def _getResponsibleParty(self, node, ns):
# TODO: manage mutiplicities (e.g. online resource, phone, etc)
individualName = getTextFromXMLNode(
node, './gmd:individualName/gco:CharacterString/', ns)
organisationName = getTextFromXMLNode(
node, './gmd:organisationName/gco:CharacterString/', ns)
roleNode = getXMLNode(node, './gmd:role/gco:CharacterString/', ns)
role = getXMLCodeText(roleNode, ns=ns)
email = getTextFromXMLNode(
node,
'./gmd:contactInfo/gmd:CI_Contact/gmd:address/gmd:CI_Address/gmd:electronicMailAddress/gco:CharacterString/',
ns)
phone = getTextFromXMLNode(
node,
'./gmd:contactInfo/gmd:CI_Contact/gmd:phone/gmd:CI_Telephone/gmd:voice/gco:CharacterString/',
ns)
url = getTextFromXMLNode(
node,
'./gmd:contactInfo/gmd:CI_Contact/gmd:onlineResource/gmd:CI_OnlineResource/gmd:linkage/gmd:URL/',
ns)
onlineResources = []
for onlineResourceNode in node.findall(
'./gmd:contactInfo/gmd:CI_Contact/gmd:onlineResource/gmd:CI_OnlineResource'
):
onlineResource = self._getOnlineResource(onlineResourceNode, ns)
onlineResources.append(onlineResource)
return {
'individualName': individualName,
'organisationName': organisationName,
'role': role,
'email': email,
'phone': phone,
'url': url
}
def _getOnlineResource(self, node, ns):
url = getTextFromXMLNode(node, './gmd:linkage/gmd:URL', ns)
protocol = getTextFromXMLNode(node, './gmd:protocol/gco:CharacterString', ns)
name = getTextFromXMLNode(node, './gmd:name/gco:CharacterString', ns)
description = getTextFromXMLNode(node, './gmd:description/gco:CharacterString', ns)
applicationProfile = getTextFromXMLNode(node, './gmd:applicationProfile/gco:CharacterString', ns)
function = getTextFromXMLNode(node, './gmd:function/gco:CharacterString', ns)
return {
'name': sanitizeXmlText(name),
'description': sanitizeXmlText(description),
'applicationProfile': sanitizeXmlText(applicationProfile),
'function': sanitizeXmlText(function),
'protocol': sanitizeXmlText(protocol),
'url': sanitizeXmlText(url)
}
def gn_get_metadata(self, metadata_id):
#curl -X DELETE --header 'Accept: */*' 'http://localhost:8080/geonetwork/srv/api/0.1/records/97769e85-2e7b-418b-a8c8-0163bfb97aac?withBackup=false'
url = self.service_url + "/srv/api/0.1/records/" + str(
metadata_id) + ""
print("Getting metadata from uuid:" + str(metadata_id) + " -> " + url)
headers = {
'Accept': 'application/xml',
'X-XSRF-TOKEN': self.get_csrf_token()
}
r = self.session.get(url,
headers=headers,
timeout=get_default_timeout(),
proxies=settings.PROXIES)
if r.status_code == 200:
try:
tree = ET.fromstring(r.content)
logger.debug(r.content)
ns = {
'gmd': 'http://www.isotc211.org/2005/gmd',
'gco': 'http://www.isotc211.org/2005/gco'
}
metadata_id = getTextFromXMLNode(tree, './gmd:fileIdentifier/',
ns)
title = getTextFromXMLNode(
tree,
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:citation/gmd:CI_Citation/gmd:title/',
ns)
abstract = getTextFromXMLNode(
tree,
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:abstract/',
ns)
publish_date = getTextFromXMLNode(
tree,
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:citation/gmd:CI_Citation/gmd:date/gmd:CI_Date/gmd:date/',
ns)
period_start = ''
period_end = ''
aux = tree.findall(
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:extent/gmd:EX_Extent/gmd:temporalElement/gmd:EX_TemporalExtent/gmd:extent/',
ns)
if len(aux) > 0 and len(aux[0]) == 2:
period_start = aux[0][0].text
period_end = aux[0][1].text
categories = []
for category in tree.findall(
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:topicCategory/gmd:MD_TopicCategoryCode/',
ns):
categories.append(sanitizeXmlText(category.text))
keywords = []
for keyword in tree.findall(
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:descriptiveKeywords/gmd:MD_Keywords/gmd:keyword/',
ns):
keywords.append(sanitizeXmlText(keyword.text))
representation_type = ''
aux = tree.findall(
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:spatialRepresentationType/',
ns)
if len(aux) > 0:
representation_type = aux[0].attrib['codeListValue']
scale = getTextFromXMLNode(
tree,
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:spatialResolution/gmd:MD_Resolution/gmd:equivalentScale/gmd:MD_RepresentativeFraction/gmd:denominator/',
ns)
srs = getTextFromXMLNode(
tree,
'./gmd:referenceSystemInfo/gmd:MD_ReferenceSystem/gmd:referenceSystemIdentifier/gmd:RS_Identifier/gmd:code/',
ns)
#https://test.gvsigonline.com/geonetwork/srv/spa/region.getmap.png?mapsrs=EPSG:3857&width=250&background=settings&geomsrs=EPSG:4326&geom=Polygon((-18.1595005217%2043.9729489023,4.96320908311%2043.9729489023,4.96320908311%2025.9993588695,-18.1595005217%2025.9993588695,-18.1595005217%2043.9729489023))
coords_w = getTextFromXMLNode(
tree,
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:extent/gmd:EX_Extent/gmd:geographicElement/gmd:EX_GeographicBoundingBox/gmd:westBoundLongitude/',
ns)
coords_e = getTextFromXMLNode(
tree,
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:extent/gmd:EX_Extent/gmd:geographicElement/gmd:EX_GeographicBoundingBox/gmd:eastBoundLongitude/',
ns)
coords_s = getTextFromXMLNode(
tree,
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:extent/gmd:EX_Extent/gmd:geographicElement/gmd:EX_GeographicBoundingBox/gmd:southBoundLatitude/',
ns)
coords_n = getTextFromXMLNode(
tree,
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:extent/gmd:EX_Extent/gmd:geographicElement/gmd:EX_GeographicBoundingBox/gmd:northBoundLatitude/',
ns)
image_url = self.service_url + '/srv/spa/region.getmap.png?mapsrs=EPSG:3857&width=250&background=osm&geomsrs=EPSG:4326&geom=Polygon((' + coords_w + ' ' + coords_s + ',' + coords_e + ' ' + coords_s + ',' + coords_e + ' ' + coords_n + ',' + coords_w + ' ' + coords_n + ',' + coords_w + ' ' + coords_s + '))'
#thumbnails
thumbnails_urls = []
for thumbnail_url in tree.findall(
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:graphicOverview/gmd:MD_BrowseGraphic/gmd:fileName/',
ns):
thumbnails_urls.append(thumbnail_url.text)
thumbnail_names = []
for thumbnail_name in tree.findall(
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:graphicOverview/gmd:MD_BrowseGraphic/gmd:fileDescription/',
ns):
thumbnail_names.append(thumbnail_name.text)
thumbnails = []
if thumbnails_urls.__len__() == thumbnail_names.__len__():
for i in range(0, thumbnails_urls.__len__()):
thumbnail = {
'url': sanitizeXmlText(thumbnails_urls[i]),
'name': sanitizeXmlText(thumbnail_names[i])
}
thumbnails.append(thumbnail)
resource_constraints = self._getXMLConstraints(
tree,
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:resourceConstraints',
ns)
metadata_constraints = self._getXMLConstraints(
tree, './gmd:metadataConstraints', ns)
#resources
resources = []
for onlineResourceNode in tree.findall(
'./gmd:distributionInfo/gmd:MD_Distribution/gmd:transferOptions/gmd:MD_DigitalTransferOptions/gmd:onLine/gmd:CI_OnlineResource',
ns):
onlineResource = self._getOnlineResource(
onlineResourceNode, ns)
resources.append(onlineResource)
resource_contacts = []
for pointOfContact in tree.findall(
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:pointOfContact',
ns):
contact = {}
organisation = pointOfContact.find(
'./gmd:CI_ResponsibleParty/gmd:organisationName/gco:CharacterString',
ns)
role = pointOfContact.find(
'./gmd:CI_ResponsibleParty/gmd:role/gmd:CI_RoleCode',
ns)
if organisation is not None:
contact['organisation'] = sanitizeXmlText(
organisation.text)
if role is not None:
contact['role'] = sanitizeXmlText(getXMLCodeText(role))
contactInfoNode = pointOfContact.find(
'./gmd:CI_ResponsibleParty/gmd:contactInfo/gmd:CI_Contact/gmd:onlineResource/gmd:CI_OnlineResource',
ns)
if contactInfoNode is not None:
onlineResource = self._getOnlineResource(
contactInfoNode, ns)
contact['onlineResource'] = onlineResource
resource_contacts.append(contact)
metadata_contacts = []
for pointOfContact in tree.findall('./gmd:contact', ns):
contact = {}
organisation = pointOfContact.find(
'./gmd:CI_ResponsibleParty/gmd:organisationName/gco:CharacterString',
ns)
role = pointOfContact.find(
'./gmd:CI_ResponsibleParty/gmd:role/gmd:CI_RoleCode',
ns)
if organisation is not None:
contact['organisation'] = sanitizeXmlText(
organisation.text)
if role is not None:
contact['role'] = sanitizeXmlText(getXMLCodeText(role))
contactInfoNode = pointOfContact.find(
'./gmd:CI_ResponsibleParty/gmd:contactInfo/gmd:CI_Contact/gmd:onlineResource/gmd:CI_OnlineResource',
ns)
if contactInfoNode is not None:
onlineResource = self._getOnlineResource(
contactInfoNode, ns)
contact['onlineResource'] = onlineResource
metadata_contacts.append(contact)
responsible_parties = []
for responsibleParty in tree.findall(
'./gmd:identificationInfo/gmd:MD_DataIdentification/gmd:citation/gmd:CI_Citation/gmd:citedResponsibleParty/gmd:CI_ResponsibleParty',
ns):
contact = {}
organisation = responsibleParty.find(
'./gmd:organisationName/gco:CharacterString', ns)
role = responsibleParty.find('./gmd:role/gmd:CI_RoleCode',
ns)
if organisation is not None:
contact['organisation'] = sanitizeXmlText(
organisation.text)
if role is not None:
contact['role'] = sanitizeXmlText(getXMLCodeText(role))
contactInfoNode = responsibleParty.find(
'./gmd:contactInfo/gmd:CI_Contact/gmd:onlineResource/gmd:CI_OnlineResource',
ns)
if contactInfoNode is not None:
onlineResource = self._getOnlineResource(
contactInfoNode, ns)
contact['onlineResource'] = onlineResource
responsible_parties.append(contact)
contacts = {
'metadata_contacts': metadata_contacts,
'resource_contacts': resource_contacts,
'responsible_parties': responsible_parties
}
resource = {
'metadata_id': sanitizeXmlText(metadata_id),
'title': sanitizeXmlText(title),
'abstract': sanitizeXmlText(abstract),
'publish_date': sanitizeXmlText(publish_date),
'period_start': sanitizeXmlText(period_start),
'period_end': sanitizeXmlText(period_end),
'categories': categories,
'keywords': keywords,
'representation_type':
sanitizeXmlText(representation_type),
'scale': sanitizeXmlText(scale),
'srs': sanitizeXmlText(srs),
'image_url': sanitizeXmlText(image_url),
'thumbnails': thumbnails,
'resources': resources,
'resource_constraints': resource_constraints,
'metadata_constraints': metadata_constraints,
'contacts': contacts
}
return resource
except Exception as e:
logger.exception(e)
#return ET.tostring(tree, encoding='UTF-8')
#return r.content
logger.error(r.status_code)
logger.error(r.content)
raise FailedRequestError(r.status_code, r.content)