def request(self,
path,
body=None,
method="GET",
identifier=None,
max_age=None):
path = self.full_path(path)
url = self.full_url(path)
if method == 'GET':
headers = {"Accept": "application/xml"}
else:
headers = {"Content-Type": "application/xml"}
self.sign(method, headers, path)
# print headers
# self.log.debug("3M request: %s %s", method, url)
if max_age and method == 'GET':
representation, cached = Representation.get(
self._db,
url,
extra_request_headers=headers,
do_get=self._simple_http_get,
max_age=max_age,
exception_handler=Representation.reraise_exception,
)
content = representation.content
return content
else:
return self._request_with_timeout(method,
url,
data=body,
headers=headers,
allow_redirects=False)
def get_library(self):
url = self.LIBRARY_ENDPOINT % dict(library_id=self.library_id)
representation, cached = Representation.get(
self._db, url, self.get,
exception_handler=Representation.reraise_exception,
)
return json.loads(representation.content)
def get_library(self):
"""Get basic information about the collection, including
a link to the titles in the collection.
"""
url = self._library_endpoint
representation, cached = Representation.get(
self._db, url, self.get,
exception_handler=Representation.reraise_exception,
)
return json.loads(representation.content)
def main():
FORMAT = '%(asctime)s %(levelname)s %(message)s'
logging.basicConfig(format=FORMAT)
logging.getLogger().setLevel(logging.INFO)
args = parse_args()
lang_map = {i: fn for i, fn in enumerate(sorted(listdir(args.lang_map)))}
if args.train.endswith('.mtx'):
mtx = mmread(args.train).todense()
t_mtx = mmread(args.test).todense()
else:
with open(args.train) as stream:
mtx = np.loadtxt(stream, np.float64)
with open(args.test) as stream:
t_mtx = np.loadtxt(stream, np.float64)
labels = np.ravel(mtx[:, 0])
test_labels = t_mtx[:, 0]
test_mtx = t_mtx[:, 1:]
if args.scale:
train = scale(mtx[:, 1:], with_mean=False)
else:
train = mtx[:, 1:]
kwargs = {}
for a in args.params:
k, v = a.split('=')
try:
v = int(v)
except:
pass
kwargs[k] = v
r = Representation(args.encoder, args.classifier, **kwargs)
r.encode(train)
logging.info('Matrix encoded')
r.train_classifier(labels)
logging.info('Model trained')
acc = 0
N = 0
for vec_ in test_mtx:
vec = np.ravel(vec_)
cl = r.classify_vector(vec, with_probs=args.with_probs)
try:
lab = test_labels[N, 0]
except IndexError:
lab = test_labels[N]
N += 1
if args.with_probs:
guess = max(enumerate(cl[0, :]), key=lambda x: x[1])[0]
print('{0}\t{1}\t{2}'.format('\t'.join(map(str, cl[0, :])), lang_map[guess], lang_map[int(lab)]))
else:
try:
guess = int(cl[0, 0])
except IndexError:
guess = int(cl + 0.5)
print('{0}\t{1}'.format(lang_map[guess], lang_map[int(lab)]))
if int(guess) == int(lab):
acc += 1
def request(self, path, identifier=None, max_age=LIST_MAX_AGE):
if not path.startswith(self.BASE_URL):
if not path.startswith("/"):
path = "/" + path
url = self.BASE_URL + path
else:
url = path
joiner = '?'
if '?' in url:
joiner = '&'
url += joiner + "api-key=" + self.api_key
representation, cached = Representation.get(self._db,
url,
do_get=self.do_get,
max_age=max_age,
debug=True,
pause_before=0.1)
content = json.loads(representation.content)
return content
def get_advantage_accounts(self):
"""Find all the Overdrive Advantage accounts managed by this library.
:yield: A sequence of OverdriveAdvantageAccount objects.
"""
library = self.get_library()
links = library.get('links', {})
advantage = links.get('advantageAccounts')
if not advantage:
return []
if advantage:
# This library has Overdrive Advantage accounts, or at
# least a link where some may be found.
advantage_url = advantage.get('href')
if not advantage_url:
return
representation, cached = Representation.get(
self._db, advantage_url, self.get,
exception_handler=Representation.reraise_exception,
)
return OverdriveAdvantageAccount.from_representation(
representation.content
)
def book_info_to_metadata(cls,
book,
include_bibliographic=True,
include_formats=True):
"""Turn Overdrive's JSON representation of a book into a Metadata
object.
Note: The json data passed into this method is from a different file/stream
from the json data that goes into the book_info_to_circulation() method.
"""
if not 'id' in book:
return None
overdrive_id = book['id']
primary_identifier = IdentifierData(Identifier.OVERDRIVE_ID,
overdrive_id)
if include_bibliographic:
title = book.get('title', None)
sort_title = book.get('sortTitle')
subtitle = book.get('subtitle', None)
series = book.get('series', None)
publisher = book.get('publisher', None)
imprint = book.get('imprint', None)
if 'publishDate' in book:
published = datetime.datetime.strptime(
book['publishDate'][:10], cls.DATE_FORMAT)
else:
published = None
languages = [l['code'] for l in book.get('languages', [])]
if 'eng' in languages or not languages:
language = 'eng'
else:
language = sorted(languages)[0]
contributors = []
for creator in book.get('creators', []):
sort_name = creator['fileAs']
display_name = creator['name']
role = creator['role']
roles = cls.parse_roles(overdrive_id,
role) or [Contributor.UNKNOWN_ROLE]
contributor = ContributorData(sort_name=sort_name,
display_name=display_name,
roles=roles,
biography=creator.get(
'bioText', None))
contributors.append(contributor)
subjects = []
for sub in book.get('subjects', []):
subject = SubjectData(type=Subject.OVERDRIVE,
identifier=sub['value'],
weight=100)
subjects.append(subject)
for sub in book.get('keywords', []):
subject = SubjectData(type=Subject.TAG,
identifier=sub['value'],
weight=1)
subjects.append(subject)
extra = dict()
if 'grade_levels' in book:
# n.b. Grade levels are measurements of reading level, not
# age appropriateness. We can use them as a measure of age
# appropriateness in a pinch, but we weight them less
# heavily than other information from Overdrive.
for i in book['grade_levels']:
subject = SubjectData(type=Subject.GRADE_LEVEL,
identifier=i['value'],
weight=10)
subjects.append(subject)
overdrive_medium = book.get('mediaType', None)
if overdrive_medium and overdrive_medium not in cls.overdrive_medium_to_simplified_medium:
cls.log.error("Could not process medium %s for %s",
overdrive_medium, overdrive_id)
medium = cls.overdrive_medium_to_simplified_medium.get(
overdrive_medium, Edition.BOOK_MEDIUM)
measurements = []
if 'awards' in book:
extra['awards'] = book.get('awards', [])
num_awards = len(extra['awards'])
measurements.append(
MeasurementData(Measurement.AWARDS, str(num_awards)))
for name, subject_type in (('ATOS', Subject.ATOS_SCORE),
('lexileScore', Subject.LEXILE_SCORE),
('interestLevel',
Subject.INTEREST_LEVEL)):
if not name in book:
continue
identifier = str(book[name])
subjects.append(
SubjectData(type=subject_type,
identifier=identifier,
weight=100))
for grade_level_info in book.get('gradeLevels', []):
grade_level = grade_level_info.get('value')
subjects.append(
SubjectData(type=Subject.GRADE_LEVEL,
identifier=grade_level,
weight=100))
identifiers = []
links = []
for format in book.get('formats', []):
for new_id in format.get('identifiers', []):
t = new_id['type']
v = new_id['value']
orig_v = v
type_key = None
if t == 'ASIN':
type_key = Identifier.ASIN
elif t == 'ISBN':
type_key = Identifier.ISBN
if len(v) == 10:
v = isbnlib.to_isbn13(v)
if v is None or not isbnlib.is_isbn13(v):
# Overdrive sometimes uses invalid values
# like "n/a" as placeholders. Ignore such
# values to avoid a situation where hundreds of
# books appear to have the same ISBN. ISBNs
# which fail check digit checks or are invalid
# also can occur. Log them for review.
cls.log.info("Bad ISBN value provided: %s", orig_v)
continue
elif t == 'DOI':
type_key = Identifier.DOI
elif t == 'UPC':
type_key = Identifier.UPC
elif t == 'PublisherCatalogNumber':
continue
if type_key and v:
identifiers.append(IdentifierData(type_key, v, 1))
# Samples become links.
if 'samples' in format:
if not format['id'] in cls.format_data_for_overdrive_format:
# Useless to us.
continue
content_type, drm_scheme = cls.format_data_for_overdrive_format.get(
format['id'])
if Representation.is_media_type(content_type):
for sample_info in format['samples']:
href = sample_info['url']
links.append(
LinkData(rel=Hyperlink.SAMPLE,
href=href,
media_type=content_type))
# A cover and its thumbnail become a single LinkData.
if 'images' in book:
images = book['images']
image_data = cls.image_link_to_linkdata(
images.get('cover'), Hyperlink.IMAGE)
for name in ['cover300Wide', 'cover150Wide', 'thumbnail']:
# Try to get a thumbnail that's as close as possible
# to the size we use.
image = images.get(name)
thumbnail_data = cls.image_link_to_linkdata(
image, Hyperlink.THUMBNAIL_IMAGE)
if not image_data:
image_data = cls.image_link_to_linkdata(
image, Hyperlink.IMAGE)
if thumbnail_data:
break
if image_data:
if thumbnail_data:
image_data.thumbnail = thumbnail_data
links.append(image_data)
# Descriptions become links.
short = book.get('shortDescription')
full = book.get('fullDescription')
if full:
links.append(
LinkData(
rel=Hyperlink.DESCRIPTION,
content=full,
media_type="text/html",
))
if short and (not full or not full.startswith(short)):
links.append(
LinkData(
rel=Hyperlink.SHORT_DESCRIPTION,
content=short,
media_type="text/html",
))
# Add measurements: rating and popularity
if book.get('starRating') is not None and book['starRating'] > 0:
measurements.append(
MeasurementData(quantity_measured=Measurement.RATING,
value=book['starRating']))
if book.get('popularity'):
measurements.append(
MeasurementData(quantity_measured=Measurement.POPULARITY,
value=book['popularity']))
metadata = Metadata(
data_source=DataSource.OVERDRIVE,
title=title,
subtitle=subtitle,
sort_title=sort_title,
language=language,
medium=medium,
series=series,
publisher=publisher,
imprint=imprint,
published=published,
primary_identifier=primary_identifier,
identifiers=identifiers,
subjects=subjects,
contributors=contributors,
measurements=measurements,
links=links,
)
else:
metadata = Metadata(
data_source=DataSource.OVERDRIVE,
primary_identifier=primary_identifier,
)
if include_formats:
formats = []
for format in book.get('formats', []):
format_id = format['id']
if format_id in cls.format_data_for_overdrive_format:
content_type, drm_scheme = cls.format_data_for_overdrive_format.get(
format_id)
formats.append(FormatData(content_type, drm_scheme))
elif format_id not in cls.ignorable_overdrive_formats:
cls.log.error(
"Could not process Overdrive format %s for %s",
format_id, overdrive_id)
# Also make a CirculationData so we can write the formats,
circulationdata = CirculationData(
data_source=DataSource.OVERDRIVE,
primary_identifier=primary_identifier,
formats=formats,
)
metadata.circulation = circulationdata
return metadata
def _do_get(self, url, headers):
"""This method is overridden in MockOverdriveAPI."""
return Representation.simple_http_get(url, headers)
def _simple_http_get(self, url, headers, *args, **kwargs):
"""This will be overridden in MockThreeMAPI."""
return Representation.simple_http_get(url, headers, *args, **kwargs)
*build_CURL_dataset(X_train, y_train, CURL_TRAIN_SIZE))
assert len(train_CURL) == CURL_TRAIN_SIZE
test_CURL = ContrastiveDataset(
*build_CURL_dataset(X_test, y_test, CURL_TEST_SIZE))
train_data = GMMDataset(X_train, y_train)
test_data = GMMDataset(X_test, y_test)
train_loader = DataLoader(train_data, shuffle=True, batch_size=BATCH_SIZE)
test_loader = DataLoader(test_data, shuffle=False, batch_size=BATCH_SIZE)
curl_train_loader = DataLoader(train_CURL, shuffle=True, batch_size=BATCH_SIZE)
curl_test_loader = DataLoader(test_CURL, shuffle=False, batch_size=BATCH_SIZE)
# Model
curl_model = Representation(INPUT_DIM, HIDDEN_DIM, OUT_DIM)
sup_model = ClassificationNet(Representation(INPUT_DIM, HIDDEN_DIM, OUT_DIM),
FCLayer(OUT_DIM, N_CENTERS))
writer_str = ("CURL/GMM-" + datetime.datetime.now().strftime("%Y%m%d-%H%M%S") +
"-" + str(INPUT_DIM) + "-" + str(HIDDEN_DIM) + "-" +
str(OUT_DIM) + "-" + str(LR) + "-" + str(BATCH_SIZE))
writer = SummaryWriter(writer_str)
"""
sup_model = train_multiclass_sup(
train_loader,
sup_model,
writer,
N_EPOCH,
LR,
verbose=True,
)
test_data = MNIST(
ROOT_DIR, train=False, download=DOWNLOAD, transform=transforms.ToTensor()
)
X_train, y_train = MNIST_pre_processing(train_data)
X_test, y_test = MNIST_pre_processing(test_data)
train_CURL = ContrastiveDataset(*build_CURL_dataset(X_train, y_train, CURL_TRAIN_SIZE))
test_CURL = ContrastiveDataset(*build_CURL_dataset(X_test, y_test, CURL_TRAIN_SIZE))
curl_train_loader = DataLoader(train_CURL, shuffle=True, batch_size=BATCH_SIZE)
curl_test_loader = DataLoader(train_CURL, shuffle=True, batch_size=BATCH_SIZE)
curl_model = Representation(INPUT_DIM, HIDDEN_DIM, OUT_DIM)
writer_str = (
"CURL/MNIST-CURL-"
+ datetime.datetime.now().strftime("%Y%m%d-%H%M%S")
+ "-"
+ str(INPUT_DIM)
+ "-"
+ str(HIDDEN_DIM)
+ "-"
+ str(OUT_DIM)
+ "-"
+ str(LR)
+ "-"
+ str(BATCH_SIZE)
def main():
r = Representation('pca', 'naive_bayes', dimension=3)
raw_mtx = numpy.array([[1, 1, 1, 0, 0, 0], [1, 0, 1, 0, 0, 0], [1, 1, 1, 0, 0, 0], [0, 0, 1, 1, 1, 0], [0, 0, 1, 1, 0, 0], [0, 0, 1, 1, 1, 0]])
r.encode(raw_mtx)
r.train_classifier([0, 0, 0, 1, 1, 1])
print r.classify_vector([1, 2, 1, 0, 1, 0])
