def test_extract_multiple_works_with_author_restriction(self):
"""We can choose to only accept works by a given author."""
xml = self.sample_data("multi_work_response.xml")
[wrong_author], ignore = Contributor.lookup(self._db, sort_name="Wrong Author")
status, swids = OCLCXMLParser.parse(
self._db, xml, languages=["eng"], authors=[wrong_author])
# This person is not listed as an author of any work in the dataset,
# so none of those works were picked up.
eq_(0, len(swids))
[melville], ignore = Contributor.lookup(self._db, sort_name="Melville, Herman")
status, swids = OCLCXMLParser.parse(
self._db, xml, languages=["eng"], authors=[melville])
# We picked up 11 of the 25 works in the dataset.
eq_(11, len(swids))
# The missing works (as you can verify by looking at
# oclc_multi_work_response.xml) either don't credit Herman
# Melville at all (the 1956 Gregory Peck movie "Moby Dick"),
# credit him as "Associated name" rather than as an author
# (four books about "Moby Dick"), or credit him as an author
# but not as the primary author (academic works and adaptations).
for missing in '10798812', '13424036', '22658644', '250604212', '474972877', '13358012', '153927888', '13206523', '46935692', "14135019", "51088077", "105446800", "164732682", "26863225":
assert missing not in swids
def test_to_edition_sets_sort_author_name_if_obvious(self):
[contributor], ignore = Contributor.lookup(self._db, u"Hawkins, Paula")
contributor.display_name = u"Paula Hawkins"
title = NYTBestSellerListTitle(self.one_list_title)
edition = title.to_edition(self._db, self.metadata_client)
eq_(contributor.sort_name, edition.sort_author)
eq_(contributor.display_name, edition.author)
assert edition.permanent_work_id is not None
def test_to_edition_sets_sort_author_name_if_obvious(self):
[contributor], ignore = Contributor.lookup(
self._db, u"Hawkins, Paula")
contributor.display_name = u"Paula Hawkins"
title = NYTBestSellerListTitle(self.one_list_title, Edition.BOOK_MEDIUM)
edition = title.to_edition(self._db, self.metadata_client)
eq_(contributor.sort_name, edition.sort_author)
eq_(contributor.display_name, edition.author)
assert edition.permanent_work_id is not None
def test_extract_multiple_works_with_author_restriction(self):
"""We can choose to only accept works by a given author."""
xml = self.sample_data("multi_work_response.xml")
[wrong_author], ignore = Contributor.lookup(self._db, sort_name="Wrong Author")
status, swids = OCLCXMLParser.parse(self._db, xml, languages=["eng"], authors=[wrong_author])
# This person is not listed as an author of any work in the dataset,
# so none of those works were picked up.
eq_(0, len(swids))
[melville], ignore = Contributor.lookup(self._db, sort_name="Melville, Herman")
status, swids = OCLCXMLParser.parse(self._db, xml, languages=["eng"], authors=[melville])
# We picked up 11 of the 25 works in the dataset.
eq_(11, len(swids))
# The missing works (as you can verify by looking at
# oclc_multi_work_response.xml) either don't credit Herman
# Melville at all (the 1956 Gregory Peck movie "Moby Dick"),
# credit him as "Associated name" rather than as an author
# (four books about "Moby Dick"), or credit him as an author
# but not as the primary author (academic works and adaptations).
for missing in (
"10798812",
"13424036",
"22658644",
"250604212",
"474972877",
"13358012",
"153927888",
"13206523",
"46935692",
"14135019",
"51088077",
"105446800",
"164732682",
"26863225",
):
assert missing not in swids
def _parse_single_author(cls, _db, author,
lc=None, viaf=None,
existing_authors=[],
default_role=Contributor.AUTHOR_ROLE,
primary_author=None):
default_role_used = False
# First find roles if present
# "Giles, Lionel, 1875-1958 [Writer of added commentary; Translator]"
author = author.strip()
m = cls.ROLES.search(author)
if m:
author = author[:m.start()].strip()
role_string = m.groups()[0]
roles = [x.strip() for x in role_string.split(";")]
elif default_role:
roles = [default_role]
default_role_used = True
else:
roles = []
# Author string now looks like
# "Giles, Lionel, 1875-1958"
m = cls.LIFESPAN.search(author)
kwargs = dict()
if m:
author = author[:m.start()].strip()
birth, death = m.groups()
if birth:
kwargs[Contributor.BIRTH_DATE] = birth
if death:
kwargs[Contributor.DEATH_DATE] = death
# Author string now looks like
# "Giles, Lionel,"
if author.endswith(","):
author = author[:-1]
contributor = None
if not author:
# No name was given for the author.
return None, roles, default_role_used
if primary_author and author == primary_author.sort_name:
if Contributor.AUTHOR_ROLE in roles:
roles.remove(Contributor.AUTHOR_ROLE)
if Contributor.UNKNOWN_ROLE in roles:
roles.remove(Contributor.UNKNOWN_ROLE)
roles.insert(0, Contributor.PRIMARY_AUTHOR_ROLE)
if existing_authors:
# Calling Contributor.lookup will result in a database
# hit, and looking up a contributor based on name may
# result in multiple results (see below). We'll have no
# way of distinguishing between those results. If
# possible, it's much more reliable to look through
# existing_authors (the authors derived from an entry's
# <authors> tag).
for x in existing_authors:
if cls._contributor_match(x, author, lc, viaf):
contributor = x
break
if contributor:
was_new = False
if not contributor:
contributor, was_new = Contributor.lookup(
_db, author, viaf, lc, extra=kwargs)
if isinstance(contributor, list):
# We asked for an author based solely on the name, which makes
# Contributor.lookup() return a list.
if len(contributor) == 1:
# Fortunately, either the database knows about only
# one author with that name, or it didn't know about
# any authors with that name and it just created one,
# so we can unambiguously use it.
contributor = contributor[0]
else:
# Uh-oh. The database knows about multiple authors
# with that name. We have no basis for deciding which
# author we mean. But we would prefer to identify with
# an author who has a known LC or VIAF number.
#
# This should happen very rarely because of our check
# against existing_authors above. But it will happen
# for authors that have a work in Project Gutenberg.
with_id = [x for x in contributor if x.lc is not None
or x.viaf is not None]
if with_id:
contributor = with_id[0]
else:
contributor = contributor[0]
return contributor, roles, default_role_used
def _parse_single_author(cls,
_db,
author,
lc=None,
viaf=None,
existing_authors=[],
default_role=Contributor.AUTHOR_ROLE,
primary_author=None):
default_role_used = False
# First find roles if present
# "Giles, Lionel, 1875-1958 [Writer of added commentary; Translator]"
author = author.strip()
m = cls.ROLES.search(author)
if m:
author = author[:m.start()].strip()
role_string = m.groups()[0]
roles = [x.strip() for x in role_string.split(";")]
elif default_role:
roles = [default_role]
default_role_used = True
else:
roles = []
# Author string now looks like
# "Giles, Lionel, 1875-1958"
m = cls.LIFESPAN.search(author)
kwargs = dict()
if m:
author = author[:m.start()].strip()
birth, death = m.groups()
if birth:
kwargs[Contributor.BIRTH_DATE] = birth
if death:
kwargs[Contributor.DEATH_DATE] = death
# Author string now looks like
# "Giles, Lionel,"
if author.endswith(","):
author = author[:-1]
contributor = None
if not author:
# No name was given for the author.
return None, roles, default_role_used
if primary_author and author == primary_author.sort_name:
if Contributor.AUTHOR_ROLE in roles:
roles.remove(Contributor.AUTHOR_ROLE)
if Contributor.UNKNOWN_ROLE in roles:
roles.remove(Contributor.UNKNOWN_ROLE)
roles.insert(0, Contributor.PRIMARY_AUTHOR_ROLE)
if existing_authors:
# Calling Contributor.lookup will result in a database
# hit, and looking up a contributor based on name may
# result in multiple results (see below). We'll have no
# way of distinguishing between those results. If
# possible, it's much more reliable to look through
# existing_authors (the authors derived from an entry's
# <authors> tag).
for x in existing_authors:
if cls._contributor_match(x, author, lc, viaf):
contributor = x
break
if contributor:
was_new = False
if not contributor:
contributor, was_new = Contributor.lookup(_db,
author,
viaf,
lc,
extra=kwargs)
if isinstance(contributor, list):
# We asked for an author based solely on the name, which makes
# Contributor.lookup() return a list.
if len(contributor) == 1:
# Fortunately, either the database knows about only
# one author with that name, or it didn't know about
# any authors with that name and it just created one,
# so we can unambiguously use it.
contributor = contributor[0]
else:
# Uh-oh. The database knows about multiple authors
# with that name. We have no basis for deciding which
# author we mean. But we would prefer to identify with
# an author who has a known LC or VIAF number.
#
# This should happen very rarely because of our check
# against existing_authors above. But it will happen
# for authors that have a work in Project Gutenberg.
with_id = [
x for x in contributor
if x.lc is not None or x.viaf is not None
]
if with_id:
contributor = with_id[0]
else:
contributor = contributor[0]
return contributor, roles, default_role_used
