def test_get_buckets(self):
# When no buckets have been set, it raises an error.
assert_raises_regexp(
CannotLoadConfiguration,
'have not been initialized and no database session',
S3Uploader.get_bucket, S3Uploader.OA_CONTENT_BUCKET_KEY)
# So let's use an ExternalIntegration to set some buckets.
integration = self._external_integration(
ExternalIntegration.S3,
goal=ExternalIntegration.STORAGE_GOAL,
username='******',
password='******',
settings={
S3Uploader.OA_CONTENT_BUCKET_KEY: 'banana',
S3Uploader.BOOK_COVERS_BUCKET_KEY: 'bucket'
})
# If an object from the database is given, the buckets
# will be initialized, even though they hadn't been yet.
identifier = self._identifier()
result = S3Uploader.get_bucket(S3Uploader.OA_CONTENT_BUCKET_KEY,
sessioned_object=identifier)
eq_('banana', result)
# Generating the S3Uploader from_config also gives us S3 buckets.
S3Uploader.__buckets__ = S3Uploader.UNINITIALIZED_BUCKETS
S3Uploader.from_config(self._db)
eq_('bucket', S3Uploader.get_bucket(S3Uploader.BOOK_COVERS_BUCKET_KEY))
# Despite our new buckets, if a requested bucket isn't set,
# an error ir raised.
assert_raises_regexp(CannotLoadConfiguration, 'No S3 bucket found',
S3Uploader.get_bucket, 'nonexistent_bucket_key')
def test_cover_image_root(self):
with self.temp_config():
gutenberg_illustrated = DataSource.lookup(
self._db, DataSource.GUTENBERG_COVER_GENERATOR)
overdrive = DataSource.lookup(
self._db, DataSource.OVERDRIVE)
eq_("http://s3.amazonaws.com/test-book-covers-s3-bucket/Gutenberg%20Illustrated/",
S3Uploader.cover_image_root(gutenberg_illustrated))
eq_("http://s3.amazonaws.com/test-book-covers-s3-bucket/Overdrive/",
S3Uploader.cover_image_root(overdrive))
eq_("http://s3.amazonaws.com/test-book-covers-s3-bucket/scaled/300/Overdrive/",
S3Uploader.cover_image_root(overdrive, 300))
def __init__(self, _db, default_data_source, **kwargs):
kwargs = dict(kwargs)
if 'mirror' not in kwargs:
kwargs['mirror'] = S3Uploader()
super(OPDSImporterWithS3Mirror, self).__init__(
_db, default_data_source, **kwargs
)
def test_from_config(self):
# If there's no configuration for S3, S3Uploader.from_config
# raises an exception.
assert_raises_regexp(CannotLoadConfiguration,
'Required S3 integration is not configured',
S3Uploader.from_config, self._db)
# If there is a configuration but it's misconfigured, an error
# is raised.
integration = self._external_integration(
ExternalIntegration.S3, goal=ExternalIntegration.STORAGE_GOAL)
assert_raises_regexp(CannotLoadConfiguration,
'without both access_key and secret_key',
S3Uploader.from_config, self._db)
# Otherwise, it builds just fine.
integration.username = '******'
integration.password = '******'
uploader = S3Uploader.from_config(self._db)
eq_(True, isinstance(uploader, S3Uploader))
# Well, unless there are multiple S3 integrations, and it
# doesn't know which one to choose!
duplicate = self._external_integration(ExternalIntegration.S3)
duplicate.goal = ExternalIntegration.STORAGE_GOAL
assert_raises_regexp(CannotLoadConfiguration,
'Multiple S3 ExternalIntegrations configured',
S3Uploader.from_config, self._db)
def test_automatic_conversion_while_mirroring(self):
edition, pool = self._edition(with_license_pool=True)
original = self._url
# Create an SVG cover for the book.
svg = """<!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1//EN"
"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd">
<svg xmlns="http://www.w3.org/2000/svg" width="100" height="50">
<ellipse cx="50" cy="25" rx="50" ry="25" style="fill:blue;"/>
</svg>"""
hyperlink, ignore = pool.add_link(Hyperlink.IMAGE,
original,
edition.data_source,
Representation.SVG_MEDIA_TYPE,
content=svg)
# 'Upload' it to S3.
s3pool = MockS3Pool()
s3 = S3Uploader(None, None, pool=s3pool)
s3.mirror_one(hyperlink.resource.representation)
[[filename, data, bucket, media_type, ignore]] = s3pool.uploads
# The thing that got uploaded was a PNG, not the original SVG
# file.
eq_(Representation.PNG_MEDIA_TYPE, media_type)
assert 'PNG' in data
assert 'svg' not in data
def cdnify(url, cdns):
if not cdns:
# No CDNs configured
return url
scheme, netloc, path, query, fragment = urlparse.urlsplit(url)
if netloc == 's3.amazonaws.com':
# This is a URL like "http://s3.amazonaws.com/bucket/foo".
# It's equivalent to "http://bucket/foo".
# i.e. treat the bucket name as the netloc.
bucket, path = S3Uploader.bucket_and_filename(url)
netloc = bucket
if netloc not in cdns:
# This domain name is not covered by any of our CDNs.
return url
cdn_host = cdns[netloc]
cdn_scheme, cdn_netloc, i1, i2, i3 = urlparse.urlsplit(cdn_host)
return urlparse.urlunsplit((cdn_scheme, cdn_netloc, path, query, fragment))
def cdnify(url, cdns=None):
"""Turn local URLs into CDN URLs"""
try:
cdns = cdns or Configuration.cdns()
except CannotLoadConfiguration, e:
pass
if not cdns:
# No CDNs configured
return url
scheme, netloc, path, query, fragment = urlparse.urlsplit(url)
if netloc == 's3.amazonaws.com':
# This is a URL like "http://s3.amazonaws.com/bucket/foo".
# It's equivalent to "http://bucket/foo".
# i.e. treat the bucket name as the netloc.
#
# Since we are using the 'filename' to generate a URL rather
# than talk to S3, we don't want it to be unquoted.
#
netloc, path = S3Uploader.bucket_and_filename(url, unquote=False)
if netloc not in cdns:
# This domain name is not covered by any of our CDNs.
return url
cdn_host = cdns[netloc]
cdn_scheme, cdn_netloc, i1, i2, i3 = urlparse.urlsplit(cdn_host)
return urlparse.urlunsplit((cdn_scheme, cdn_netloc, path, query, fragment))
def test_content_root(self):
bucket = u'test-open-access-s3-bucket'
eq_("http://s3.amazonaws.com/test-open-access-s3-bucket/",
S3Uploader.content_root(bucket))
def test_custom_client_class(self):
"""You can specify a client class to use instead of boto3.client."""
integration = self._integration()
uploader = S3Uploader(integration, MockS3Client)
assert isinstance(uploader.client, MockS3Client)
def test_key_join(self):
"""Test the code used to build S3 keys from parts."""
parts = ["Gutenberg", "Gutenberg ID", 1234, "Die Flügelmaus.epub"]
eq_('Gutenberg/Gutenberg+ID/1234/Die+Fl%C3%BCgelmaus.epub',
S3Uploader.key_join(parts))
def test_content_root(self):
with self.temp_config():
eq_("http://s3.amazonaws.com/test-open-access-s3-bucket/",
S3Uploader.content_root())
def cdnify(url, cdns=None):
"""Turn local URLs into CDN URLs"""
try:
cdns = cdns or Configuration.cdns()
except CannotLoadConfiguration, e:
pass
if not cdns:
# No CDNs configured
return url
scheme, netloc, path, query, fragment = urlparse.urlsplit(url)
if netloc == 's3.amazonaws.com':
# This is a URL like "http://s3.amazonaws.com/bucket/foo".
# It's equivalent to "http://bucket/foo".
# i.e. treat the bucket name as the netloc.
#
# Since we are using the 'filename' to generate a URL rather
# than talk to S3, we don't want it to be unquoted.
#
netloc, path = S3Uploader.bucket_and_filename(url, unquote=False)
if netloc not in cdns:
# This domain name is not covered by any of our CDNs.
return url
cdn_host = cdns[netloc]
cdn_scheme, cdn_netloc, i1, i2, i3 = urlparse.urlsplit(cdn_host)
return urlparse.urlunsplit((cdn_scheme, cdn_netloc, path, query, fragment))
from config import Configuration, CannotLoadConfiguration
from s3 import S3Uploader
def cdnify(url, cdns=None):
"""Turn local URLs into CDN URLs"""
try:
cdns = cdns or Configuration.cdns()
except CannotLoadConfiguration, e:
pass
if not cdns:
# No CDNs configured
return url
scheme, netloc, path, query, fragment = urlparse.urlsplit(url)
if netloc == 's3.amazonaws.com':
# This is a URL like "http://s3.amazonaws.com/bucket/foo".
# It's equivalent to "http://bucket/foo".
# i.e. treat the bucket name as the netloc.
bucket, path = S3Uploader.bucket_and_filename(url)
netloc = bucket
if netloc not in cdns:
# This domain name is not covered by any of our CDNs.
return url
cdn_host = cdns[netloc]
cdn_scheme, cdn_netloc, i1, i2, i3 = urlparse.urlsplit(cdn_host)
return urlparse.urlunsplit((cdn_scheme, cdn_netloc, path, query, fragment))
