def _publisher_graph(self, dataset_ref, dataset_dict):
if any([
self._get_dataset_value(dataset_dict, 'publisher_uri'),
self._get_dataset_value(dataset_dict, 'publisher_name'),
dataset_dict.get('organization'),
]):
publisher_uri = publisher_uri_from_dataset_dict(dataset_dict)
if publisher_uri:
publisher_details = URIRef(publisher_uri)
else:
# No organization nor publisher_uri
publisher_details = BNode()
self.g.add((publisher_details, RDF.type, SCHEMA.Organization))
self.g.add((dataset_ref, SCHEMA.publisher, publisher_details))
self.g.add((dataset_ref, SCHEMA.sourceOrganization,
publisher_details)) # noqa
publisher_name = self._get_dataset_value(dataset_dict,
'publisher_name')
if not publisher_name and dataset_dict.get('organization'):
publisher_name = dataset_dict['organization']['title']
self._add_multilang_value(publisher_details,
SCHEMA.name,
multilang_values=publisher_name)
else:
g.add((publisher_details, SCHEMA.name,
Literal(publisher_name))) # noqa
contact_point = BNode()
self.g.add((publisher_details, SCHEMA.contactPoint, contact_point))
self.g.add((contact_point, SCHEMA.contactType,
Literal('customer service'))) # noqa
publisher_url = self._get_dataset_value(dataset_dict,
'publisher_url') # noqa
if not publisher_url and dataset_dict.get('organization'):
publisher_url = dataset_dict['organization'].get(
'url') or config.get('ckan.site_url', '') # noqa
self.g.add((contact_point, SCHEMA.url, Literal(publisher_url)))
items = [
('publisher_email', SCHEMA.email,
['contact_email', 'maintainer_email',
'author_email'], Literal), # noqa
('publisher_name', SCHEMA.name,
['contact_name', 'maintainer', 'author'], Literal), # noqa
]
self._add_triples_from_dict(dataset_dict, contact_point, items)
def graph_from_dataset(self, dataset_dict, dataset_ref):
g = self.g
for prefix, namespace in namespaces.iteritems():
g.bind(prefix, namespace)
# -- start
g.add((dataset_ref, RDF.type, DCAT.Dataset))
# Basic fields
items = [
('title', DCT.title, None, Literal),
('notes', DCT.description, None, Literal),
('url', DCAT.landingPage, None, URIRef),
('identifier', DCT.identifier, ['guid', 'id'], Literal),
('version', OWL.versionInfo, ['dcat_version'], Literal),
('version_notes', ADMS.versionNotes, None, Literal),
('frequency', DCT.accrualPeriodicity, None, URIRef),
('subject', DCT.subject, None,
URIRef), # Mentioned in the vocabulary
('provenance', DCT.provenance, None, URIRef)
]
self._add_triples_from_dict(dataset_dict, dataset_ref, items)
# Tags
for tag in dataset_dict.get('tags', []):
g.add((dataset_ref, DCAT.keyword, Literal(tag['name'])))
# Dates
items = [
('issued', DCT.issued, ['metadata_created'], Literal),
('modified', DCT.modified, ['metadata_modified'], Literal),
]
self._add_date_triples_from_dict(dataset_dict, dataset_ref, items)
# Lists
items = [('language', DCT.language, None, URIRef),
('theme', DCAT.theme, None, URIRef),
('spatial_uri', DCT.spatial, None, URIRef),
('conforms_to', DCT.conformsTo, None, URIRef),
('alternate_identifier', ADMS.identifier, None, Literal),
('documentation', FOAF.page, None, URIRef),
('access_rights', DCT.accessRights, None, URIRef),
('related_resource', DCT.relation, None, URIRef),
('has_version', DCT.hasVersion, None, Literal),
('is_version_of', DCT.isVersionOf, None, Literal),
('source', DCT.source, None, Literal),
('sample', ADMS.sample, None, Literal)]
self._add_list_triples_from_dict(dataset_dict, dataset_ref, items)
# Contact details
if any([
self._get_dataset_value(dataset_dict, 'contact_uri'),
self._get_dataset_value(dataset_dict, 'contact_name'),
self._get_dataset_value(dataset_dict, 'contact_email'),
self._get_dataset_value(dataset_dict, 'maintainer'),
self._get_dataset_value(dataset_dict, 'maintainer_email'),
]):
contact_uri = self._get_dataset_value(dataset_dict, 'contact_uri')
if contact_uri:
contact_details = URIRef(contact_uri)
else:
contact_details = BNode()
g.add((contact_details, RDF.type, VCARD.Kind))
g.add((dataset_ref, DCAT.contactPoint, contact_details))
items = [
('contact_name', VCARD.fn, ['maintainer'], Literal),
('contact_email', VCARD.hasEmail, ['maintainer_email'],
Literal),
]
self._add_triples_from_dict(dataset_dict, contact_details, items)
# Publisher
if any([
self._get_dataset_value(dataset_dict, 'publisher_uri'),
self._get_dataset_value(dataset_dict, 'publisher_name'),
self._get_dataset_value(dataset_dict, 'publisher_identifier'),
dataset_dict.get('organization'),
]):
publisher_uri = publisher_uri_from_dataset_dict(dataset_dict)
if publisher_uri:
publisher_details = URIRef(publisher_uri)
else:
# No organization nor publisher_uri
publisher_details = BNode()
g.add((publisher_details, RDF.type, FOAF.Agent))
g.add((dataset_ref, DCT.publisher, publisher_details))
publisher_name = self._get_dataset_value(dataset_dict,
'publisher_name')
if not publisher_name and dataset_dict.get('organization'):
publisher_name = dataset_dict['organization']['title']
g.add((publisher_details, FOAF.name, Literal(publisher_name)))
# TODO: It would make sense to fallback these to organization
# fields but they are not in the default schema and the
# `organization` object in the dataset_dict does not include
# custom fields
items = [('publisher_email', FOAF.mbox, None, Literal),
('publisher_identifier', DCT.identifier, None, Literal),
('publisher_url', FOAF.homepage, None, URIRef),
('publisher_type', DCT.type, None, Literal)]
self._add_triples_from_dict(dataset_dict, publisher_details, items)
# Temporal
start = self._get_dataset_value(dataset_dict, 'temporal_start')
end = self._get_dataset_value(dataset_dict, 'temporal_end')
if start or end:
temporal_extent = BNode()
g.add((temporal_extent, RDF.type, DCT.PeriodOfTime))
if start:
self._add_date_triple(temporal_extent, SCHEMA.startDate, start)
if end:
self._add_date_triple(temporal_extent, SCHEMA.endDate, end)
g.add((dataset_ref, DCT.temporal, temporal_extent))
# parts - has part/is part of
if any([
self._get_dataset_value(dataset_dict, 'has_part'),
self._get_dataset_value(dataset_dict, 'is_part_of')
]):
items = [('has_part', DCT.hasPart, None, URIRef),
('is_part_of', DCT.isPartOf, None, URIRef)]
self._add_list_triples_from_dict(dataset_dict, dataset_ref, items)
# Spatial
spatial_uri = self._get_dataset_value(dataset_dict, 'spatial_uri')
spatial_text = self._get_dataset_value(dataset_dict, 'spatial_text')
spatial_geom = self._get_dataset_value(dataset_dict, 'spatial')
if spatial_uri:
spatial_uri = get_spatial_uri(spatial_uri) # map from code to URI
if spatial_uri or spatial_text or spatial_geom:
if spatial_uri:
spatial_ref = URIRef(spatial_uri)
else:
spatial_ref = BNode()
g.add((spatial_ref, RDF.type, DCT.Location))
g.add((dataset_ref, DCT.spatial, spatial_ref))
if spatial_text:
g.add((spatial_ref, SKOS.prefLabel, Literal(spatial_text)))
if spatial_geom:
# GeoJSON
g.add((spatial_ref, LOCN.geometry,
Literal(spatial_geom, datatype=GEOJSON_IMT)))
# WKT, because GeoDCAT-AP says so
try:
g.add((spatial_ref, LOCN.geometry,
Literal(wkt.dumps(json.loads(spatial_geom),
decimals=4),
datatype=GSP.wktLiteral)))
except (TypeError, ValueError, InvalidGeoJSONException):
pass
# Resources
for resource_dict in dataset_dict.get('resources', []):
distribution = URIRef(resource_uri(resource_dict))
g.add((dataset_ref, DCAT.distribution, distribution))
g.add((distribution, RDF.type, DCAT.Distribution))
if 'license' not in resource_dict and 'license_id' in dataset_dict:
lr = LicenseRegister()
_license = lr.get(dataset_dict['license_id'])
if _license:
resource_dict['license'] = _license.url
# Simple values
items = [
('name', DCT.title, None, Literal),
('description', DCT.description, None, Literal),
('status', ADMS.status, None, Literal),
('rights', DCT.rights, None, Literal),
('license', DCT.license, None, URIRef),
]
self._add_triples_from_dict(resource_dict, distribution, items)
# Lists
items = [
('documentation', FOAF.page, None, URIRef),
('language', DCT.language, None, URIRef),
('conforms_to', DCT.conformsTo, None, URIRef),
]
self._add_list_triples_from_dict(resource_dict, distribution,
items)
# Format
if '/' in resource_dict.get('format', ''):
g.add((distribution, DCAT.mediaType,
Literal(resource_dict['format'])))
else:
if resource_dict.get('format'):
g.add((distribution, DCT['format'],
Literal(resource_dict['format'])))
if resource_dict.get('mimetype'):
g.add((distribution, DCAT.mediaType,
Literal(resource_dict['mimetype'])))
# URL
url = resource_dict.get('url')
download_url = resource_dict.get('download_url')
if download_url:
g.add((distribution, DCAT.downloadURL, URIRef(download_url)))
if (url and not download_url) or (url and url != download_url):
g.add((distribution, DCAT.accessURL, URIRef(url)))
# Dates
items = [
('issued', DCT.issued, None, Literal),
('modified', DCT.modified, None, Literal),
]
self._add_date_triples_from_dict(resource_dict, distribution,
items)
# Numbers
if resource_dict.get('size'):
try:
g.add((distribution, DCAT.byteSize,
Literal(float(resource_dict['size']),
datatype=XSD.decimal)))
except (ValueError, TypeError):
g.add((distribution, DCAT.byteSize,
Literal(resource_dict['size'])))
# Checksum
if resource_dict.get('hash'):
checksum = BNode()
g.add((checksum, SPDX.checksumValue,
Literal(resource_dict['hash'], datatype=XSD.hexBinary)))
if resource_dict.get('hash_algorithm'):
if resource_dict['hash_algorithm'].startswith('http'):
g.add((checksum, SPDX.algorithm,
URIRef(resource_dict['hash_algorithm'])))
else:
g.add((checksum, SPDX.algorithm,
Literal(resource_dict['hash_algorithm'])))
g.add((distribution, SPDX.checksum, checksum))
def graph_from_dataset(self, dataset_dict, dataset_ref):
g = self.g
for prefix, namespace in namespaces.iteritems():
g.bind(prefix, namespace)
g.add((dataset_ref, RDF.type, DCAT.Dataset))
# Basic fields
items = [
('title', DCT.title, None),
('notes', DCT.description, None),
('url', DCAT.landingPage, None),
('identifier', DCT.identifier, ['guid', 'id']),
('version', OWL.versionInfo, ['dcat_version']),
('version_notes', ADMS.versionNotes, None),
('frequency', DCT.accrualPeriodicity, None),
('access_rights', DCT.accessRights, None),
]
self._add_triples_from_dict(dataset_dict, dataset_ref, items)
# Tags
for tag in dataset_dict.get('tags', []):
g.add((dataset_ref, DCAT.keyword, Literal(tag['name'])))
# Dates
items = [
('issued', DCT.issued, ['metadata_created']),
('modified', DCT.modified, ['metadata_modified']),
]
self._add_date_triples_from_dict(dataset_dict, dataset_ref, items)
# Lists
items = [
('language', DCT.language, None),
('theme', DCAT.theme, None),
('conforms_to', DCT.conformsTo, None),
('alternate_identifier', ADMS.identifier, None),
('documentation', FOAF.page, None),
('related_resource', DCT.relation, None),
('has_version', DCT.hasVersion, None),
('is_version_of', DCT.isVersionOf, None),
('source', DCT.source, None),
('sample', ADMS.sample, None),
]
self._add_list_triples_from_dict(dataset_dict, dataset_ref, items)
# Contact details
if any([
self._get_dataset_value(dataset_dict, 'contact_uri'),
self._get_dataset_value(dataset_dict, 'contact_name'),
self._get_dataset_value(dataset_dict, 'contact_email'),
self._get_dataset_value(dataset_dict, 'maintainer'),
self._get_dataset_value(dataset_dict, 'maintainer_email'),
self._get_dataset_value(dataset_dict, 'author'),
self._get_dataset_value(dataset_dict, 'author_email'),
]):
contact_uri = self._get_dataset_value(dataset_dict, 'contact_uri')
if contact_uri:
contact_details = URIRef(contact_uri)
else:
contact_details = BNode()
g.add((contact_details, RDF.type, VCARD.Organization))
g.add((dataset_ref, DCAT.contactPoint, contact_details))
items = [
('contact_name', VCARD.fn, ['maintainer', 'author']),
('contact_email', VCARD.hasEmail, ['maintainer_email',
'author_email']),
]
self._add_triples_from_dict(dataset_dict, contact_details, items)
# Publisher
if any([
self._get_dataset_value(dataset_dict, 'publisher_uri'),
self._get_dataset_value(dataset_dict, 'publisher_name'),
dataset_dict.get('organization'),
]):
publisher_uri = publisher_uri_from_dataset_dict(dataset_dict)
if publisher_uri:
publisher_details = URIRef(publisher_uri)
else:
# No organization nor publisher_uri
publisher_details = BNode()
g.add((publisher_details, RDF.type, FOAF.Organization))
g.add((dataset_ref, DCT.publisher, publisher_details))
publisher_name = self._get_dataset_value(dataset_dict, 'publisher_name')
if not publisher_name and dataset_dict.get('organization'):
publisher_name = dataset_dict['organization']['title']
g.add((publisher_details, FOAF.name, Literal(publisher_name)))
# TODO: It would make sense to fallback these to organization
# fields but they are not in the default schema and the
# `organization` object in the dataset_dict does not include
# custom fields
items = [
('publisher_email', FOAF.mbox, None),
('publisher_url', FOAF.homepage, None),
('publisher_type', DCT.type, None),
]
self._add_triples_from_dict(dataset_dict, publisher_details, items)
# Temporal
start = self._get_dataset_value(dataset_dict, 'temporal_start')
end = self._get_dataset_value(dataset_dict, 'temporal_end')
if start or end:
temporal_extent = BNode()
g.add((temporal_extent, RDF.type, DCT.PeriodOfTime))
if start:
self._add_date_triple(temporal_extent, SCHEMA.startDate, start)
if end:
self._add_date_triple(temporal_extent, SCHEMA.endDate, end)
g.add((dataset_ref, DCT.temporal, temporal_extent))
# Spatial
spatial_uri = self._get_dataset_value(dataset_dict, 'spatial_uri')
spatial_text = self._get_dataset_value(dataset_dict, 'spatial_text')
spatial_geom = self._get_dataset_value(dataset_dict, 'spatial')
if spatial_uri or spatial_text or spatial_geom:
if spatial_uri:
spatial_ref = URIRef(spatial_uri)
else:
spatial_ref = BNode()
g.add((spatial_ref, RDF.type, DCT.Location))
g.add((dataset_ref, DCT.spatial, spatial_ref))
if spatial_text:
g.add((spatial_ref, SKOS.prefLabel, Literal(spatial_text)))
if spatial_geom:
# GeoJSON
g.add((spatial_ref,
LOCN.geometry,
Literal(spatial_geom, datatype=GEOJSON_IMT)))
# WKT, because GeoDCAT-AP says so
try:
g.add((spatial_ref,
LOCN.geometry,
Literal(wkt.dumps(json.loads(spatial_geom),
decimals=4),
datatype=GSP.wktLiteral)))
except (TypeError, ValueError, InvalidGeoJSONException):
pass
# Resources
for resource_dict in dataset_dict.get('resources', []):
distribution = URIRef(resource_uri(resource_dict))
g.add((dataset_ref, DCAT.distribution, distribution))
g.add((distribution, RDF.type, DCAT.Distribution))
# Simple values
items = [
('name', DCT.title, None),
('description', DCT.description, None),
('status', ADMS.status, None),
('rights', DCT.rights, None),
('license', DCT.license, None),
]
self._add_triples_from_dict(resource_dict, distribution, items)
# Lists
items = [
('documentation', FOAF.page, None),
('language', DCT.language, None),
('conforms_to', DCT.conformsTo, None),
]
self._add_list_triples_from_dict(resource_dict, distribution, items)
# Format
if '/' in resource_dict.get('format', ''):
g.add((distribution, DCAT.mediaType,
Literal(resource_dict['format'])))
else:
if resource_dict.get('format'):
g.add((distribution, DCT['format'],
Literal(resource_dict['format'])))
if resource_dict.get('mimetype'):
g.add((distribution, DCAT.mediaType,
Literal(resource_dict['mimetype'])))
# URL
url = resource_dict.get('url')
download_url = resource_dict.get('download_url')
if download_url:
g.add((distribution, DCAT.downloadURL, Literal(download_url)))
if (url and not download_url) or (url and url != download_url):
g.add((distribution, DCAT.accessURL, Literal(url)))
# Dates
items = [
('issued', DCT.issued, None),
('modified', DCT.modified, None),
]
self._add_date_triples_from_dict(resource_dict, distribution, items)
# Numbers
if resource_dict.get('size'):
try:
g.add((distribution, DCAT.byteSize,
Literal(float(resource_dict['size']),
datatype=XSD.decimal)))
except (ValueError, TypeError):
g.add((distribution, DCAT.byteSize,
Literal(resource_dict['size'])))
# Checksum
if resource_dict.get('hash'):
checksum = BNode()
g.add((checksum, SPDX.checksumValue,
Literal(resource_dict['hash'],
datatype=XSD.hexBinary)))
if resource_dict.get('hash_algorithm'):
if resource_dict['hash_algorithm'].startswith('http'):
g.add((checksum, SPDX.algorithm,
URIRef(resource_dict['hash_algorithm'])))
else:
g.add((checksum, SPDX.algorithm,
Literal(resource_dict['hash_algorithm'])))
g.add((distribution, SPDX.checksum, checksum))
def graph_from_dataset(self, dataset_dict, dataset_ref):
g = self.g
for prefix, namespace in namespaces.iteritems():
g.bind(prefix, namespace)
g.add((dataset_ref, RDF.type, DCAT.Dataset))
# Basic fields
items = [
('title', DCT.title, None),
('notes', DCT.description, None),
('url', DCAT.landingPage, None),
('identifier', DCT.identifier, ['guid', 'id']),
('version', OWL.versionInfo, ['dcat_version']),
('alternate_identifier', ADMS.identifier, None),
('version_notes', ADMS.versionNotes, None),
('frequency', DCT.accrualPeriodicity, None),
('accrualPeriodicity', DCT.accrualPeriodicity, None)
]
self._add_triples_from_dict(dataset_dict, dataset_ref, items)
# Tags
for tag in dataset_dict.get('tags', []):
g.add((dataset_ref, DCAT.keyword, Literal(tag['name'])))
# Dates
items = [
('issued', DCT.issued, ['metadata_created']),
('modified', DCT.modified, ['metadata_modified']),
]
self._add_date_triples_from_dict(dataset_dict, dataset_ref, items)
# Lists
items = [
('language', DCT.language, None),
('theme-primary', DCAT.theme, None),
('theme-secondary', DCAT.theme, None),
('conforms-to', DCAT.conformsTo, None),
('lineage', DCT.provenance, None)
]
self._add_list_triples_from_dict(dataset_dict, dataset_ref, items)
# Contact details
if any([
self._get_dataset_value(dataset_dict, 'contact_uri'),
self._get_dataset_value(dataset_dict, 'contact_name'),
self._get_dataset_value(dataset_dict, 'contact_email'),
self._get_dataset_value(dataset_dict, 'maintainer'),
self._get_dataset_value(dataset_dict, 'maintainer_email'),
self._get_dataset_value(dataset_dict, 'author'),
self._get_dataset_value(dataset_dict, 'author_email'),
]):
contact_uri = self._get_dataset_value(dataset_dict, 'contact_uri')
if contact_uri:
contact_details = URIRef(contact_uri)
else:
contact_details = BNode()
g.add((contact_details, RDF.type, VCARD.Organization))
g.add((dataset_ref, DCAT.contactPoint, contact_details))
items = [
('contact_name', VCARD.fn, ['maintainer', 'author']),
('contact_email', VCARD.hasEmail, ['maintainer_email',
'author_email']),
]
self._add_triples_from_dict(dataset_dict, contact_details, items)
license_id = self._get_dataset_value(dataset_dict, 'license_id')
if (license_id == 'cc-by'):
g.add((dataset_ref, DCT.license, Literal('https://creativecommons.org/licenses/by/4.0/')))
else:
g.add((dataset_ref, DCT.license, Literal(license_id)))
# Publisher
if any([
self._get_dataset_value(dataset_dict, 'publisher_uri'),
self._get_dataset_value(dataset_dict, 'publisher_name'),
dataset_dict.get('organization'),
]):
publisher_uri = publisher_uri_from_dataset_dict(dataset_dict)
if publisher_uri:
publisher_details = URIRef(publisher_uri)
else:
# No organization nor publisher_uri
publisher_details = BNode()
g.add((publisher_details, RDF.type, FOAF.Organization))
g.add((dataset_ref, DCT.publisher, publisher_details))
publisher_name = self._get_dataset_value(dataset_dict, 'publisher_name')
if not publisher_name and dataset_dict.get('organization'):
publisher_name = dataset_dict['organization']['title']
g.add((publisher_details, FOAF.name, Literal(publisher_name)))
# TODO: It would make sense to fallback these to organization
# fields but they are not in the default schema and the
# `organization` object in the dataset_dict does not include
# custom fields
items = [
('publisher_email', FOAF.mbox, None),
('publisher_url', FOAF.homepage, None),
('publisher_type', DCT.type, None),
]
self._add_triples_from_dict(dataset_dict, publisher_details, items)
# Update Frequency
# accrualPeriodicity
update_freq = self._get_dataset_value(dataset_dict, 'update_frequency')
if (update_freq):
has_uri = False
# check if there exists a URI for the update_frequency value
from ckanext.dgu.forms.dataset_form import update_frequency_uri
for freq_name, freq_uri in update_frequency_uri:
if freq_name.lower() == update_freq.lower():
has_uri = True
break
g.add((dataset_ref, DCT.accrualPeriodicity, URIRef(freq_uri) if has_uri else Literal(update_freq)))
# Temporal
start = self._get_dataset_value(dataset_dict, 'temporal_start')
end = self._get_dataset_value(dataset_dict, 'temporal_end')
if start or end:
temporal_extent = BNode()
g.add((temporal_extent, RDF.type, DCT.PeriodOfTime))
if start:
self._add_date_triple(temporal_extent, SCHEMA.startDate, start)
if end:
self._add_date_triple(temporal_extent, SCHEMA.endDate, end)
g.add((dataset_ref, DCT.temporal, temporal_extent))
# Spatial
spatial_uri = self._get_dataset_value(dataset_dict, 'spatial_uri')
spatial_text = self._get_dataset_value(dataset_dict, 'spatial_text')
spatial_geom = self._get_dataset_value(dataset_dict, 'spatial')
if spatial_uri or spatial_text or spatial_geom:
if spatial_uri:
spatial_ref = URIRef(spatial_uri)
else:
spatial_ref = BNode()
g.add((spatial_ref, RDF.type, DCT.Location))
g.add((dataset_ref, DCT.spatial, spatial_ref))
if spatial_text:
g.add((spatial_ref, SKOS.prefLabel, Literal(spatial_text)))
if spatial_geom:
# GeoJSON
g.add((spatial_ref,
LOCN.geometry,
Literal(spatial_geom, datatype=GEOJSON_IMT)))
# WKT, because GeoDCAT-AP says so
try:
g.add((spatial_ref,
LOCN.geometry,
Literal(wkt.dumps(json.loads(spatial_geom),
decimals=4),
datatype=GSP.wktLiteral)))
except (TypeError, ValueError, InvalidGeoJSONException):
pass
# Resources
for resource_dict in dataset_dict.get('resources', []):
distribution = URIRef(resource_uri(resource_dict))
g.add((dataset_ref, DCAT.distribution, distribution))
g.add((distribution, RDF.type, DCAT.Distribution))
# Simple values
items = [
('name', DCT.title, None),
('description', DCT.description, None),
('status', ADMS.status, None),
('rights', DCT.rights, None),
('license', DCT.license, None),
]
self._add_triples_from_dict(resource_dict, distribution, items)
# Format
_format = resource_dict['format']
if _format:
if '/' in _format:
# add dct:format
dctFormat = _format.strip().replace("/", ".").replace(" ", "").lower()
g.add((distribution, DCT['format'], Literal(dctFormat)))
else:
g.add((distribution, DCT['format'], Literal(_format.lower())))
# add dcat:mediaType
fmt = formats.Formats.match(_format.strip().lower())
mime_types = fmt['mime_types'] if fmt else None
if mime_types:
g.add((distribution, DCAT.mediaType, Literal(mime_types)))
license_id = self._get_dataset_value(dataset_dict, 'license_id')
if (license_id == 'cc-by'):
g.add((distribution, DCT.license, Literal('https://creativecommons.org/licenses/by/4.0/')))
else:
g.add((distribution, DCT.license, Literal(license_id)))
# URL
url = resource_dict.get('url')
download_url = resource_dict.get('download_url')
if download_url:
g.add((distribution, DCAT.downloadURL, Literal(download_url)))
if (url and not download_url) or (url and url != download_url):
g.add((distribution, DCAT.accessURL, Literal(url)))
# Dates
items = [
('issued', DCT.issued, None),
('modified', DCT.modified, None),
]
self._add_date_triples_from_dict(resource_dict, distribution, items)
# Numbers
if resource_dict.get('size'):
try:
g.add((distribution, DCAT.byteSize,
Literal(float(resource_dict['size']),
datatype=XSD.decimal)))
except (ValueError, TypeError):
g.add((distribution, DCAT.byteSize,
Literal(resource_dict['size'])))
def graph_from_dataset(self, dataset_dict, dataset_ref):
g = self.g
for prefix, namespace in namespaces.iteritems():
g.bind(prefix, namespace)
g.add((dataset_ref, RDF.type, DCAT.Dataset))
# Basic fields
items = [
('title', DCT.title, None, Literal),
('notes', DCT.description, None, Literal),
('url', DCAT.landingPage, None, URIRef),
('identifier', DCT.identifier, ['guid', 'id'], Literal),
('version', OWL.versionInfo, ['dcat_version'], Literal),
('version_notes', ADMS.versionNotes, None, Literal),
('frequency', DCT.accrualPeriodicity, None, URIRef),
('access_rights', DCT.accessRights, None, Literal),
('dcat_type', DCT.type, None, Literal),
('provenance', DCT.provenance, None, Literal),
]
self._add_triples_from_dict(dataset_dict, dataset_ref, items)
# Tags
for tag in dataset_dict.get('tags', []):
g.add((dataset_ref, DCAT.keyword, Literal(tag['name'])))
# Dates
items = [
('issued', DCT.issued, ['metadata_created'], Literal),
('modified', DCT.modified, ['metadata_modified'], Literal),
]
self._add_date_triples_from_dict(dataset_dict, dataset_ref, items)
# Lists
items = [
('language', DCT.language, None, Literal),
('theme', DCAT.theme, None, URIRef),
('conforms_to', DCT.conformsTo, None, Literal),
('alternate_identifier', ADMS.identifier, None, Literal),
('documentation', FOAF.page, None, URIRef),
('related_resource', DCT.relation, None, URIRef),
('has_version', DCT.hasVersion, None, URIRef),
('is_version_of', DCT.isVersionOf, None, URIRef),
('source', DCT.source, None, Literal),
('sample', ADMS.sample, None, Literal),
]
self._add_list_triples_from_dict(dataset_dict, dataset_ref, items)
# Contact details
if any([
self._get_dataset_value(dataset_dict, 'contact_uri'),
self._get_dataset_value(dataset_dict, 'contact_name'),
self._get_dataset_value(dataset_dict, 'contact_email'),
self._get_dataset_value(dataset_dict, 'maintainer'),
self._get_dataset_value(dataset_dict, 'maintainer_email'),
self._get_dataset_value(dataset_dict, 'author'),
self._get_dataset_value(dataset_dict, 'author_email'),
]):
contact_uri = self._get_dataset_value(dataset_dict, 'contact_uri')
if contact_uri:
contact_details = URIRef(self._removeWhitespaces(contact_uri))
else:
contact_details = BNode()
g.add((contact_details, RDF.type, VCARD.Organization))
g.add((dataset_ref, DCAT.contactPoint, contact_details))
self._add_triple_from_dict(dataset_dict, contact_details, VCARD.fn,
'contact_name',
['maintainer', 'author'])
# Add mail address as URIRef, and ensure it has a mailto: prefix
self._add_triple_from_dict(dataset_dict,
contact_details,
VCARD.hasEmail,
'contact_email',
['maintainer_email', 'author_email'],
_type=URIRef,
value_modifier=self._add_mailto)
# Publisher
if any([
self._get_dataset_value(dataset_dict, 'publisher_uri'),
self._get_dataset_value(dataset_dict, 'publisher_name'),
dataset_dict.get('organization'),
]):
publisher_uri = publisher_uri_from_dataset_dict(dataset_dict)
if publisher_uri:
publisher_details = URIRef(
self._removeWhitespaces(publisher_uri))
else:
# No organization nor publisher_uri
publisher_details = BNode()
g.add((publisher_details, RDF.type, FOAF.Organization))
g.add((dataset_ref, DCT.publisher, publisher_details))
publisher_name = self._get_dataset_value(dataset_dict,
'publisher_name')
if not publisher_name and dataset_dict.get('organization'):
publisher_name = dataset_dict['organization']['title']
g.add((publisher_details, FOAF.name, Literal(publisher_name)))
# TODO: It would make sense to fallback these to organization
# fields but they are not in the default schema and the
# `organization` object in the dataset_dict does not include
# custom fields
items = [
('publisher_email', FOAF.mbox, None, Literal),
('publisher_url', FOAF.homepage, None, URIRef),
('publisher_type', DCT.type, None, URIRef),
]
self._add_triples_from_dict(dataset_dict, publisher_details, items)
# Temporal
start = self._get_dataset_value(dataset_dict, 'temporal_start')
end = self._get_dataset_value(dataset_dict, 'temporal_end')
if start or end:
temporal_extent = BNode()
g.add((temporal_extent, RDF.type, DCT.PeriodOfTime))
if start:
self._add_date_triple(temporal_extent, SCHEMA.startDate, start)
if end:
self._add_date_triple(temporal_extent, SCHEMA.endDate, end)
g.add((dataset_ref, DCT.temporal, temporal_extent))
# Spatial
spatial_uri = self._get_dataset_value(dataset_dict, 'spatial_uri')
spatial_text = self._get_dataset_value(dataset_dict, 'spatial_text')
spatial_geom = self._get_dataset_value(dataset_dict, 'spatial')
if spatial_uri or spatial_text or spatial_geom:
if spatial_uri:
spatial_ref = URIRef(self._removeWhitespaces(spatial_uri))
else:
spatial_ref = BNode()
g.add((spatial_ref, RDF.type, DCT.Location))
g.add((dataset_ref, DCT.spatial, spatial_ref))
if spatial_text:
g.add((spatial_ref, SKOS.prefLabel, Literal(spatial_text)))
if spatial_geom:
# GeoJSON
g.add((spatial_ref, LOCN.geometry,
Literal(spatial_geom, datatype=GEOJSON_IMT)))
# WKT, because GeoDCAT-AP says so
try:
g.add((spatial_ref, LOCN.geometry,
Literal(wkt.dumps(json.loads(spatial_geom),
decimals=4),
datatype=GSP.wktLiteral)))
except (TypeError, ValueError, InvalidGeoJSONException):
pass
# Resources
for resource_dict in dataset_dict.get('resources', []):
distribution = URIRef(
self._removeWhitespaces(resource_uri(resource_dict)))
g.add((dataset_ref, DCAT.distribution, distribution))
g.add((distribution, RDF.type, DCAT.Distribution))
# Simple values
items = [('name', DCT.title, None, Literal),
('description', DCT.description, None, Literal),
('status', ADMS.status, None, URIRef),
('rights', DCT.rights, None, URIRef),
('license', DCT.license, None, URIRef),
('access_url', DCAT.accessURL, None, URIRef),
('download_url', DCAT.downloadURL, None, URIRef)]
self._add_triples_from_dict(resource_dict, distribution, items)
# Lists
items = [
('documentation', FOAF.page, None, URIRef),
('language', DCT.language, None, Literal),
('conforms_to', DCT.conformsTo, None, Literal),
]
self._add_list_triples_from_dict(resource_dict, distribution,
items)
# Format
mimetype = resource_dict.get('mimetype')
fmt = resource_dict.get('format')
# IANA media types (either URI or Literal) should be mapped as mediaType.
# In case format is available and mimetype is not set or identical to format,
# check which type is appropriate.
if fmt and (not mimetype or mimetype == fmt):
if ('iana.org/assignments/media-types' in fmt
or not fmt.startswith('http') and '/' in fmt):
# output format value as dcat:mediaType instead of dct:format
mimetype = fmt
fmt = None
else:
# Use dct:format
mimetype = None
if mimetype:
if mimetype.startswith('http'):
g.add((distribution, DCAT.mediaType,
URIRef(self._removeWhitespaces(mimetype))))
else:
g.add((distribution, DCAT.mediaType, Literal(mimetype)))
if fmt:
if fmt.startswith('http'):
g.add((distribution, DCT['format'],
URIRef(self._removeWhitespaces(fmt))))
else:
g.add((distribution, DCT['format'], Literal(fmt)))
# URL fallback and old behavior
url = resource_dict.get('url')
download_url = resource_dict.get('download_url')
access_url = resource_dict.get('access_url')
# Use url as fallback for access_url if access_url is not set and download_url is not equal
if (url and ((not (access_url or download_url)) or
((not access_url) and
(download_url and url != download_url)))):
self._add_triple_from_dict(resource_dict,
distribution,
DCAT.accessURL,
'url',
_type=URIRef)
# Dates
items = [
('issued', DCT.issued, None, Literal),
('modified', DCT.modified, None, Literal),
]
self._add_date_triples_from_dict(resource_dict, distribution,
items)
# Numbers
if resource_dict.get('size'):
try:
g.add((distribution, DCAT.byteSize,
Literal(float(resource_dict['size']),
datatype=XSD.decimal)))
except (ValueError, TypeError):
g.add((distribution, DCAT.byteSize,
Literal(resource_dict['size'])))
# Checksum
if resource_dict.get('hash'):
checksum = BNode()
g.add((checksum, RDF.type, SPDX.Checksum))
g.add((checksum, SPDX.checksumValue,
Literal(resource_dict['hash'], datatype=XSD.hexBinary)))
if resource_dict.get('hash_algorithm'):
if resource_dict['hash_algorithm'].startswith('http'):
g.add((checksum, SPDX.algorithm,
URIRef(
self._removeWhitespaces(
resource_dict['hash_algorithm']))))
else:
g.add((checksum, SPDX.algorithm,
Literal(resource_dict['hash_algorithm'])))
g.add((distribution, SPDX.checksum, checksum))
