def _save_object_error(self,message,obj,stage=u'Fetch'):
'''
Helper function to create an error during the fetch or import stage.
'''
err = HarvestObjectError(message=message,object=obj,stage=stage)
err.save()
log.error(message)
def _save_object_error(self, message, obj, stage=u"Fetch", line=None):
err = HarvestObjectError(message=message, object=obj, stage=stage, line=line)
try:
err.save()
except InvalidRequestError, e:
Session.rollback()
err.save()
def _save_object_error(self, message, obj, stage=u'Fetch'):
'''
Helper function to create an error during the fetch or import stage.
'''
err = HarvestObjectError(message=message, object=obj, stage=stage)
err.save()
log.error(message)
def test_error_mail_sent_with_object_error(self, mock_mailer_mail_recipient):
context, harvest_source, harvest_job = self._create_harvest_source_and_job_if_not_existing()
data_dict = {
'guid': 'guid',
'content': 'content',
'job_id': harvest_job['id'],
'extras': {'a key': 'a value'},
'source_id': harvest_source['id']
}
harvest_object = toolkit.get_action('harvest_object_create')(
context, data_dict)
harvest_object_model = HarvestObject.get(harvest_object['id'])
# create a HarvestObjectError
msg = 'HarvestObjectError occured: %s' % harvest_job['id']
harvest_object_error = HarvestObjectError(message=msg, object=harvest_object_model)
harvest_object_error.save()
status = toolkit.get_action('harvest_source_show_status')(context, {'id': harvest_source['id']})
send_error_mail(
context,
harvest_source['id'],
status
)
assert_equal(1, status['last_job']['stats']['errored'])
assert mock_mailer_mail_recipient.called
def import_stage(self, harvest_object):
''' save to CKAN '''
logger.info('Importing {}'.format(harvest_object.id))
self.set_paths()
package_dict = json.loads(harvest_object.content)
action = package_dict.pop('action')
extras = package_dict.get('extras', [])
extras = self.update_extras(extras, harvest_object)
package_dict['extras'] = extras
resources = package_dict.pop('resources', [])
# Save (create or update) to CKAN
# Using base class function ._create_or_update_package
# seems no useful to deal with resources
user_name = self._get_user_name()
context = {'model': model, 'session': model.Session, 'user': user_name}
if action == 'create':
try:
pkg = p.toolkit.get_action('package_create')(context, package_dict)
except Exception, e:
logger.error('Error creating package {}: {}'.format(str(e), package_dict))
# TODO, no debería suceder
if str(e).find('already in use') > 0:
action = 'update'
else:
msg = 'Import CREATE error. pkg name: {}. \n\tError: {}'.format(package_dict.get('name', 'unnamed'), e)
harvest_object_error = HarvestObjectError(message=msg, object=harvest_object)
harvest_object_error.save()
return False
def is_part_of_to_package_id(self, ipo, harvest_object):
""" Get an identifier from external source using isPartOf
and returns the parent dataset or raises an ParentNotHarvestedException.
Only search for datasets that are the parent of a collection.
"""
ps = p.toolkit.get_action('package_search')
query = 'extras_identifier:{} AND extras_collection_metadata:true'.format(
ipo)
results = ps(self.context(), {"fq": query})
log.info('Package search results {}'.format(results))
if results[
'count'] > 0: # event if we have only one we need to be sure is the parent I need
# possible check identifier collision
# check the URL of the source to validate
datasets = results['results']
harvest_source = harvest_object.source
for dataset in datasets:
extras = dataset.get('extras', [])
identifiers = [
extra['value'] for extra in extras
if extra['key'] == 'identifier'
]
if ipo not in identifiers:
log.error('BAD SEARCH for {}:{}'.format(ipo, identifiers))
continue
dataset_harvest_source_id = self.get_harvest_source_id(
dataset['id'])
if harvest_source.id == dataset_harvest_source_id:
log.info('Parent dataset identified correctly')
return dataset
else:
log.info('{} not found at {} for {}'.format(
harvest_source.id, dataset_harvest_source_id, ipo))
# we have 0 o bad results
msg = 'Parent identifier not found: "{}"'.format(ipo)
log.error(msg)
try:
harvest_object_error = HarvestObjectError(message=msg,
object=harvest_object)
harvest_object_error.save()
harvest_object.state = "ERROR"
harvest_object.save()
except:
pass
raise ParentNotHarvestedException(
'Unable to find parent dataset. Raising error to allow re-run later'
)
def _save_object_error(self, message, obj, stage=u'Fetch', line=None):
err = HarvestObjectError(message=message,
object=obj,
stage=stage,
line=line)
try:
err.save()
except InvalidRequestError, e:
Session.rollback()
err.save()
class SIUTransparenciaHarvester(HarvesterBase):
def set_paths(self):
here = os.path.dirname(os.path.abspath(__file__))
base = os.environ.get('CKAN_STORAGE_PATH', here)
self.results_folder_path = os.path.join(base, 'siu-harvester-results')
if not os.path.isdir(self.results_folder_path):
os.makedirs(self.results_folder_path)
# librearia que gestiona los datos en el portal de SIU
self.siu_data_lib = SIUPoratlTransparenciaData()
## IHarvester
def info(self):
'''
:returns: A dictionary with the harvester descriptors
'''
return {
'name': 'siu_transp',
'title': 'SIU Portal de transparencia',
'description': 'Extraer y publicar datos del portal de transparecnia de SIU',
'form_config_interface': 'Text'
}
def validate_config(self, config):
'''
[optional]
Harvesters can provide this method to validate the configuration
entered in the form. It should return a single string, which will be
stored in the database. Exceptions raised will be shown in the form's
error messages.
:param config: Config string coming from the form
:returns: A string with the validated configuration options
'''
if not config:
raise ValueError('Set up the required configuration settings')
try:
config_obj = json.loads(config)
except ValueError as e:
raise e
# allow to get config from URL
# Sample: https://raw.githubusercontent.com/avdata99/ckan-env/develop/docs/full_config.json
config_from_url = config_obj.get('from_url', None)
if config_from_url is not None:
logger.info('Updating config from URL')
response = requests.get(config_from_url)
update_config = response.json()
config_obj.update(update_config)
required_cfg = ['username', 'password'] # , 'owner_org']
faileds = []
for req in required_cfg:
if req not in config_obj:
faileds.append(req)
if len(faileds) > 0:
raise ValueError('Missing configs: {}'.format(faileds))
return config
def gather_stage(self, harvest_job):
'''
analyze the source, return a list of IDs
and create one HarvestObject per dataset
'''
logger.info('Starts Gather SIU Transp')
# load paths
self.set_paths()
self.siu_data_lib.get_query_files()
# basic things you'll need
self.source = harvest_job.source
self.source_config = json.loads(self.source.config)
# allow to get config from URL
# Sample: https://raw.githubusercontent.com/avdata99/ckan-env/develop/docs/full_config.json
config_from_url = self.source_config.get('from_url', None)
if config_from_url is not None:
logger.info('Updating config from URL')
response = requests.get(config_from_url)
update_config = response.json()
self.source_config.update(update_config)
self.siu_data_lib.base_url = self.source.url
self.siu_data_lib.username = self.source_config['username']
self.siu_data_lib.password = self.source_config['password']
# ####################################
# get previous harvested packages
pfr = self.get_packages_for_source(harvest_source_id=self.source.id)
prev_names = [pkg['name'] for pkg in pfr['results']]
logger.info('Get previous harvested objects {}'.format(prev_names))
# TODO
# ####################################
object_ids = [] # lista de IDs a procesar, esto se devuelve en esta funcion
self.source_dataset = get_harvest_source(self.source.id)
owner_org = self.source_dataset.get('owner_org')
logger.info('Gather SIU Transp to ORG {}'.format(owner_org))
# Iterar por cada query para obtener diferentes conjuntos de datos
# Por cada archivo en siu_transp_data/queries se generarán múltiples datasets para publicar
report = [] # resumen de todos los resultados
logger.info('Iter files')
# ver si la config me pide sobreescribir metadatos en los datasets de cada archivo
override = self.source_config.get('override', {})
logger.info("General override {}".format(override))
for qf in self.siu_data_lib.query_files:
only_files = self.source_config.get('only_files', None)
query_file_name = qf.split('/')[-1]
if only_files is not None:
if query_file_name not in only_files:
logger.info('Skipping file by config {}'.format(query_file_name))
continue
logger.info('Gather SIU Transp FILE {}'.format(qf))
stqf = SIUTranspQueryFile(portal=self.siu_data_lib, path=qf)
# open to read query params
stqf.open()
# request all data
stqf.request_all(results_folder_path=self.results_folder_path)
for err in stqf.errors:
hgerr = HarvestGatherError(message=err, job=harvest_job)
hgerr.save()
# ====== Prepare dict to override datasets metadata ============
override_this = override.get(query_file_name, {})
logger.info("To override {}: {}".format(query_file_name, override_this))
# extras need to be {"key": "extra name", "value": "extra value"}
extras = override_this.get('extras', {})
new_extras = []
for extra_key, extra_value in extras.iteritems():
logger.info("Override extra found {}: {}".format(extra_key, extra_value))
if not isinstance(extra_value, str):
extra_value = str(extra_value)
new_extras.append({"key": extra_key, "value": extra_value})
if len(new_extras) > 0:
override_this['extras'] = new_extras
# tags need to be {"name": "tag name"}
tags = override_this.get('tags', [])
new_tags = []
for tag in tags:
logger.info("Override tag found {}".format(unicode(tag).encode("utf-8")))
new_tags.append({"name": tag})
if len(new_tags) > 0:
override_this['tags'] = new_tags
# groups need to be {"name": "tag name"}
groups = override_this.get('groups', [])
new_groups = []
for group in groups:
logger.info("Override group found {}".format(group))
# check if groups must be created
context = {'model': model, 'session': model.Session, 'user': self._get_user_name()}
try:
p.toolkit.get_action('group_create')(context, {"name": group})
except Exception as e:
logger.error('Error creating group (skipped) {}: {}'.format(group, e))
new_groups.append({"name": group})
if len(new_groups) > 0:
override_this['groups'] = new_groups
# ================================
report += stqf.requests
for dataset in stqf.datasets:
if dataset['name'] in prev_names:
action = 'update'
# leave this list just with packages to remove
prev_names.remove(dataset['name'])
else:
action = 'create'
logger.info('Dataset {} to {}'.format(dataset['name'], action))
ho_dict = {
'title': dataset['title'],
'name': dataset['name'],
'owner_org': owner_org,
'notes': dataset['notes'],
'tags': dataset['tags'],
'resources': dataset['resources'],
'action': action
}
# fix extras if they exists
ho_dict.update(override_this)
logger.info("Overrided ho_dict {}".format(ho_dict))
# Each harvest object will be passed to other stages in harvest process
obj = HarvestObject(guid=dataset['name'],
job=harvest_job,
content=json.dumps(ho_dict))
obj.save()
logger.info('Objects ID appends {}'.format(obj.id))
object_ids.append(obj.id)
# TODO compare with previous harvested data to remove dataset no more at harvest source
# resumen final
logger.info('REQUESTS: \n{}'.format('\n\t'.join(report)))
return object_ids
def fetch_stage(self, harvest_object):
''' donwload and get what you need before import to CKAN
Already downloaded in Gather stage '''
logger.info('Fetching {}'.format(harvest_object.id))
return True
def import_stage(self, harvest_object):
''' save to CKAN '''
logger.info('Importing {}'.format(harvest_object.id))
self.set_paths()
package_dict = json.loads(harvest_object.content)
action = package_dict.pop('action')
extras = package_dict.get('extras', [])
extras = self.update_extras(extras, harvest_object)
package_dict['extras'] = extras
resources = package_dict.pop('resources', [])
# Save (create or update) to CKAN
# Using base class function ._create_or_update_package
# seems no useful to deal with resources
user_name = self._get_user_name()
context = {'model': model, 'session': model.Session, 'user': user_name}
if action == 'create':
try:
pkg = p.toolkit.get_action('package_create')(context, package_dict)
except Exception, e:
logger.error('Error creating package {}: {}'.format(str(e), package_dict))
# TODO, no debería suceder
if str(e).find('already in use') > 0:
action = 'update'
else:
msg = 'Import CREATE error. pkg name: {}. \n\tError: {}'.format(package_dict.get('name', 'unnamed'), e)
harvest_object_error = HarvestObjectError(message=msg, object=harvest_object)
harvest_object_error.save()
return False
if action == 'update':
try:
pkg = p.toolkit.get_action('package_update')(context, package_dict)
except Exception, e:
msg = 'Import UPDATE error. pkg name: {}. \n\tError: {}'.format(package_dict.get('name', 'unnamed'), e)
harvest_object_error = HarvestObjectError(message=msg, object=harvest_object)
harvest_object_error.save()
logger.error(msg)
return False
def _save_object_error(self,message,obj,stage=u'Fetch'):
err = HarvestObjectError(message=message,object=obj,stage=stage)
err.save()
log.error(message)
def _get_xml_url_content(xml_url, urlopen_timeout, harvest_object):
try:
try:
r = requests.get(xml_url, timeout=urlopen_timeout)
ET.XML(r.content) # test for valid xml
return r
except ET.ParseError as e:
msg = '%s: %s. From external XML content at %s' % (
type(e).__name__, str(e), xml_url)
log.warn(msg)
err = HarvestObjectError(message=msg,
object=harvest_object,
stage='Import')
err.save()
except requests.exceptions.Timeout as e:
msg = '%s: %s. From external XML content at %s' % (
type(e).__name__, str(e), xml_url)
log.warn(msg)
err = HarvestObjectError(message=msg,
object=harvest_object,
stage='Import')
err.save()
except requests.exceptions.TooManyRedirects as e:
msg = 'HTTP too many redirects: %s' % e.code
log.warn(msg)
err = HarvestObjectError(message=msg,
object=harvest_object,
stage='Import')
err.save()
except requests.exceptions.RequestException as e:
msg = 'HTTP request exception: %s' % e.code
log.warn(msg)
err = HarvestObjectError(message=msg,
object=harvest_object,
stage='Import')
err.save()
except Exception as e:
msg = '%s: %s. From external XML content at %s' % (
type(e).__name__, str(e), xml_url)
log.warn(msg)
err = HarvestObjectError(message=msg,
object=harvest_object,
stage='Import')
err.save()
finally:
return ''
except StaleDataError as e:
log.warn('Harvest object %s is stail. Error object not created. %s' %
(harvest_object.id, str(e)))
