authors -- one with
# a sort name + VIAF, and one with a VIAF but no sort name.
contributor1 = ContributorData(viaf="1")
contributor2 = ContributorData(viaf="2", sort_name="Jordan, Robert")
contributor3 = ContributorData(sort_name="Rice, Anne",
display_name="Anne Rice")
idata = IdentifierData(type=identifier.type,
identifier=identifier.identifier)
metadata = Metadata(
DataSource.OCLC_LINKED_DATA,
contributors=[contributor1, contributor2, contributor3],
primary_identifier=idata,
title=u"foo")
oclc.queue_info_for(metadata)
# Our OCLC Linked Data client is going to try to fill in the
# data, asking VIAF about the contributors that have VIAF data,
# and not those who do not.
lookup1 = (ContributorData(viaf="1",
display_name="Display Name",
family_name="Family",
sort_name="Name, Sort",
wikipedia_name="Wikipedia_Name"), None,
None)
lookup2 = (ContributorData(viaf="2",
wikipedia_name="Robert_Jordan_(Author)",
biography="That guy."), None, None)
viaf.queue_lookup(lookup1, lookup2, "Unrequested lookup")
provider.process_item(identifier)
# Both VIAF-identified authors have had their information updated
# with the VIAF results.
filled_in = sorted([(x.sort_name, x.display_name, x.viaf,
x.wikipedia_name, x.biography)
for x in edition.contributors])
eq_([(u'Jordan, Robert', None, u'2', u'Robert_Jordan_(Author)',
u'That guy.'),
(u'Name, Sort', u'Display Name', u'1', u'Wikipedia_Name', None),
(u'Rice, Anne', u'Anne Rice', None, None, None)], filled_in)
# The author without VIAF data didn't request a VIAF lookup.
# Instead, that result is still in the mock VIAF queue.
eq_(viaf.results, ["Unrequested lookup"])
