def _fill_metadata_db_model(self, metadata: Metadata):
""" Fills a Metadata db record from the ISOMetadata data
Args:
metadata (Metadata): The old/empty object
Returns:
metadata (Metadata): The metadata object
"""
metadata.identifier = self.file_identifier
metadata.abstract = self.abstract
metadata.access_constraints = self.access_constraints
# Take the polygon with the largest area as bounding geometry
if len(self.polygonal_extent_exterior) > 0:
max_area_poly = None
for poly in self.polygonal_extent_exterior:
if max_area_poly is None:
max_area_poly = poly
if max_area_poly.area < poly.area:
max_area_poly = poly
metadata.bounding_geometry = max_area_poly
try:
metadata.contact = Organization.objects.get_or_create(
organization_name=self.responsible_party,
email=self.contact_email,
)[0]
except MultipleObjectsReturned:
# okay, we need to create a unique organization
# "unique" since it will only be identified using organization_name and email
metadata.contact = Organization.objects.get_or_create(
organization_name="{}#1".format(self.responsible_party),
email=self.contact_email,
)[0]
metadata.is_inspire_conform = self.inspire_interoperability
metadata.metadata_url = self.uri
metadata.last_remote_change = self.last_change_date
metadata.spatial_res_type = self.spatial_res_type
metadata.spatial_res_value = self.spatial_res_val
if self.title is None:
self.title = "BROKEN"
metadata.title = self.title
metadata.origin = self.origin
metadata.is_broken = self.is_broken
metadata.save()
# save legal dates and reports
for report in self.legal_reports:
report.date.save()
report.save()
metadata.legal_reports.add(report)
for date in self.legal_dates:
date.save()
metadata.legal_dates.add(date)
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))