def test_dataset_to_resource(self):
data_dict = copy.deepcopy(self.test_data)
assert 'resources' not in data_dict
utils.dataset_to_resource(data_dict)
assert 'resources' in data_dict
utils.dataset_to_resource(data_dict)
assert 'resources' in data_dict
def test_dataset_to_resource_invalid(self):
data_dict = copy.deepcopy(self.test_data)
data_dict.pop('direct_download_URL')
data_dict.pop('checksum')
data_dict.pop('mimetype')
assert 'resources' not in data_dict
utils.dataset_to_resource(data_dict)
# dataset_to_resource can handle missing data, so resources is created
assert 'resources' in data_dict
def package_update(context, data_dict):
'''
Updates the dataset.
Extends ckan's similar method to instantly re-index the SOLR index.
Otherwise the changes would only be added during a re-index (a rebuild of search index,
to be specific).
:type context: dict
:param context: context
:type data_dict: dict
:param data_dict: dataset as dictionary
:rtype: dictionary
'''
# Get all resources here since we get only 'dataset' resources from WUI.
package_context = {'model': model, 'ignore_auth': True, 'validate': True,
'extras_as_string': True}
user = model.User.get(context['user'])
if not user.name == "harvest":
_remove_extras_from_data_dict(data_dict)
package_data = package_show(package_context, data_dict)
if not 'resources' in data_dict:
# When this is reached, we are updating a dataset, not creating a new resource
old_resources = package_data.get('resources', [])
data_dict['resources'] = old_resources
data_dict = utils.dataset_to_resource(data_dict)
else:
data_dict['accept-terms'] = 'yes' # This is not needed when adding a resource
_handle_pids(data_dict)
_add_ida_download_url(data_dict)
if asbool(data_dict.get('private')) and not data_dict.get('persist_schema'):
context['schema'] = Schemas.private_package_schema()
data_dict.pop('persist_schema', False)
if package_data.get('type') == 'harvest':
context['schema'] = Schemas.harvest_source_update_package_schema()
pkg_dict1 = ckan.logic.action.update.package_update(context, data_dict)
# Logging for production use
_log_action('Package', 'update', context['user'], data_dict['id'])
context = {'model': model, 'ignore_auth': True, 'validate': False,
'extras_as_string': True}
pkg_dict = ckan.logic.action.get.package_show(context, pkg_dict1)
index = index_for('package')
# update_dict calls index_package, so it would basically be the same
index.update_dict(pkg_dict)
return pkg_dict1
def test_resource_handling_2(self):
data_dict = copy.deepcopy(self.test_data3)
utils.resource_to_dataset(data_dict)
assert 'direct_download_URL' in data_dict
assert 'resources' in data_dict
data_dict['availability'] = 'direct_download'
utils.dataset_to_resource(data_dict)
assert 'resources' in data_dict
utils.resource_to_dataset(data_dict)
assert 'resources' in data_dict
assert data_dict.get('algorithm') == self.test_data3['resources'][0]['algorithm']
assert data_dict.get('checksum') == self.test_data3['resources'][0]['hash']
assert data_dict.get('mimetype') == self.test_data3['resources'][0]['mimetype']
assert data_dict.get('direct_download_URL') == self.test_data3['resources'][0]['url']
def package_create(context, data_dict):
"""
Creates a new dataset.
Extends ckan's similar method to instantly reindex the SOLR index,
so that this newly added package emerges in search results instantly instead of
during the next timed reindexing.
:param context: context
:param data_dict: data dictionary (package data)
:rtype: dictionary
"""
user = model.User.get(context['user'])
if data_dict.get('type') == 'harvest' and not user.sysadmin:
ckan.lib.base.abort(401, _('Unauthorized to add a harvest source'))
if not user.name == "harvest":
_remove_extras_from_data_dict(data_dict)
data_dict = utils.dataset_to_resource(data_dict)
if not user.name == 'harvest':
_handle_package_id_on_create(data_dict)
_handle_pids(data_dict)
_add_ida_download_url(data_dict)
if asbool(data_dict.get('private')) and not data_dict.get('persist_schema'):
context['schema'] = Schemas.private_package_schema()
data_dict.pop('persist_schema', False)
if data_dict.get('type') == 'harvest':
context['schema'] = Schemas.harvest_source_create_package_schema()
pkg_dict1 = ckan.logic.action.create.package_create(context, data_dict)
# Logging for production use
_log_action('Package', 'create', context['user'], pkg_dict1['id'])
context = {'model': model, 'ignore_auth': True, 'validate': False,
'extras_as_string': False}
pkg_dict = ckan.logic.action.get.package_show(context, pkg_dict1)
index = index_for('package')
index.index_package(pkg_dict)
return pkg_dict1
def package_create(context, data_dict):
"""
Creates a new dataset.
Extends ckan's similar method to instantly reindex the SOLR index,
so that this newly added package emerges in search results instantly instead of
during the next timed reindexing.
:param context: context
:param data_dict: data dictionary (package data)
:rtype: dictionary
"""
user = model.User.get(context['user'])
if data_dict.get('type') == 'harvest' and not user.sysadmin:
ckan.lib.base.abort(401, _('Unauthorized to add a harvest source'))
data_dict = utils.dataset_to_resource(data_dict)
_handle_pids(context, data_dict)
_add_ida_download_url(context, data_dict)
if asbool(data_dict.get('private')) and not data_dict.get('persist_schema'):
context['schema'] = Schemas.private_package_schema()
data_dict.pop('persist_schema', False)
if data_dict.get('type') == 'harvest':
context['schema'] = Schemas.harvest_source_create_package_schema()
pkg_dict1 = ckan.logic.action.create.package_create(context, data_dict)
# Logging for production use
_log_action('Package', 'create', context['user'], pkg_dict1['id'])
context = {'model': model, 'ignore_auth': True, 'validate': False,
'extras_as_string': False}
pkg_dict = ckan.logic.action.get.package_show(context, pkg_dict1)
index = index_for('package')
index.index_package(pkg_dict)
return pkg_dict1
def package_create(context, data_dict):
"""
Creates a new dataset.
Extends ckan's similar method to instantly reindex the SOLR index,
so that this newly added package emerges in search results instantly instead of
during the next timed reindexing.
:param context: context
:param data_dict: data dictionary (package data)
:rtype: dictionary
"""
user = model.User.get(context['user'])
try:
if data_dict['type'] == 'harvest' and not user.sysadmin:
ckan.lib.base.abort(401, _('Unauthorized to add a harvest source'))
except KeyError:
log.debug("Tried to check the package type, but it wasn't present!")
# TODO: JUHO: Dubious to let pass without checking user.sysadmin
pass
data_dict = utils.dataset_to_resource(data_dict)
_handle_pids(context, data_dict)
_add_ida_download_url(context, data_dict)
if data_dict.get('type') == 'harvest':
context['schema'] = Schemas.harvest_source_create_package_schema()
pkg_dict1 = ckan.logic.action.create.package_create(context, data_dict)
# Logging for production use
_log_action('Package', 'create', context['user'], pkg_dict1['id'])
context = {'model': model, 'ignore_auth': True, 'validate': False,
'extras_as_string': False}
pkg_dict = ckan.logic.action.get.package_show(context, pkg_dict1)
index = index_for('package')
index.index_package(pkg_dict)
return pkg_dict1
def package_update(context, data_dict):
'''
Updates the dataset.
Extends ckan's similar method to instantly re-index the SOLR index.
Otherwise the changes would only be added during a re-index (a rebuild of search index,
to be specific).
:type context: dict
:param context: context
:type data_dict: dict
:param data_dict: dataset as dictionary
:rtype: dictionary
'''
# Get all resources here since we get only 'dataset' resources from WUI.
package_context = {'model': model, 'ignore_auth': True, 'validate': True,
'extras_as_string': True}
package_data = package_show(package_context, data_dict)
# package_data = ckan.logic.action.get.package_show(package_context, data_dict)
old_resources = package_data.get('resources', [])
if not 'resources' in data_dict:
# When this is reached, we are updating a dataset, not creating a new resource
data_dict['resources'] = old_resources
data_dict = utils.dataset_to_resource(data_dict)
else:
data_dict['accept-terms'] = 'yes' # This is not needed when adding a resource
_handle_pids(context, data_dict)
_add_ida_download_url(context, data_dict)
# # Check if data version has changed and if so, generate a new version_PID
# if not data_dict['version'] == temp_pkg_dict['version']:
# data_dict['pids'].append(
# {
# u'provider': u'kata',
# u'id': utils.generate_pid(),
# u'type': u'version',
# })
if asbool(data_dict.get('private')) and not data_dict.get('persist_schema'):
context['schema'] = Schemas.private_package_schema()
data_dict.pop('persist_schema', False)
if package_data.get('type') == 'harvest':
context['schema'] = Schemas.harvest_source_update_package_schema()
pkg_dict1 = ckan.logic.action.update.package_update(context, data_dict)
# Logging for production use
_log_action('Package', 'update', context['user'], data_dict['id'])
context = {'model': model, 'ignore_auth': True, 'validate': False,
'extras_as_string': True}
pkg_dict = ckan.logic.action.get.package_show(context, pkg_dict1)
index = index_for('package')
# update_dict calls index_package, so it would basically be the same
index.update_dict(pkg_dict)
return pkg_dict1
def package_update(context, data_dict):
'''
Updates the dataset.
Extends ckan's similar method to instantly re-index the SOLR index.
Otherwise the changes would only be added during a re-index (a rebuild of search index,
to be specific).
:type context: dict
:param context: context
:type data_dict: dict
:param data_dict: dataset as dictionary
:rtype: dictionary
'''
# Get all resources here since we get only 'dataset' resources from WUI.
package_context = {'model': model, 'ignore_auth': True, 'validate': True,
'extras_as_string': True}
package_data = package_show(package_context, data_dict)
# package_data = ckan.logic.action.get.package_show(package_context, data_dict)
old_resources = package_data.get('resources', [])
if not 'resources' in data_dict:
# When this is reached, we are updating a dataset, not creating a new resource
data_dict['resources'] = old_resources
data_dict = utils.dataset_to_resource(data_dict)
_handle_pids(context, data_dict)
_add_ida_download_url(context, data_dict)
# # Check if data version has changed and if so, generate a new version_PID
# if not data_dict['version'] == temp_pkg_dict['version']:
# data_dict['pids'].append(
# {
# u'provider': u'kata',
# u'id': utils.generate_pid(),
# u'type': u'version',
# })
# This fixes extras fields being cleared when adding a resource. This is be because the extras are not properly
# cleared in show_package_schema conversions. Some fields stay in extras and they cause all other fields to be
# dropped in package_update(). When updating a dataset via UI or API, the conversion to extras occur in
# package_update() and popping extras here should have no effect.
data_dict.pop('extras', None)
# TODO: MIKKO: Get rid of popping extras here and rather pop the additional extras in converters so we could remove the
# popping and the above "context['allow_partial_update'] = True" which causes the extras to be processed in a way
# that nothing gets added to extras from the converters and everything not initially present in extras gets removed.
# TODO: JUHO: Apply correct schema depending on dataset
# This is quick resolution. More robust way would be to check through
# model.Package to which harvest source the dataset belongs and then get the
# type of the harvester (eg. DDI)
# if data_dict['name'].startswith('FSD'):
# context['schema'] = schemas.update_package_schema_ddi()
if package_data.get('type') == 'harvest':
context['schema'] = Schemas.harvest_source_update_package_schema()
pkg_dict1 = ckan.logic.action.update.package_update(context, data_dict)
# Logging for production use
_log_action('Package', 'update', context['user'], data_dict['id'])
context = {'model': model, 'ignore_auth': True, 'validate': False,
'extras_as_string': True}
pkg_dict = ckan.logic.action.get.package_show(context, pkg_dict1)
index = index_for('package')
# update_dict calls index_package, so it would basically be the same
index.update_dict(pkg_dict)
return pkg_dict1
