def create_missing_books(self, subset=None):
"""Finds books present in the PG catalog but missing from Edition.
Yields (Edition, LicensePool) 2-tuples.
"""
books = list(self.all_books())
for pg_id, archive, archive_item in books:
if subset is not None and not subset(pg_id, archive, archive_item):
continue
self.log.info("Considering %s" % pg_id)
# Find an existing Edition for the book.
book = Edition.for_foreign_id(
self._db, self.source, Identifier.GUTENBERG_ID, pg_id,
create_if_not_exists=False)
if not book:
# Create a new Edition object with bibliographic
# information from the Project Gutenberg RDF file.
fh = archive.extractfile(archive_item)
data = fh.read()
fake_fh = StringIO(data)
book, license, new = GutenbergRDFExtractor.book_in(
self.collection, pg_id, fake_fh)
if book and license:
yield (book, license)
def update_licensepool_with_book_info(self, book, license_pool, is_new_pool):
"""Update a book's LicensePool with information from a JSON
representation of its circulation info.
Then, create an Edition and make sure it has bibliographic
coverage. If the new Edition is the only candidate for the
pool's presentation_edition, promote it to presentation
status.
"""
circulation = OverdriveRepresentationExtractor.book_info_to_circulation(
book
)
license_pool, circulation_changed = circulation.apply(license_pool)
edition, is_new_edition = Edition.for_foreign_id(
self._db, self.source, license_pool.identifier.type,
license_pool.identifier.identifier)
# If the pool does not already have a presentation edition,
# and if this edition is newly made, then associate pool and edition
# as presentation_edition
if ((not license_pool.presentation_edition) and is_new_edition):
edition_changed = license_pool.set_presentation_edition(
policy=None
)
if is_new_pool:
license_pool.open_access = False
self.log.info("New Overdrive book discovered: %r", edition)
return license_pool, is_new_pool, circulation_changed
def update_licensepool_with_book_info(self, book, license_pool,
is_new_pool):
"""Update a book's LicensePool with information from a JSON
representation of its circulation info.
Then, create an Edition and make sure it has bibliographic
coverage. If the new Edition is the only candidate for the
pool's presentation_edition, promote it to presentation
status.
"""
circulation = OverdriveRepresentationExtractor.book_info_to_circulation(
book)
license_pool, circulation_changed = circulation.apply(
self._db, license_pool.collection)
edition, is_new_edition = Edition.for_foreign_id(
self._db, self.source, license_pool.identifier.type,
license_pool.identifier.identifier)
# If the pool does not already have a presentation edition,
# and if this edition is newly made, then associate pool and edition
# as presentation_edition
if ((not license_pool.presentation_edition) and is_new_edition):
edition_changed = license_pool.set_presentation_edition()
if is_new_pool:
license_pool.open_access = False
self.log.info("New Overdrive book discovered: %r", edition)
return license_pool, is_new_pool, circulation_changed
def process_item(self, identifier):
"""Associate bibliographic metadata with the given Identifier.
:param Identifier: Look up this Identifier on Content Cafe.
"""
try:
# Create a Metadata object.
metadata = self.content_cafe.create_metadata(identifier)
if not metadata:
# TODO: The only time this is really a transient error
# is when the book is too new for Content Cafe to know
# about it, which isn't often. It would be best to
# keep this as a transient failure but give it a relatively
# long and exponentially increasing retry time.
return self.failure(
identifier,
"Content Cafe has no knowledge of this identifier.",
transient=True
)
edition, is_new = Edition.for_foreign_id(
self._db, self.data_source, identifier.type,
identifier.identifier
)
# We're passing in collection=None even though we
# technically have a Collection available, because our
# goal is to add metadata for the book without reference
# to any particular collection.
metadata.apply(
edition, collection=None, replace=self.replacement_policy
)
return identifier
except Exception as e:
self.log.error('Coverage error for %r', identifier, exc_info=e)
return self.failure(identifier, repr(e), transient=True)
def process_item(self, identifier):
"""Associate bibliographic metadata with the given Identifier.
:param Identifier: Look up this Identifier on Content Cafe.
"""
try:
# Create a Metadata object.
metadata = self.content_cafe.create_metadata(identifier)
if not metadata:
# TODO: The only time this is really a transient error
# is when the book is too new for Content Cafe to know
# about it, which isn't often. It would be best to
# keep this as a transient failure but give it a relatively
# long and exponentially increasing retry time.
return self.failure(
identifier,
"Content Cafe has no knowledge of this identifier.",
transient=True)
edition, is_new = Edition.for_foreign_id(self._db,
self.data_source,
identifier.type,
identifier.identifier)
# We're passing in collection=None even though we
# technically have a Collection available, because our
# goal is to add metadata for the book without reference
# to any particular collection.
metadata.apply(edition,
collection=None,
replace=self.replacement_policy)
return identifier
except Exception as e:
self.log.error('Coverage error for %r', identifier, exc_info=e)
return self.failure(identifier, repr(e), transient=True)
def test_finalize_edition(self):
provider_no_presentation_ready = self._provider(presentation_ready_on_success=False)
provider_presentation_ready = self._provider(presentation_ready_on_success=True)
identifier = self._identifier()
source = DataSource.lookup(self._db, DataSource.GUTENBERG)
# Here's an Edition with no LicensePool.
edition, is_new = Edition.for_foreign_id(
self._db, source, identifier.type, identifier.identifier
)
edition.title = self._str
# This will effectively do nothing.
provider_no_presentation_ready.finalize_edition(edition)
# No Works have been created.
eq_(0, self._db.query(Work).count())
# But if there's also a LicensePool...
pool, is_new = LicensePool.for_foreign_id(
self._db, source, identifier.type, identifier.identifier
)
# finalize_edition() will create a Work.
provider_no_presentation_ready.finalize_edition(edition)
work = pool.work
eq_(work, edition.work)
eq_(False, work.presentation_ready)
# If the provider is configured to do so, finalize_edition()
# will also set the Work as presentation-ready.
provider_presentation_ready.finalize_edition(edition)
eq_(True, work.presentation_ready)
def _edition(self, licensepool):
"""Find or create the Edition that would be used to contain
Overdrive metadata for the given LicensePool.
"""
return Edition.for_foreign_id(
self._db, self.source, licensepool.identifier.type,
licensepool.identifier.identifier
)
def _edition(self, licensepool):
"""Find or create the Edition that would be used to contain
Overdrive metadata for the given LicensePool.
"""
return Edition.for_foreign_id(
self._db, self.source, licensepool.identifier.type,
licensepool.identifier.identifier
)
def edition(self, identifier):
"""Finds or creates an edition with license-offering DataSource.GUTENBERG
instead of the class's local DataSource.GUTENBERG_EPUB_GENERATOR
"""
edition, is_new = Edition.for_foreign_id(
self._db, DataSource.GUTENBERG,
identifier.type, identifier.identifier
)
return edition
def items_that_need_coverage(self):
"""Returns Editions associated with an open-access LicensePool but
with no open-access download URL.
"""
q = Edition.missing_coverage_from(self._db, [], self.coverage_source)
clause = and_(Edition.data_source_id==LicensePool.data_source_id,
Edition.primary_identifier_id==LicensePool.identifier_id)
q = q.join(LicensePool, clause)
q = q.filter(LicensePool.open_access == True).filter(
Edition.open_access_download_url==None
)
return q
def handle_event(self, bibliotheca_id, isbn, foreign_patron_id, start_time,
end_time, internal_event_type):
# Find or lookup the LicensePool for this event.
license_pool, is_new = LicensePool.for_foreign_id(
self._db,
self.api.source,
Identifier.BIBLIOTHECA_ID,
bibliotheca_id,
collection=self.collection)
if is_new:
# This is a new book. Immediately acquire bibliographic
# coverage for it. This will set the
# DistributionMechanisms and make the book
# presentation-ready. However, its circulation information
# might not be up to date until we process some more
# events.
record = self.bibliographic_coverage_provider.ensure_coverage(
license_pool.identifier, force=True)
bibliotheca_identifier = license_pool.identifier
isbn, ignore = Identifier.for_foreign_id(self._db, Identifier.ISBN,
isbn)
edition, ignore = Edition.for_foreign_id(self._db, self.api.source,
Identifier.BIBLIOTHECA_ID,
bibliotheca_id)
# The ISBN and the Bibliotheca identifier are exactly equivalent.
bibliotheca_identifier.equivalent_to(self.api.source, isbn, strength=1)
# Log the event.
start = start_time or CirculationEvent.NO_DATE
# Make sure the effects of the event reported by Bibliotheca
# are made visible on the LicensePool and turned into
# analytics events. This is not 100% reliable, but it
# should be mostly accurate, and the BibliothecaCirculationSweep
# will periodically correct the errors.
license_pool.update_availability_from_delta(internal_event_type,
start_time, 1,
self.analytics)
if is_new:
# This is our first time seeing this LicensePool. Log its
# occurance as a separate event.
license_pool.collect_analytics_event(
self.analytics, CirculationEvent.DISTRIBUTOR_TITLE_ADD,
license_pool.last_checked or start_time, 0, 1)
title = edition.title or "[no title]"
self.log.info("%r %s: %s", start_time, title, internal_event_type)
return start_time
def handle_event(self, threem_id, isbn, foreign_patron_id, start_time,
end_time, internal_event_type):
# Find or lookup the LicensePool for this event.
license_pool, is_new = LicensePool.for_foreign_id(
self._db, self.api.source, Identifier.THREEM_ID, threem_id)
if is_new:
# Immediately acquire bibliographic coverage for this book.
# This will set the DistributionMechanisms and make the
# book presentation-ready. However, its circulation information
# might not be up to date until we process some more events.
record = self.bibliographic_coverage_provider.ensure_coverage(
license_pool.identifier, force=True)
threem_identifier = license_pool.identifier
isbn, ignore = Identifier.for_foreign_id(self._db, Identifier.ISBN,
isbn)
edition, ignore = Edition.for_foreign_id(self._db, self.api.source,
Identifier.THREEM_ID,
threem_id)
# The ISBN and the 3M identifier are exactly equivalent.
threem_identifier.equivalent_to(self.api.source, isbn, strength=1)
# Log the event.
event, was_new = get_one_or_create(self._db,
CirculationEvent,
license_pool=license_pool,
type=internal_event_type,
start=start_time,
foreign_patron_id=foreign_patron_id,
create_method_kwargs=dict(
delta=1, end=end_time))
# If this is our first time seeing this LicensePool, log its
# occurance as a separate event
if is_new:
event = get_one_or_create(
self._db,
CirculationEvent,
type=CirculationEvent.DISTRIBUTOR_TITLE_ADD,
license_pool=license_pool,
create_method_kwargs=dict(
start=license_pool.last_checked or start_time,
delta=1,
end=license_pool.last_checked or end_time,
))
title = edition.title or "[no title]"
self.log.info("%r %s: %s", start_time, title, internal_event_type)
return start_time
def handle_event(self, threem_id, isbn, foreign_patron_id,
start_time, end_time, internal_event_type):
# Find or lookup the LicensePool for this event.
license_pool, is_new = LicensePool.for_foreign_id(
self._db, self.api.source, Identifier.THREEM_ID, threem_id)
if is_new:
# Immediately acquire bibliographic coverage for this book.
# This will set the DistributionMechanisms and make the
# book presentation-ready. However, its circulation information
# might not be up to date until we process some more events.
record = self.bibliographic_coverage_provider.ensure_coverage(
license_pool.identifier, force=True
)
threem_identifier = license_pool.identifier
isbn, ignore = Identifier.for_foreign_id(
self._db, Identifier.ISBN, isbn)
edition, ignore = Edition.for_foreign_id(
self._db, self.api.source, Identifier.THREEM_ID, threem_id)
# The ISBN and the 3M identifier are exactly equivalent.
threem_identifier.equivalent_to(self.api.source, isbn, strength=1)
# Log the event.
event, was_new = get_one_or_create(
self._db, CirculationEvent, license_pool=license_pool,
type=internal_event_type, start=start_time,
foreign_patron_id=foreign_patron_id,
create_method_kwargs=dict(delta=1,end=end_time)
)
# If this is our first time seeing this LicensePool, log its
# occurance as a separate event
if is_new:
event = get_one_or_create(
self._db, CirculationEvent,
type=CirculationEvent.TITLE_ADD,
license_pool=license_pool,
create_method_kwargs=dict(
start=license_pool.last_checked or start_time,
delta=1,
end=license_pool.last_checked or end_time,
)
)
title = edition.title or "[no title]"
self.log.info("%r %s: %s", start_time, title, internal_event_type)
return start_time
def test_register_equivalency(self):
data_source = DataSource.lookup(self._db, DataSource.GUTENBERG)
id = "549"
# We've got a record.
record, was_new = Edition.for_foreign_id(self._db, data_source,
Identifier.GUTENBERG_ID, id)
# Then we look it up and discover another identifier for it.
data_source_2 = DataSource.lookup(self._db, DataSource.OCLC)
record2, was_new = Edition.for_foreign_id(self._db, data_source_2,
Identifier.OCLC_NUMBER, "22")
eq = record.primary_identifier.equivalent_to(
data_source_2, record2.primary_identifier, 1)
assert eq.input == record.primary_identifier
assert eq.output == record2.primary_identifier
assert eq.data_source == data_source_2
assert [eq] == record.primary_identifier.equivalencies
assert set([record,
record2]) == set(record.equivalent_editions().all())
def test_finalize_license_pool(self):
provider_no_presentation_ready = self._provider(
presentation_ready_on_success=False)
provider_presentation_ready = self._provider(
presentation_ready_on_success=True)
identifier = self._identifier()
license_source = DataSource.lookup(self._db, DataSource.GUTENBERG)
data_source = DataSource.lookup(self._db, DataSource.OA_CONTENT_SERVER)
# Here's a LicensePool with no presentation edition.
pool, is_new = LicensePool.for_foreign_id(self._db, license_source,
identifier.type,
identifier.identifier)
eq_(None, pool.presentation_edition)
# Calling finalize_license_pool() here won't do much.
provider_no_presentation_ready.finalize_license_pool(pool)
# A presentation edition has been created for the LicensePool,
# but it has no title (in fact it has no data at all), so no
# Work was created.
eq_(None, pool.presentation_edition.title)
eq_(0, self._db.query(Work).count())
# Here's an Edition for the same book as the LicensePool but
# from a different data source.
edition, is_new = Edition.for_foreign_id(self._db, data_source,
identifier.type,
identifier.identifier)
edition.title = self._str
# Although Edition and LicensePool share an identifier, they
# are not otherwise related.
eq_(None, pool.presentation_edition.title)
# finalize_license_pool() will create a Work and update the
# LicensePool's presentation edition, based on the brand-new
# Edition.
provider_no_presentation_ready.finalize_license_pool(pool)
work = pool.work
eq_(edition.title, pool.presentation_edition.title)
eq_(False, work.presentation_ready)
# If the provider is configured to do so, finalize_license_pool()
# will also set the Work as presentation-ready.
provider_presentation_ready.finalize_license_pool(pool)
eq_(True, work.presentation_ready)
def update_licensepool_with_book_info(self, book, license_pool, is_new):
"""Update a book's LicensePool with information from a JSON
representation of its circulation info.
Also creates an Edition and gives it very basic bibliographic
information (the title), if possible.
"""
circulation = OverdriveRepresentationExtractor.book_info_to_circulation(
book
)
circulation_changed = circulation.update(license_pool, is_new)
edition, ignore = Edition.for_foreign_id(
self._db, self.source, license_pool.identifier.type,
license_pool.identifier.identifier)
edition.title = edition.title or book.get('title')
if is_new:
license_pool.open_access = False
self.log.info("New Overdrive book discovered: %r", edition)
return license_pool, is_new, circulation_changed
def test_finalize_license_pool(self):
identifier = self._identifier()
license_source = DataSource.lookup(self._db, DataSource.GUTENBERG)
data_source = DataSource.lookup(self._db, DataSource.OA_CONTENT_SERVER)
# Here's a LicensePool with no presentation edition.
pool, is_new = LicensePool.for_foreign_id(
self._db,
license_source,
identifier.type,
identifier.identifier,
collection=self._default_collection)
eq_(None, pool.presentation_edition)
# Here's an Edition for the same book as the LicensePool but
# from a different data source.
edition, is_new = Edition.for_foreign_id(self._db, data_source,
identifier.type,
identifier.identifier)
edition.title = self._str
# Although Edition and LicensePool share an identifier, they
# are not otherwise related.
eq_(None, pool.presentation_edition)
# finalize_license_pool() will create a Work and update the
# LicensePool's presentation edition, based on the brand-new
# Edition.
lookup = MockSimplifiedOPDSLookup(self._url)
provider = ContentServerBibliographicCoverageProvider(
self._default_collection, lookup)
provider.finalize_license_pool(pool)
work = pool.work
eq_(edition.title, pool.presentation_edition.title)
eq_(True, work.presentation_ready)
def _extract_publication_metadata(self, feed, publication, data_source_name):
"""Extract a Metadata object from webpub-manifest-parser's publication.
:param publication: Feed object
:type publication: opds2_ast.OPDS2Feed
:param publication: Publication object
:type publication: opds2_ast.OPDS2Publication
:param data_source_name: Data source's name
:type data_source_name: str
:return: Publication's metadata
:rtype: Metadata
"""
metadata = super(ODL2Importer, self)._extract_publication_metadata(
feed, publication, data_source_name
)
formats = []
licenses = []
medium = None
with self._get_configuration(self._db) as configuration:
skipped_license_formats = configuration.skipped_license_formats
if skipped_license_formats:
skipped_license_formats = set(skipped_license_formats)
if publication.licenses:
for odl_license in publication.licenses:
identifier = odl_license.metadata.identifier
checkout_link = first_or_default(
odl_license.links.get_by_rel(OPDS2LinkRelationsRegistry.BORROW.key)
)
if checkout_link:
checkout_link = checkout_link.href
license_info_document_link = first_or_default(
odl_license.links.get_by_rel(OPDS2LinkRelationsRegistry.SELF.key)
)
if license_info_document_link:
license_info_document_link = license_info_document_link.href
expires = (
to_utc(odl_license.metadata.terms.expires)
if odl_license.metadata.terms
else None
)
concurrency = (
int(odl_license.metadata.terms.concurrency)
if odl_license.metadata.terms
else None
)
if not license_info_document_link:
parsed_license = None
else:
parsed_license = ODLImporter.get_license_data(
license_info_document_link,
checkout_link,
identifier,
expires,
concurrency,
self.http_get,
)
if parsed_license is not None:
licenses.append(parsed_license)
# DPLA feed doesn't have information about a DRM protection used for audiobooks.
# We want to try to extract that information from the License Info Document it's present there.
license_formats = set(odl_license.metadata.formats)
if parsed_license and parsed_license.content_types:
license_formats |= set(parsed_license.content_types)
for license_format in license_formats:
if (
skipped_license_formats
and license_format in skipped_license_formats
):
continue
if not medium:
medium = Edition.medium_from_media_type(license_format)
if license_format in ODLImporter.LICENSE_FORMATS:
# Special case to handle DeMarque audiobooks which
# include the protection in the content type
drm_schemes = [
ODLImporter.LICENSE_FORMATS[license_format][
ODLImporter.DRM_SCHEME
]
]
license_format = ODLImporter.LICENSE_FORMATS[license_format][
ODLImporter.CONTENT_TYPE
]
else:
drm_schemes = (
odl_license.metadata.protection.formats
if odl_license.metadata.protection
else []
)
for drm_scheme in drm_schemes or [None]:
formats.append(
FormatData(
content_type=license_format,
drm_scheme=drm_scheme,
rights_uri=RightsStatus.IN_COPYRIGHT,
)
)
metadata.circulation.licenses = licenses
metadata.circulation.licenses_owned = None
metadata.circulation.licenses_available = None
metadata.circulation.licenses_reserved = None
metadata.circulation.patrons_in_hold_queue = None
metadata.circulation.formats.extend(formats)
metadata.medium = medium
return metadata
def test_recursively_equivalent_identifiers(self):
# We start with a Gutenberg book.
gutenberg = DataSource.lookup(self._db, DataSource.GUTENBERG)
record, ignore = Edition.for_foreign_id(self._db, gutenberg,
Identifier.GUTENBERG_ID, "100")
gutenberg_id = record.primary_identifier
# We use OCLC Classify to do a title/author lookup.
oclc = DataSource.lookup(self._db, DataSource.OCLC)
search_id, ignore = Identifier.for_foreign_id(self._db,
Identifier.OCLC_WORK,
"60010")
gutenberg_id.equivalent_to(oclc, search_id, 1)
# The title/author lookup associates the search term with two
# different OCLC Numbers.
oclc_id, ignore = Identifier.for_foreign_id(self._db,
Identifier.OCLC_NUMBER,
"9999")
oclc_id_2, ignore = Identifier.for_foreign_id(self._db,
Identifier.OCLC_NUMBER,
"1000")
search_id.equivalent_to(oclc, oclc_id, 1)
search_id.equivalent_to(oclc, oclc_id_2, 1)
# We then use OCLC Linked Data to connect one of the OCLC
# Numbers with an ISBN.
linked_data = DataSource.lookup(self._db, DataSource.OCLC_LINKED_DATA)
isbn_id, ignore = Identifier.for_foreign_id(self._db, Identifier.ISBN,
"900100434X")
oclc_id.equivalent_to(linked_data, isbn_id, 1)
# As it turns out, we have an Overdrive work record...
overdrive = DataSource.lookup(self._db, DataSource.OVERDRIVE)
overdrive_record, ignore = Edition.for_foreign_id(
self._db, overdrive, Identifier.OVERDRIVE_ID, "{111-222}")
overdrive_id = overdrive_record.primary_identifier
# ...which is tied (by Overdrive) to the same ISBN.
overdrive_id.equivalent_to(overdrive, isbn_id, 1)
# Finally, here's a completely unrelated Edition, which
# will not be showing up.
gutenberg2, ignore = Edition.for_foreign_id(self._db, gutenberg,
Identifier.GUTENBERG_ID,
"200")
gutenberg2.title = "Unrelated Gutenberg record."
levels = [
record.equivalent_identifiers(policy=PresentationCalculationPolicy(
equivalent_identifier_levels=i)) for i in range(0, 5)
]
# At level 0, the only identifier found is the Gutenberg ID.
assert set([gutenberg_id]) == set(levels[0])
# At level 1, we pick up the title/author lookup.
assert set([gutenberg_id, search_id]) == set(levels[1])
# At level 2, we pick up the title/author lookup and the two
# OCLC Numbers.
assert set([gutenberg_id, search_id, oclc_id,
oclc_id_2]) == set(levels[2])
# At level 3, we also pick up the ISBN.
assert set([gutenberg_id, search_id, oclc_id, oclc_id_2,
isbn_id]) == set(levels[3])
# At level 4, the recursion starts to go in the other
# direction: we pick up the Overdrive ID that's equivalent to
# the same ISBN as the OCLC Number.
assert set([
gutenberg_id, search_id, oclc_id, oclc_id_2, isbn_id, overdrive_id
]) == set(levels[4])
