def test_syntax_analyzer_returns_ast(self):
# Arrange
syntax_analyzer = RWPMSyntaxAnalyzer()
input_steam = six.StringIO(RWPM_MANIFEST)
manifest_json = ManifestParser.get_manifest_json(input_steam)
# Act
manifest = syntax_analyzer.analyze(manifest_json)
# Assert
self.assertIsInstance(manifest, RWPMManifest)
self.assertIsInstance(manifest.context, list)
self.assertEqual(1, len(manifest.context))
[context] = manifest.context
self.assertEqual(context,
"https://readium.org/webpub-manifest/context.jsonld")
self.assertIsInstance(manifest.metadata, Metadata)
self.assertEqual("http://schema.org/Book", manifest.metadata.type)
self.assertEqual("Moby-Dick", manifest.metadata.title)
self.assertEqual(
[Contributor(name="Herman Melville", roles=[], links=LinkList())],
manifest.metadata.authors,
)
self.assertEqual("urn:isbn:978031600000X",
manifest.metadata.identifier)
self.assertEqual(["en"], manifest.metadata.languages)
self.assertEqual(
datetime.datetime(2015, 9, 29, 17, 0, 0, tzinfo=tzutc()),
manifest.metadata.modified,
)
self.assertIsInstance(manifest.links, list)
self.assertEqual(1, len(manifest.links))
[link] = manifest.links
self.assertIsInstance(link.rels, list)
self.assertEqual(1, len(link.rels))
self.assertEqual(RWPMLinkRelationsRegistry.SELF.key, link.rels[0])
self.assertEqual("https://example.com/manifest.json", link.href)
self.assertEqual(RWPMMediaTypesRegistry.MANIFEST.key, link.type)
self.assertIsInstance(manifest.reading_order, CompactCollection)
self.assertIsInstance(manifest.reading_order.links, list)
self.assertEqual(1, len(manifest.reading_order.links))
[reading_order_link] = manifest.reading_order.links
self.assertEqual("https://example.com/c001.html",
reading_order_link.href)
self.assertEqual(RWPMMediaTypesRegistry.HTML.key,
reading_order_link.type)
self.assertEqual("Chapter 1", reading_order_link.title)
class RWPMSyntaxAnalyzerTest(TestCase):
@parameterized.expand([
(
"when_manifest_does_not_contain_metadata",
RWPM_MANIFEST_WITHOUT_METADATA,
RWPMManifest,
RWPMManifest.metadata.key,
),
(
"when_manifest_metadata_does_not_contain_title",
RWPM_MANIFEST_WITH_MISSING_METADATA_TITLE_PROPERTY,
PresentationMetadata,
PresentationMetadata.title.key,
),
(
"when_metadata_contributor_does_not_contain_name",
RWPM_MANIFEST_WITH_MISSING_CONTRIBUTOR_NAME_PROPERTY,
Contributor,
Contributor.name.key,
),
(
"when_metadata_subject_does_not_contain_name",
RWPM_MANIFEST_WITH_MISSING_SUBJECT_NAME_PROPERTY,
Subject,
Subject.name.key,
),
(
"when_manifest_does_not_contain_links",
RWPM_MANIFEST_WITHOUT_LINKS,
RWPMManifest,
RWPMManifest.links.key,
),
(
"when_manifest_link_does_not_contain_href",
RWPM_MANIFEST_WITH_MISSING_LINK_HREF_PROPERTY,
Link,
Link.href.key,
),
(
"when_manifest_link_does_not_contain_reading_order_sub_collection",
RWPM_MANIFEST_WITH_MISSING_READING_ORDER,
RWPMManifest,
RWPMManifest.reading_order.key,
),
])
def test_syntax_analyzer_raises_missing_property_error_correctly(
self,
_,
rwpm_manifest_content,
expected_class_with_missing_property,
expected_missing_property,
):
# Arrange
syntax_analyzer = RWPMSyntaxAnalyzer()
input_steam = six.StringIO(rwpm_manifest_content)
manifest_json = DocumentParser.get_manifest_json(input_steam)
# Act
with self.assertRaises(MissingPropertyError) as assert_raises_context:
syntax_analyzer.analyze(manifest_json)
# Assert
self.assertEqual(
expected_class_with_missing_property,
assert_raises_context.exception.cls,
)
self.assertEqual(
expected_missing_property,
assert_raises_context.exception.object_property.key,
)
@parameterized.expand([
(
"when_manifest_context_has_duplicated_items",
RWPM_MANIFEST_WINT_INCORRECT_CONTEXT,
"Item 'context1' is not unique",
),
(
"when_metadata_modified_property_has_incorrect_format",
RWPM_MANIFEST_WITH_INCORRECT_METADATA_MODIFIED_PROPERTY,
"'2015-09-29T17:00:00Z---' is not a 'date-time'",
),
(
"when_metadata_modified_language_has_incorrect_format",
RWPM_MANIFEST_WITH_INCORRECT_METADATA_LANGUAGE_PROPERTY,
"String value 'eng;dat' does not match regular expression ^((?P<grandfathered>(en-GB-oed|i-ami|i-bnn|i-default|i-enochian|i-hak|i-klingon|i-lux|i-mingo|i-navajo|i-pwn|i-tao|i-tay|i-tsu|sgn-BE-FR|sgn-BE-NL|sgn-CH-DE)|(art-lojban|cel-gaulish|no-bok|no-nyn|zh-guoyu|zh-hakka|zh-min|zh-min-nan|zh-xiang))|((?P<language>([A-Za-z]{2,3}(-(?P<extlang>[A-Za-z]{3}(-[A-Za-z]{3}){0,2}))?)|[A-Za-z]{4}|[A-Za-z]{5,8})(-(?P<script>[A-Za-z]{4}))?(-(?P<region>[A-Za-z]{2}|[0-9]{3}))?(-(?P<variant>[A-Za-z0-9]{5,8}|[0-9][A-Za-z0-9]{3}))*(-(?P<extension>[0-9A-WY-Za-wy-z](-[A-Za-z0-9]{2,8})+))*(-(?P<privateUse>x(-[A-Za-z0-9]{1,8})+))?)|(?P<privateUse2>x(-[A-Za-z0-9]{1,8})+))$",
),
(
"when_contributor_identifier_has_incorrect_format",
RWPM_MANIFEST_WITH_INCORRECT_CONTRIBUTOR_IDENTIFIER_PROPERTY,
"'x' is not a 'uri'",
),
(
"when_link_height_less_or_equal_than_the_exclusive_minimum",
RWPM_MANIFEST_WITH_INCORRECT_LINK_HEIGHT_PROPERTY,
"Value -10 is less or equal than the exclusive minimum (0)",
),
])
def test_syntax_analyzer_raises_value_parsing_error_when_property_has_incorrect_value(
self, _, rwpm_manifest_content, expected_error_message):
# Arrange
syntax_analyzer = RWPMSyntaxAnalyzer()
input_steam = six.StringIO(rwpm_manifest_content)
manifest_json = DocumentParser.get_manifest_json(input_steam)
# Act
with self.assertRaises(ValueParsingError) as assert_raises_context:
syntax_analyzer.analyze(manifest_json)
# Assert
self.assertEqual(
expected_error_message,
six.text_type(assert_raises_context.exception).strip("u"),
)
@parameterized.expand([
(
"when_manifest_has_author_encoded_as_string",
RWPM_MANIFEST_WITH_AUTHOR_ENCODED_AS_STRING,
[Contributor(name="Herman Melville", roles=[], links=LinkList())],
),
(
"when_manifest_has_authors_encoded_as_array_of_strings",
RWPM_MANIFEST_WITH_AUTHORS_ENCODED_AS_ARRAY_OF_STRINGS,
[
Contributor(name="Herman Melville 1",
roles=[],
links=LinkList()),
Contributor(name="Herman Melville 2",
roles=[],
links=LinkList()),
],
),
(
"when_manifest_has_author_encoded_as_object",
RWPM_MANIFEST_WITH_AUTHOR_ENCODED_AS_OBJECT,
[Contributor(name="Herman Melville", roles=[], links=LinkList())],
),
(
"when_manifest_has_authors_encoded_as_array_of_objects",
RWPM_MANIFEST_WITH_AUTHORS_ENCODED_AS_ARRAY_OF_OBJECTS,
[
Contributor(name="Herman Melville 1",
roles=[],
links=LinkList()),
Contributor(name="Herman Melville 2",
roles=[],
links=LinkList()),
],
),
(
"when_manifest_has_authors_encoded_as_mixed_array_of_strings_and_objects",
RWPM_MANIFEST_WITH_AUTHORS_ENCODED_AS_MIXED_ARRAY_OF_STRINGS_AND_OBJECTS,
[
Contributor(name="Herman Melville 1",
roles=[],
links=LinkList()),
Contributor(name="Herman Melville 2",
roles=[],
links=LinkList()),
],
),
])
def test_syntax_analyzer_correctly_parses_contributor_metadata_as_contributor_objects(
self, _, rwpm_manifest_content, expected_authors):
# Arrange
syntax_analyzer = RWPMSyntaxAnalyzer()
input_steam = six.StringIO(rwpm_manifest_content)
manifest_json = DocumentParser.get_manifest_json(input_steam)
# Act
manifest = syntax_analyzer.analyze(manifest_json)
# Assert
self.assertEqual(expected_authors, manifest.metadata.authors)
def test_syntax_analyzer_returns_ast(self):
# Arrange
syntax_analyzer = RWPMSyntaxAnalyzer()
input_steam = six.StringIO(RWPM_MANIFEST)
manifest_json = DocumentParser.get_manifest_json(input_steam)
# Act
manifest = syntax_analyzer.analyze(manifest_json)
# Assert
self.assertIsInstance(manifest, RWPMManifest)
self.assertIsInstance(manifest.context, list)
self.assertEqual(1, len(manifest.context))
[context] = manifest.context
self.assertEqual(context,
"https://readium.org/webpub-manifest/context.jsonld")
self.assertIsInstance(manifest.metadata, Metadata)
self.assertEqual("http://schema.org/Book", manifest.metadata.type)
self.assertEqual("Moby-Dick", manifest.metadata.title)
self.assertEqual(
[Contributor(name="Herman Melville", roles=[], links=LinkList())],
manifest.metadata.authors,
)
self.assertEqual("urn:isbn:978031600000X",
manifest.metadata.identifier)
self.assertEqual(["en"], manifest.metadata.languages)
self.assertEqual(datetime.datetime(2015, 9, 29, 17, 0, 0),
manifest.metadata.modified)
self.assertIsInstance(manifest.links, list)
self.assertEqual(1, len(manifest.links))
[link] = manifest.links
self.assertIsInstance(link.rels, list)
self.assertEqual(1, len(link.rels))
self.assertEqual(RWPMLinkRelationsRegistry.SELF.key, link.rels[0])
self.assertEqual("https://example.com/manifest.json", link.href)
self.assertEqual(RWPMMediaTypesRegistry.MANIFEST.key, link.type)
self.assertIsInstance(manifest.reading_order, CompactCollection)
self.assertIsInstance(manifest.reading_order.links, list)
self.assertEqual(1, len(manifest.reading_order.links))
[reading_order_link] = manifest.reading_order.links
self.assertEqual("https://example.com/c001.html",
reading_order_link.href)
self.assertEqual(RWPMMediaTypesRegistry.HTML.key,
reading_order_link.type)
self.assertEqual("Chapter 1", reading_order_link.title)
def test(self):
# Arrange
parser_factory = OPDS2FeedParserFactory()
parser = parser_factory.create()
input_file_path = os.path.join(os.path.dirname(__file__),
"../../files/opds2/feed.json")
# Act
result = parser.parse_file(input_file_path)
# Assert
self.assertIsInstance(result, ManifestParserResult)
self.assertEqual(0, len(result.errors))
feed = result.root
self.assertIsInstance(feed, OPDS2Feed)
self.assertIsInstance(feed.metadata, OPDS2FeedMetadata)
self.assertEqual("Example listing publications", feed.metadata.title)
self.assertIsInstance(feed.links, list)
self.assertEqual(1, len(feed.links))
[manifest_link] = feed.links
self.assertEqual(OPDS2LinkRelationsRegistry.SELF.key,
manifest_link.rels[0])
self.assertEqual("http://example.com/new", manifest_link.href)
self.assertEqual(OPDS2MediaTypesRegistry.OPDS_FEED.key,
manifest_link.type)
self.assertIsInstance(feed.publications, list)
self.assertEqual(2, len(feed.publications))
publication = feed.publications[0]
self.assertIsInstance(publication.metadata, PresentationMetadata)
self.assertEqual("http://schema.org/Book", publication.metadata.type)
self.assertEqual("Moby-Dick", publication.metadata.title)
self.assertEqual(
[Contributor(name="Herman Melville", roles=[], links=LinkList())],
publication.metadata.authors,
)
self.assertEqual("urn:isbn:978-3-16-148410-0",
publication.metadata.identifier)
self.assertEqual(["en"], publication.metadata.languages)
self.assertEqual(
datetime.datetime(2015, 9, 29, 17, 0, tzinfo=tzutc()),
publication.metadata.modified,
)
self.assertIsInstance(publication.links, list)
self.assertEqual(len(publication.links), 2)
publication_self_link = first_or_default(
publication.links.get_by_rel(OPDS2LinkRelationsRegistry.SELF.key))
self.assertEqual(OPDS2LinkRelationsRegistry.SELF.key,
publication_self_link.rels[0])
self.assertEqual("http://example.org/publication.json",
publication_self_link.href)
self.assertEqual(OPDS2MediaTypesRegistry.OPDS_PUBLICATION.key,
publication_self_link.type)
publication_acquisition_link = first_or_default(
publication.links.get_by_rel(
OPDS2LinkRelationsRegistry.OPEN_ACCESS.key))
self.assertEqual(
OPDS2LinkRelationsRegistry.OPEN_ACCESS.key,
publication_acquisition_link.rels[0],
)
self.assertEqual("http://example.org/file.epub",
publication_acquisition_link.href)
self.assertEqual(
OPDS2MediaTypesRegistry.EPUB_PUBLICATION_PACKAGE.key,
publication_acquisition_link.type,
)
self.assertIsInstance(publication.images, CompactCollection)
self.assertIsInstance(publication.images.links, list)
self.assertEqual(3, len(publication.images.links))
jpeg_cover_link = first_or_default(
publication.images.links.get_by_href(
"http://example.org/cover.jpg"))
self.assertEqual([], jpeg_cover_link.rels)
self.assertEqual("http://example.org/cover.jpg", jpeg_cover_link.href)
self.assertEqual(OPDS2MediaTypesRegistry.JPEG.key,
jpeg_cover_link.type)
self.assertEqual(1400, jpeg_cover_link.height)
self.assertEqual(800, jpeg_cover_link.width)
small_jpeg_cover_link = first_or_default(
publication.images.links.get_by_href(
"http://example.org/cover-small.jpg"))
self.assertEqual("http://example.org/cover-small.jpg",
small_jpeg_cover_link.href)
self.assertEqual(OPDS2MediaTypesRegistry.JPEG.key,
small_jpeg_cover_link.type)
self.assertEqual(700, small_jpeg_cover_link.height)
self.assertEqual(400, small_jpeg_cover_link.width)
svg_cover_link = first_or_default(
publication.images.links.get_by_href(
"http://example.org/cover.svg"))
self.assertEqual(svg_cover_link.href, "http://example.org/cover.svg")
self.assertEqual(svg_cover_link.type,
OPDS2MediaTypesRegistry.SVG_XML.key)
publication = feed.publications[1]
self.assertIsInstance(publication.metadata, PresentationMetadata)
self.assertEqual("http://schema.org/Book", publication.metadata.type)
self.assertEqual("Adventures of Huckleberry Finn",
publication.metadata.title)
self.assertEqual(
[
Contributor(name="Mark Twain", roles=[], links=LinkList()),
Contributor(name="Samuel Langhorne Clemens",
roles=[],
links=LinkList()),
],
publication.metadata.authors,
)
self.assertEqual("urn:isbn:978-3-16-148410-0",
publication.metadata.identifier)
self.assertEqual(["eng", "fre"], publication.metadata.languages)
self.assertEqual(
datetime.datetime(2015, 9, 29, 0, 0, tzinfo=tzutc()),
publication.metadata.published,
)
self.assertEqual(
datetime.datetime(2015, 9, 29, 17, 0, 0, tzinfo=tzutc()),
publication.metadata.modified,
)
self.assertIsInstance(publication.links, list)
publication_acquisition_link = first_or_default(
publication.links.get_by_rel(
OPDS2LinkRelationsRegistry.BORROW.key))
self.assertEqual(OPDS2LinkRelationsRegistry.BORROW.key,
publication_acquisition_link.rels[0])
self.assertEqual(
OPDS2MediaTypesRegistry.OPDS_PUBLICATION.key,
publication_acquisition_link.type,
)
link_properties = publication_acquisition_link.properties
self.assertIsInstance(link_properties, OPDS2LinkProperties)
self.assertEqual(OPDS2AvailabilityType.AVAILABLE.value,
link_properties.availability.state)
self.assertEqual(2, len(link_properties.indirect_acquisition))
indirect_acquisition_object = link_properties.indirect_acquisition[0]
self.assertEqual("application/vnd.adobe.adept+xml",
indirect_acquisition_object.type)
self.assertEqual(1, len(indirect_acquisition_object.child))
self.assertIsInstance(indirect_acquisition_object.child[0],
OPDS2AcquisitionObject)
self.assertEqual("application/epub+zip",
indirect_acquisition_object.child[0].type)
indirect_acquisition_object = link_properties.indirect_acquisition[1]
self.assertEqual(
"application/vnd.readium.lcp.license.v1.0+json",
indirect_acquisition_object.type,
)
self.assertEqual(1, len(indirect_acquisition_object.child))
self.assertIsInstance(indirect_acquisition_object.child[0],
OPDS2AcquisitionObject)
self.assertEqual("application/epub+zip",
indirect_acquisition_object.child[0].type)
def test(self):
# Arrange
parser_factory = RWPMDocumentParserFactory()
parser = parser_factory.create()
input_file_path = os.path.join(
os.path.dirname(__file__), "../../files/rwpm/spec_example.json"
)
# Act
manifest = parser.parse_file(input_file_path)
# Assert
self.assertIsInstance(manifest.context, list)
self.assertEqual(1, len(manifest.context))
[context] = manifest.context
self.assertEqual(context, "https://readium.org/webpub-manifest/context.jsonld")
self.assertIsInstance(manifest.metadata, Metadata)
self.assertEqual("http://schema.org/Book", manifest.metadata.type)
self.assertEqual("Moby-Dick", manifest.metadata.title)
self.assertEqual(
[Contributor(name="Herman Melville", roles=[], links=LinkList())],
manifest.metadata.authors,
)
self.assertEqual("urn:isbn:978031600000X", manifest.metadata.identifier)
self.assertEqual(["en"], manifest.metadata.languages)
self.assertEqual(
datetime.datetime(2015, 9, 29, 17, 0, 0), manifest.metadata.modified
)
self.assertIsInstance(manifest.links, list)
self.assertEqual(3, len(manifest.links))
self_link = first_or_default(
manifest.links.get_by_rel(RWPMLinkRelationsRegistry.SELF.key)
)
self.assertIsNotNone(self_link)
self.assertIn(RWPMLinkRelationsRegistry.SELF.key, self_link.rels)
self.assertEqual("https://example.com/manifest.json", self_link.href)
self.assertEqual(RWPMMediaTypesRegistry.MANIFEST.key, self_link.type)
alternate_link = first_or_default(
manifest.links.get_by_rel(RWPMLinkRelationsRegistry.ALTERNATE.key)
)
self.assertIsNotNone(alternate_link)
self.assertIn(RWPMLinkRelationsRegistry.ALTERNATE.key, alternate_link.rels)
self.assertEqual("https://example.com/publication.epub", alternate_link.href)
self.assertEqual(
RWPMMediaTypesRegistry.EPUB_PUBLICATION_PACKAGE.key, alternate_link.type
)
search_link = first_or_default(
manifest.links.get_by_rel(RWPMLinkRelationsRegistry.SEARCH.key)
)
self.assertIsNotNone(search_link)
self.assertIn(RWPMLinkRelationsRegistry.SEARCH.key, search_link.rels)
self.assertEqual("https://example.com/search{?query}", search_link.href)
self.assertEqual(RWPMMediaTypesRegistry.HTML.key, search_link.type)
self.assertIsInstance(manifest.reading_order, CompactCollection)
self.assertIsInstance(manifest.reading_order.links, list)
self.assertEqual(2, len(manifest.reading_order.links))
reading_order_link = manifest.reading_order.links[0]
self.assertEqual("https://example.com/c001.html", reading_order_link.href)
self.assertEqual(RWPMMediaTypesRegistry.HTML.key, reading_order_link.type)
self.assertEqual("Chapter 1", reading_order_link.title)
reading_order_link = manifest.reading_order.links[1]
self.assertEqual("https://example.com/c002.html", reading_order_link.href)
self.assertEqual(RWPMMediaTypesRegistry.HTML.key, reading_order_link.type)
self.assertEqual("Chapter 2", reading_order_link.title)
resources_sub_collection = manifest.resources
self.assertEqual(5, len(resources_sub_collection.links))
self.assertEqual(
[RWPMLinkRelationsRegistry.COVER.key],
resources_sub_collection.links[0].rels,
)
self.assertEqual(
"https://example.com/cover.jpg", resources_sub_collection.links[0].href
)
self.assertEqual(
RWPMMediaTypesRegistry.JPEG.key, resources_sub_collection.links[0].type
)
self.assertEqual(600, resources_sub_collection.links[0].height)
self.assertEqual(400, resources_sub_collection.links[0].width)
self.assertEqual(
"https://example.com/style.css", resources_sub_collection.links[1].href
)
self.assertEqual(
RWPMMediaTypesRegistry.CSS.key, resources_sub_collection.links[1].type
)
self.assertEqual(
"https://example.com/whale.jpg", resources_sub_collection.links[2].href
)
self.assertEqual(
RWPMMediaTypesRegistry.JPEG.key, resources_sub_collection.links[2].type
)
self.assertEqual(
"https://example.com/boat.svg", resources_sub_collection.links[3].href
)
self.assertEqual(
RWPMMediaTypesRegistry.SVG_XML.key, resources_sub_collection.links[3].type
)
self.assertEqual(
"https://example.com/notes.html", resources_sub_collection.links[4].href
)
self.assertEqual(
RWPMMediaTypesRegistry.HTML.key, resources_sub_collection.links[4].type
)