def get_or_create_artist(name):
if ' ' in name:
parts = name.split()
if len(parts) == 2:
last, first = parts
else:
last, first = parts[-1], ' '.join(parts[:-1]) # Diana Wynne Jones
else:
last, first = name, ''
try:
contestants = []
if Artist.objects.filter(first_name=first, last_name=last).count():
contestants = Artist.objects.filter(first_name=first,
last_name=last)
return Artist.objects.get(first_name=first, last_name=last)
elif first != '' and Artist.objects.filter(first_name=first).count():
contestants = Artist.objects.filter(first_name=first)
elif Artist.objects.filter(last_name=last).count():
contestants = Artist.objects.filter(last_name=last)
except:
pass
if contestants:
choice = pick_among(contestants)
if choice:
return choice
artist = Artist(first_name=first, last_name=last)
artist.save()
return artist
def get_or_create_artist(name):
if ' ' in name:
parts = name.split()
if len(parts) == 2:
last, first = parts
else:
last, first = parts[-1], ' '.join(parts[:-1]) # Diana Wynne Jones
else:
last, first = name, ''
try:
contestants = []
if Artist.objects.filter(first_name=first, last_name=last).count():
contestants = Artist.objects.filter(first_name=first, last_name=last)
return Artist.objects.get(first_name=first, last_name=last)
elif first != '' and Artist.objects.filter(first_name=first).count():
contestants = Artist.objects.filter(first_name=first)
elif Artist.objects.filter(last_name=last).count():
contestants = Artist.objects.filter(last_name=last)
except:
pass
if contestants:
choice = pick_among(contestants)
if choice:
return choice
artist = Artist(first_name=first, last_name=last)
artist.save()
return artist
def run():
with open('../data/manga-news/manga.csv') as f:
next(f)
artists = {}
hipsters = Counter()
for i, line in enumerate(f):
# print(len(line.split(';;')))
title, vo_title, writer, mangaka, editor, origin, genre1, genre2, manga_type, synopsis, poster = line.split(';;')
for artist in [writer, mangaka]:
if artist in artists:
continue
m = re.match('^([A-ZÔÛÏ\'-]+) (.*)$', writer)
if m:
last_name, first_name = m.groups()
last_name = last_name.lower().capitalize()
if not m:
first_name = ''
last_name = artist
if Artist.objects.filter(first_name=first_name, last_name=last_name).count() == 0:
a = Artist(first_name=first_name, last_name=last_name)
a.save()
else:
a = Artist.objects.get(first_name=first_name, last_name=last_name)
artists[artist] = a
with open('../data/manga-news/manga.csv') as f:
next(f)
for i, line in enumerate(f):
title, vo_title, writer, mangaka, editor, origin, genre1, genre2, manga_type, synopsis, poster = line.split(';;')
try:
if Manga.objects.filter(title=title, vo_title=vo_title).count() == 0:
manga = Manga(title=title, vo_title=vo_title, mangaka=artists[mangaka], writer=artists[writer], editor=editor, origin=origin.lower().replace('hong kong', 'hong-kong').replace('international', 'intl'), manga_type=manga_type.lower(), source='', poster=poster, synopsis=synopsis)
manga.save()
else:
manga = Manga.objects.get(title=title, vo_title=vo_title)
if genre1:
manga.genre.add(Genre.objects.get(title=genre1))
if genre2:
manga.genre.add(Genre.objects.get(title=genre2))
except IntegrityError as err:
print(line)
print(writer)
print(err)
break
except DataError as err:
print(line)
print(origin)
print(err)
break
except Genre.DoesNotExist as err:
print(line)
print('Genres: [%s] [%s]' % (genre1, genre2))
print(err)
break
def test_anidb_get_animes(self):
# Fake an artist entry with no AniDB creator ID that will be filled by retrieving Sangatsu
artist = Artist(name="Shinbou Akiyuki").save()
filenames = ['anidb/sangatsu_no_lion.xml', 'anidb/sangatsu_no_lion.xml', 'anidb/hibike_euphonium.xml']
with responses.RequestsMock(assert_all_requests_are_fired=True) as rsps:
for filename in filenames:
rsps.add(
responses.GET,
AniDB.BASE_URL,
body=self.read_fixture(filename),
status=200,
content_type='application/xml'
)
sangatsu = self.anidb.get_or_update_work(11606)
tags_sangatsu_from_anidb = self.anidb.get_tags(11606)
tags_diff_sangatsu = diff_between_anidb_and_local_tags(sangatsu, tags_sangatsu_from_anidb)
hibike = self.anidb.get_or_update_work(10889)
# Retrieve tags
tags_sangatsu = set(Work.objects.get(pk=sangatsu.pk).taggedwork_set.all().values_list('tag__title', flat=True))
tags_hibike = set(Work.objects.get(pk=hibike.pk).taggedwork_set.all().values_list('tag__title', flat=True))
shared_tags = tags_sangatsu.intersection(tags_hibike)
# Checks on tags
self.assertEqual(len(tags_sangatsu), 30)
self.assertEqual(len(tags_hibike), 38)
self.assertEqual(len(shared_tags), 18)
# Check for Sangatsu's informations
self.assertEqual(sangatsu.title, 'Sangatsu no Lion')
self.assertEqual(sangatsu.nb_episodes, 22)
self.assertEqual(sangatsu.studio.title, 'Shaft')
self.assertEqual(sangatsu.date, datetime(2016, 10, 8, 0, 0))
self.assertEqual(sangatsu.end_date, datetime(2017, 3, 18, 0, 0))
# Check for Sangatsu's staff
staff_sangatsu = Work.objects.get(pk=sangatsu.pk).staff_set.all().values_list('artist__name', flat=True)
self.assertCountEqual(staff_sangatsu, ['Umino Chika', 'Hashimoto Yukari', 'Shinbou Akiyuki', 'Okada Kenjirou'])
# Check retrieved tags from AniDB
self.assertEqual(len(tags_diff_sangatsu["deleted_tags"]), 0)
self.assertEqual(len(tags_diff_sangatsu["added_tags"]), 0)
self.assertEqual(len(tags_diff_sangatsu["updated_tags"]), 0)
self.assertEqual(len(tags_diff_sangatsu["kept_tags"]), len(tags_sangatsu))
# Check for no artist duplication
artist = Artist.objects.filter(name="Shinbou Akiyuki")
self.assertEqual(artist.count(), 1)
self.assertEqual(artist.first().anidb_creator_id, 59)
def test_insert_into_database(self):
self.add_fake_auth()
artist = Artist(name='Ishihara Tatsuya').save()
# Test insert AniListEntry into database
responses.add(
responses.GET,
urljoin(AniList.BASE_URL, 'browse/anime'),
body=self.read_fixture('anilist/airing_summer_2017_trimmed.json'),
status=200,
content_type='application/json')
seasonal = list(
self.anilist.list_seasonal_animes(year=2017, season='summer'))
self.assertEqual(
len(insert_works_into_database_from_anilist(seasonal)), 7)
# Test insert AniListRichEntry into database
responses.add(responses.GET,
urljoin(AniList.BASE_URL, 'anime/20912/page'),
body=self.read_fixture('anilist/hibike_euphonium.json'),
status=200,
content_type='application/json')
hibike_entry = self.anilist.get_work_by_id(AniListWorks.animes, 20912)
hibike = insert_work_into_database_from_anilist(hibike_entry)
titles_hibike = WorkTitle.objects.filter(work=hibike)
genres_hibike = hibike.genre.values_list('title', flat=True)
related_hibike = RelatedWork.objects.filter(parent_work=hibike)
staff_hibike = Work.objects.get(
pk=hibike.pk).staff_set.all().values_list('artist__name',
flat=True)
self.assertEqual(hibike.studio.title, 'Kyoto Animation')
self.assertEqual(len(titles_hibike), 3)
self.assertEqual(len(related_hibike), 4)
self.assertCountEqual(genres_hibike,
['Slice of Life', 'Music', 'Drama'])
self.assertCountEqual(
staff_hibike,
['Ishihara Tatsuya', 'Matsuda Akito', 'Takeda Ayano'])
# Check for no artist duplication
artist = Artist.objects.filter(name='Ishihara Tatsuya')
self.assertEqual(artist.count(), 1)
self.assertEqual(artist.first().anilist_creator_id, 100055)
# Try adding this work to the DB again
hibike_again = insert_work_into_database_from_anilist(hibike_entry)
self.assertEqual(hibike, hibike_again)
def build_staff(work: Work,
staff: List[AniListStaff]) -> List[Staff]:
"""
Insert Artist and Staff objects for a given Work when required into Mangaki's database.
:param work: a work
:param staff: a list of staff (and artists) informations
:type work: Work
:type staff: List[AniListStaff]
:return: a list of Staff objects that were inserted in Mangaki's database
:rtype: List[Staff]
"""
if not staff:
return []
anilist_roles_map = {
'Director': 'director',
'Music': 'composer',
'Original Creator': 'author'
}
artists_to_add = {}
artists = {}
for creator in staff:
name = '{} {}'.format(creator.name_last or '', creator.name_first or '').strip()
try: # This artist exists : prevent duplicates by updating with the AniList id
artist = Artist.objects.get(Q(name=name) | Q(anilist_creator_id=creator.id))
artist.name = name
artist.anilist_creator_id = creator.id
artist.save()
artists[creator.id] = artist
except Artist.DoesNotExist: # This artist does not yet exist : will be bulk created
artist = Artist(name=name, anilist_creator_id=creator.id)
artists_to_add[creator.id] = artist
artists.update(
{artist.anilist_creator_id: artist for artist in Artist.objects.bulk_create(list(artists_to_add.values()))}
)
artists_ids = [artist.id for artist in artists.values()]
existing_staff_artists = set(Staff.objects.filter(work=work, artist__in=artists_ids).values_list('work__id',
'artist__id',
'role'))
role_map = staff_roles.role_map
missing_staff = {creator.id: Staff(
work=work,
role=role_map.get(anilist_roles_map[creator.role]),
artist=artists[creator.id]
) for creator in staff if anilist_roles_map.get(creator.role)
and (work.id, artists[creator.id].id, role_map.get(anilist_roles_map[creator.role]).id) not in existing_staff_artists
}
Staff.objects.bulk_create(list(missing_staff.values()))
return list(missing_staff.values())
def run():
with open('../data/manga-news/anime.csv') as f:
next(f)
artists = {}
anime_data = []
for i, line in enumerate(f):
title, vo_title, studio, author, editor, anime_type, genre1, genre2, nb_episodes, origin, synopsis, poster = line.split(
';;')
anime_data.append(AnimeData(*line.split(';;')))
with open('../data/PAIR.log') as f:
for line in f:
mangaki_id = int(line.split('::')[0])
mn_id = int(line.split('::')[1])
anime = Anime.objects.get(id=mangaki_id)
data = anime_data[mn_id - 1]
anime.studio = create_if_not_exists(Studio, data.studio)
anime.editor = create_if_not_exists(Editor, data.editor)
m = re.match('^([A-ZÔÛÏ\'-]+) (.*)$', data.author)
if m:
last_name, first_name = m.groups()
last_name = last_name.lower().capitalize()
if not m:
first_name = ''
last_name = data.author
if Artist.objects.filter(first_name=first_name,
last_name=last_name).count() == 0:
print('NEW')
a = Artist(first_name=first_name, last_name=last_name)
a.save()
else:
a = Artist.objects.get(first_name=first_name,
last_name=last_name)
print('Exists', a)
anime.author = a
anime.anime_type = data.anime_type
genre1 = create_if_not_exists(Genre, data.genre1)
genre2 = create_if_not_exists(Genre, data.genre2)
anime.genre.add(genre1)
anime.genre.add(genre2)
anime.nb_episodes = data.nb_episodes
anime.origin = data.origin
anime.synopsis = data.synopsis # The most important!
anime.save()
def build_staff(work: Work, staff: List[AniListStaff]) -> List[Staff]:
"""
Insert Artist and Staff objects for a given Work when required into Mangaki's database.
:param work: a work
:param staff: a list of staff (and artists) informations
:type work: Work
:type staff: List[AniListStaff]
:return: a list of Staff objects that were inserted in Mangaki's database
:rtype: List[Staff]
"""
if not staff:
return []
anilist_roles_map = {
'Director': 'director',
'Music': 'composer',
'Original Creator': 'author'
}
artists_to_add = []
artists = []
for creator in staff:
name = '{} {}'.format(creator.name_last or '', creator.name_first
or '').strip()
try: # This artist exists : prevent duplicates by updating with the AniList id
artist = Artist.objects.get(
Q(name=name) | Q(anilist_creator_id=creator.id))
artist.name = name
artist.anilist_creator_id = creator.id
artist.save()
artists.append(artist)
except Artist.DoesNotExist: # This artist does not yet exist : will be bulk created
artist = Artist(name=name, anilist_creator_id=creator.id)
artists_to_add.append(artist)
artists.extend(Artist.objects.bulk_create(artists_to_add))
existing_staff_artists = set(
s.artist for s in Staff.objects.filter(work=work, artist__in=artists))
role_map = staff_roles.role_map
missing_staff = [
Staff(work=work,
role=role_map.get(anilist_roles_map[creator.role]),
artist=artists[index]) for index, creator in enumerate(staff)
if anilist_roles_map.get(creator.role)
and artists[index] not in existing_staff_artists
]
Staff.objects.bulk_create(missing_staff)
return missing_staff
def run():
with open("../data/manga-news/anime.csv") as f:
next(f)
artists = {}
anime_data = []
for i, line in enumerate(f):
title, vo_title, studio, author, editor, anime_type, genre1, genre2, nb_episodes, origin, synopsis, poster = line.split(
";;"
)
anime_data.append(AnimeData(*line.split(";;")))
with open("../data/PAIR.log") as f:
for line in f:
mangaki_id = int(line.split("::")[0])
mn_id = int(line.split("::")[1])
anime = Anime.objects.get(id=mangaki_id)
data = anime_data[mn_id - 1]
anime.studio = create_if_not_exists(Studio, data.studio)
anime.editor = create_if_not_exists(Editor, data.editor)
m = re.match("^([A-ZÔÛÏ'-]+) (.*)$", data.author)
if m:
last_name, first_name = m.groups()
last_name = last_name.lower().capitalize()
if not m:
first_name = ""
last_name = data.author
if Artist.objects.filter(first_name=first_name, last_name=last_name).count() == 0:
print("NEW")
a = Artist(first_name=first_name, last_name=last_name)
a.save()
else:
a = Artist.objects.get(first_name=first_name, last_name=last_name)
print("Exists", a)
anime.author = a
anime.anime_type = data.anime_type
genre1 = create_if_not_exists(Genre, data.genre1)
genre2 = create_if_not_exists(Genre, data.genre2)
anime.genre.add(genre1)
anime.genre.add(genre2)
anime.nb_episodes = data.nb_episodes
anime.origin = data.origin
anime.synopsis = data.synopsis # The most important!
anime.save()
def test_insert_into_database(self):
artist = Artist(name='Ishihara Tatsuya').save()
# Test insert AniListEntry into database
responses.add(responses.POST,
self.anilist.BASE_URL,
body=self.read_fixture('anilist/hibike_euphonium.json'),
status=200,
content_type='application/json')
hibike_entry = self.anilist.get_work(search_id=20912)
hibike = insert_work_into_database_from_anilist(hibike_entry,
build_related=False)
titles_hibike = WorkTitle.objects.filter(work=hibike).values_list(
'title', flat=True)
genres_hibike = hibike.genre.values_list('title', flat=True)
related_hibike = RelatedWork.objects.filter(parent_work=hibike)
staff_hibike = Work.objects.get(
pk=hibike.pk).staff_set.all().values_list('artist__name',
flat=True)
self.assertEqual(hibike.studio.title, 'Kyoto Animation')
self.assertCountEqual(
titles_hibike,
['Hibike! Euphonium', 'Sound! Euphonium', '響け！ユーフォニアム'])
self.assertCountEqual(genres_hibike,
['Slice of Life', 'Music', 'Drama'])
self.assertCountEqual(
staff_hibike,
['Ishihara Tatsuya', 'Matsuda Akito', 'Takeda Ayano'])
# Check for no artist duplication
artist = Artist.objects.filter(name='Ishihara Tatsuya')
self.assertEqual(artist.count(), 1)
self.assertEqual(artist.first().anilist_creator_id, 100055)
# Try adding this work to the DB again
hibike_again = insert_work_into_database_from_anilist(
hibike_entry, build_related=False)
self.assertEqual(hibike, hibike_again)
def build_staff(work: Work, staff: List[Tuple[AniListEntry,
str]]) -> List[Staff]:
anilist_roles_map = {
'Director': 'director',
'Music': 'composer',
'Original Creator': 'author'
}
artists_to_add = []
artists = []
for creator in staff:
name = '{} {}'.format(creator.name_last or '', creator.name_first
or '').strip()
try: # This artist exists : prevent duplicates by updating with the AniList id
artist = Artist.objects.get(
Q(name=name) | Q(anilist_creator_id=creator.id))
artist.name = name
artist.anilist_creator_id = creator.id
artist.save()
artists.append(artist)
except Artist.DoesNotExist: # This artist does not yet exist : will be bulk created
artist = Artist(name=name, anilist_creator_id=creator.id)
artists_to_add.append(artist)
artists.extend(Artist.objects.bulk_create(artists_to_add))
existing_staff_artists = set(
s.artist for s in Staff.objects.filter(work=work, artist__in=artists))
missing_staff = [
Staff(work=work,
role=role_map.get(anilist_roles_map[creator.role]),
artist=artists[index]) for index, creator in enumerate(staff)
if anilist_roles_map.get(creator.role)
and artists[index] not in existing_staff_artists
]
Staff.objects.bulk_create(missing_staff)
return missing_staff
def _build_staff(self,
work: Work,
creators: List[Dict[str, Any]],
reload_role_cache: bool = False) -> List[Staff]:
if reload_role_cache:
del self.role_map
processed_anidb_aids = []
artists_to_add = []
artists_list = []
for nc in creators:
artist = Artist.objects.filter(
Q(name=nc["name"])
| Q(anidb_creator_id=nc["anidb_creator_id"])).first()
if nc["anidb_creator_id"] in processed_anidb_aids: # Skip if this artist has more than one role
continue
if not artist: # This artist does not yet exist : will be bulk created
artist = Artist(name=nc["name"],
anidb_creator_id=nc["anidb_creator_id"])
artists_to_add.append(artist)
else: # This artist exists : prevent duplicates by updating with the AniDB id
artist.name = nc["name"]
artist.anidb_creator_id = nc["anidb_creator_id"]
artist.save()
artists_list.append(artist)
processed_anidb_aids.append(nc["anidb_creator_id"])
artists_list.extend(Artist.objects.bulk_create(artists_to_add))
artists = {artist.name: artist for artist in artists_list}
existing_staff = set(
Staff.objects.filter(
work=work,
role__in=(nc["role"] for nc in creators),
artist__name__in=(nc["name"] for nc in creators)).values_list(
'work', 'role', 'artist'))
missing_staff = [
Staff(work=work, role=nc["role"], artist=artists[nc["name"]])
for nc in creators
if (work.pk, nc["role"].pk,
artists[nc["name"]].pk) not in existing_staff
]
Staff.objects.bulk_create(missing_staff)
return missing_staff
def _build_staff(self,
work: Work,
creators: List[Dict[str, Any]],
reload_role_cache: bool = False) -> List[Staff]:
if reload_role_cache:
del self.role_map
artists_to_add = []
artists = []
for nc in creators:
artist = Artist.objects.filter(
Q(name=nc["name"])
| Q(anidb_creator_id=nc["anidb_creator_id"])).first()
if not artist: # This artist does not yet exist : will be bulk created
artist = Artist(name=nc["name"],
anidb_creator_id=nc["anidb_creator_id"])
artists_to_add.append(artist)
else: # This artist exists : prevent duplicates by updating with the AniDB id
artist.name = nc["name"]
artist.anidb_creator_id = nc["anidb_creator_id"]
artist.save()
artists.append(artist)
artists.extend(Artist.objects.bulk_create(artists_to_add))
staffs = []
for index, nc in enumerate(creators):
staffs.append(
Staff(work=work, role=nc["role"], artist=artists[index]))
existing_staff = set(
Staff.objects.filter(work=work,
role__in=[nc["role"] for nc in creators],
artist__in=[artist for artist in artists
]).values_list(
'work', 'role', 'artist'))
missing_staff = [
staff for staff in staffs
if (staff.work, staff.role, staff.artist) not in existing_staff
]
Staff.objects.bulk_create(missing_staff)
return missing_staff
def run():
with open('../data/manga-news/manga.csv') as f:
next(f)
artists = {}
hipsters = Counter()
for i, line in enumerate(f):
# print(len(line.split(';;')))
title, vo_title, writer, mangaka, editor, origin, genre1, genre2, manga_type, synopsis, poster = line.split(
';;')
for artist in [writer, mangaka]:
if artist in artists:
continue
m = re.match('^([A-ZÔÛÏ\'-]+) (.*)$', writer)
if m:
last_name, first_name = m.groups()
last_name = last_name.lower().capitalize()
if not m:
first_name = ''
last_name = artist
if Artist.objects.filter(first_name=first_name,
last_name=last_name).count() == 0:
a = Artist(first_name=first_name, last_name=last_name)
a.save()
else:
a = Artist.objects.get(first_name=first_name,
last_name=last_name)
artists[artist] = a
with open('../data/manga-news/manga.csv') as f:
next(f)
for i, line in enumerate(f):
title, vo_title, writer, mangaka, editor, origin, genre1, genre2, manga_type, synopsis, poster = line.split(
';;')
try:
if Manga.objects.filter(title=title,
vo_title=vo_title).count() == 0:
manga = Manga(title=title,
vo_title=vo_title,
mangaka=artists[mangaka],
writer=artists[writer],
editor=editor,
origin=origin.lower().replace(
'hong kong', 'hong-kong').replace(
'international', 'intl'),
manga_type=manga_type.lower(),
source='',
poster=poster,
synopsis=synopsis)
manga.save()
else:
manga = Manga.objects.get(title=title, vo_title=vo_title)
if genre1:
manga.genre.add(Genre.objects.get(title=genre1))
if genre2:
manga.genre.add(Genre.objects.get(title=genre2))
except IntegrityError as err:
print(line)
print(writer)
print(err)
break
except DataError as err:
print(line)
print(origin)
print(err)
break
except Genre.DoesNotExist as err:
print(line)
print('Genres: [%s] [%s]' % (genre1, genre2))
print(err)
break