def _create_metadata_record(self, contact: Organization,
group: MrMapGroup):
""" Creates a Metadata record from the OGCWebMapService
Args:
contact (Organization): The contact organization for this metadata record
group (MrMapGroup): The owner/creator group
Returns:
metadata (Metadata): The persisted metadata record
"""
metadata = Metadata()
md_type = MetadataEnum.SERVICE.value
metadata.metadata_type = md_type
if self.service_file_iso_identifier is None:
# We didn't found any file identifier in the document -> we create one
self.service_file_iso_identifier = uuid.uuid4()
metadata.title = self.service_identification_title
metadata.abstract = self.service_identification_abstract
metadata.online_resource = self.service_provider_onlineresource_linkage
metadata.capabilities_original_uri = self.service_connect_url
metadata.access_constraints = self.service_identification_accessconstraints
metadata.fees = self.service_identification_fees
if self.service_bounding_box is not None:
metadata.bounding_geometry = self.service_bounding_box
metadata.identifier = self.service_file_identifier
metadata.is_active = False
metadata.created_by = group
metadata.contact = contact
# Save metadata instance to be able to add M2M entities
metadata.save()
return metadata
def to_db_model(self,
type=MetadataEnum.DATASET.value,
created_by: MrMapGroup = None):
""" Get corresponding metadata object from database or create it if not found!
Returns:
metadata (Metadata): A db model Metadata object
"""
update = False
new = False
# try to find the object by uuid and uri. If not existing yet, create a new record
try:
metadata = Metadata.objects.get(identifier=self.file_identifier,
metadata_url=self.uri)
# check if the parsed metadata might be newer
# make sure both date time objects will be comparable
persisted_change = metadata.last_remote_change.replace(tzinfo=utc)
new_change = self.last_change_date.replace(tzinfo=utc)
if persisted_change > new_change:
# Nothing to do here
return metadata
else:
update = True
except ObjectDoesNotExist:
# object does not seem to exist -> create it!
metadata = Metadata()
md_type = type
metadata.metadata_type = md_type
if metadata.is_dataset_metadata:
metadata.dataset = Dataset()
metadata.dataset.created_by = created_by
metadata.created_by = created_by
new = True
if update or new:
# In case of a dataset, we need to fill the information into the dataset object
if metadata.is_dataset_metadata:
metadata.dataset = self._fill_dataset_db_model(
metadata.dataset)
metadata = self._fill_metadata_db_model(metadata)
metadata.save()
metadata.dataset.save()
orig_document = Document.objects.get_or_create(
metadata=metadata,
document_type=DocumentEnum.METADATA.value,
is_original=True,
)[0]
orig_document.content = self.raw_metadata
orig_document.save()
if update:
metadata.keywords.clear()
for kw in self.keywords:
keyword = Keyword.objects.get_or_create(keyword=kw)[0]
metadata.keywords.add(keyword)
return metadata
def create_service_model_instance(self, user: MrMapUser, register_group,
register_for_organization, external_auth,
is_update_candidate_for):
""" Map all data from the OGCCatalogueService class to their database models
Args:
user (MrMapUser): The user which performs the action
register_group (Group): The group which is used to register this service
register_for_organization (Organization): The organization for which this service is being registered
external_auth (ExternalAuthentication): The external authentication object
Returns:
service (Service): Service instance, contains all information, ready for persisting!
"""
if current_task:
current_task.update_state(state=states.STARTED,
meta={
'current':
PROGRESS_STATUS_AFTER_PARSING,
'phase': 'Persisting...',
})
md = Metadata()
md_type = MetadataEnum.CATALOGUE.value
md.metadata_type = md_type
md.title = self.service_identification_title
md.identifier = self.service_file_identifier
md.abstract = self.service_identification_abstract
md.online_resource = self.service_provider_onlineresource_linkage
md.contact = self._create_contact_organization_record()
md.authority_url = self.service_provider_url
md.access_constraints = self.service_identification_accessconstraints
md.fees = self.service_identification_fees
md.created_by = register_group
md.capabilities_original_uri = self.service_connect_url
if self.service_bounding_box is not None:
md.bounding_geometry = self.service_bounding_box
# Save metadata record so we can use M2M or id of record later
md.save()
md.identifier = str(md.id) if md.identifier is None else md.identifier
# Keywords
for kw in self.service_identification_keywords:
if kw is None:
continue
keyword = Keyword.objects.get_or_create(keyword=kw)[0]
md.keywords.add(keyword)
md.formats.add(*self.formats_list)
md.save()
service = self._create_service_record(register_group,
register_for_organization, md,
is_update_candidate_for)
return service
def _create_metadata_record(self, parent_service: Service,
group: MrMapGroup):
""" Creates a Metadata record from the OGCLayer object
Args:
self (OGCLayer): The OGCLayer object (result of parsing)
parent_service (Service): The parent Service object
group (MrMapGroup): The creator/owner group
Returns:
metadata (Metadata): The persisted metadata object
"""
metadata = Metadata()
md_type = MetadataEnum.LAYER.value
metadata.metadata_type = md_type
metadata.title = self.title
metadata.abstract = self.abstract
metadata.online_resource = parent_service.metadata.online_resource
metadata.capabilities_original_uri = parent_service.metadata.capabilities_original_uri
metadata.identifier = self.identifier
metadata.contact = parent_service.metadata.contact
metadata.access_constraints = parent_service.metadata.access_constraints
metadata.is_active = False
metadata.created_by = group
# Save metadata to use id afterwards
metadata.save()
# create bounding box polygon
bounding_points = ((float(self.capability_bbox_lat_lon["minx"]),
float(self.capability_bbox_lat_lon["miny"])),
(float(self.capability_bbox_lat_lon["minx"]),
float(self.capability_bbox_lat_lon["maxy"])),
(float(self.capability_bbox_lat_lon["maxx"]),
float(self.capability_bbox_lat_lon["maxy"])),
(float(self.capability_bbox_lat_lon["maxx"]),
float(self.capability_bbox_lat_lon["miny"])),
(float(self.capability_bbox_lat_lon["minx"]),
float(self.capability_bbox_lat_lon["miny"])))
metadata.bounding_geometry = Polygon(bounding_points)
metadata.save()
return metadata
def _persist_metadata(self, md_data_entry: dict):
""" Creates real Metadata model records from the parsed data
Args:
md_data_entry (dict):
Returns:
metadata (Metadata): The persisted metadata object
"""
_id = md_data_entry["id"]
# Remove this id from the set of metadata which shall be deleted in the end.
try:
self.deleted_metadata.remove(_id)
except KeyError:
pass
try:
md = Metadata.objects.get(identifier=_id, )
is_new = False
if md.last_remote_change == md_data_entry["date_stamp"]:
# Nothing to do here!
return
except ObjectDoesNotExist:
md = Metadata(identifier=_id)
is_new = True
md.access_constraints = md_data_entry.get("access_constraints", None)
md.created_by = self.harvesting_group
md.origin = ResourceOriginEnum.CATALOGUE.value
md.last_remote_change = md_data_entry.get("date_stamp", None)
md.title = md_data_entry.get("title", None)
md.contact = md_data_entry.get("contact", None)
md.language_code = md_data_entry.get("language_code", None)
md.metadata_type = md_data_entry.get("metadata_type", None)
md.abstract = md_data_entry.get("abstract", None)
md.bounding_geometry = md_data_entry.get("bounding_geometry", None)
formats = md_data_entry.get("formats", [])
md.is_active = True
md.capabilities_original_uri = md_data_entry.get(
"capabilities_original_url", None)
try:
# Improve speed for keyword get-create by fetching (filter) all existing ones and only perform
# get_or_create on the ones that do not exist yet. Speed up by ~50% for large amount of data
existing_kws = Keyword.objects.filter(
keyword__in=md_data_entry["keywords"])
existing_kws = [kw.keyword for kw in existing_kws]
new_kws = [
kw for kw in md_data_entry["keywords"]
if kw not in existing_kws
]
[Keyword.objects.get_or_create(keyword=kw)[0] for kw in new_kws]
kws = Keyword.objects.filter(keyword__in=md_data_entry["keywords"])
# Same for MimeTypes
existing_formats = MimeType.objects.filter(
mime_type__in=md_data_entry["formats"])
existing_formats = [
_format.mime_type for _format in existing_formats
]
new_formats = [
_format for _format in md_data_entry["formats"]
if _format not in existing_formats
]
[
MimeType.objects.get_or_create(mime_type=_format)[0]
for _format in new_formats
]
formats = MimeType.objects.filter(
mime_type__in=md_data_entry["formats"])
with transaction.atomic():
if len(md_data_entry["categories"]) > 0:
q = Q()
for cat in md_data_entry["categories"]:
q |= Q(title_EN__iexact=cat)
categories = Category.objects.filter(q)
else:
categories = []
for link in md_data_entry.get("links", []):
url = link.get("link", None)
if url is None:
continue
generic_url = GenericUrl()
generic_url.description = "[HARVESTED URL] \n{}".format(
link.get("description", ""))
generic_url.method = "Get"
generic_url.url = url
generic_url.save()
md.additional_urls.add(generic_url)
md.save(add_monitoring=False)
md.keywords.add(*kws)
md.categories.add(*categories)
md.formats.add(*formats)
# To reduce runtime, we only create a new MetadataRelation if we are sure there hasn't already been one.
# Using get_or_create increases runtime on existing metadata too much!
if is_new:
md.add_metadata_relation(
to_metadata=self.metadata,
relation_type=MetadataRelationEnum.HARVESTED_THROUGH.
value,
origin=ResourceOriginEnum.CATALOGUE.value)
parent_id = md_data_entry["parent_id"]
# Add the found parent_id to the parent_child map!
if parent_id is not None:
if self.parent_child_map.get(parent_id, None) is None:
self.parent_child_map[parent_id] = [md]
else:
self.parent_child_map[parent_id].append(md)
except (IntegrityError, DataError) as e:
csw_logger.error(
CSW_ERROR_LOG_TEMPLATE.format(md.identifier,
self.metadata.title, e))
def update_metadata(old: Metadata, new: Metadata, keep_custom_md: bool):
""" Overwrites existing metadata (old) with newer content (new).
Database related information like id, created_by, and so on is saved before and written back after overwriting.
Args:
old (Metadata): The existing metadata, that shall be overwritten
new (Metadata): The new metadata that is used for overwriting
Returns:
old (Metadata): The overwritten metadata
"""
# reset update candidate
new.is_update_candidate_for = None
# Save important persistance information
_id = old.id
created_by = old.created_by
created_on = old.created
activated = old.is_active
metadata_type = old.metadata_type
metadata_is_custom = old.is_custom
# If needed, cache custom metadata
custom_md = {}
if keep_custom_md:
custom_md_fields = MetadataEditorForm._meta.fields
for field in custom_md_fields:
custom_md[field] = old.__getattribute__(field)
del custom_md_fields
# Overwrite old information with new one
old = deepcopy(new)
old.id = _id
old.created = created_on
old.created_by = created_by
old.is_active = activated
old.metadata_type = metadata_type
# reference systems
old.reference_system.clear()
old.reference_system.add(*new.reference_system.all())
# Dimensions
old.dimensions.clear()
old.dimensions.add(*new.dimensions.all())
# formats
old.formats.clear()
old.formats.add(*new.formats_list)
# Restore custom metadata if needed
if keep_custom_md:
old.is_custom = metadata_is_custom
for key, val in custom_md.items():
# ManyRelatedManagers have to be handled differently
try:
old.__setattr__(key, val)
except TypeError:
# If the above simple attribute setter fails, we are dealing with a ManyRelatedManager, that has to be
# handled differently
field = val.prefetch_cache_name
old_manager = old.__getattribute__(field)
old_manager_elems = old_manager.all()
custom_m2m_elements = val.all()
for elem in custom_m2m_elements:
if elem not in old_manager_elems:
old_manager.add(elem)
else:
# Keywords updating without keeping custom md
old.keywords.clear()
old.keywords.add(*new.keywords.all())
old.last_modified = timezone.now()
return old