def init_rabbit_mq(self):
try:
self.logger.info("Initializing RabbitMQ message consumer...")
schedule_exchange = Exchange("airtime-media-monitor",
"direct",
durable=True,
auto_delete=True)
schedule_queue = Queue("media-monitor",
exchange=schedule_exchange,
key="filesystem")
self.connection = BrokerConnection(self.cfg["rabbitmq_host"],
self.cfg["rabbitmq_user"],
self.cfg["rabbitmq_password"],
self.cfg["rabbitmq_vhost"])
channel = self.connection.channel()
self.simple_queue = SimpleQueue(channel, schedule_queue)
self.logger.info("Initialized RabbitMQ consumer.")
except Exception as e:
self.logger.info("Failed to initialize RabbitMQ consumer")
self.logger.error(e)
return False
return True
def init_rabbit_mq(self):
self.logger.info("Initializing RabbitMQ stuff")
try:
self.logger.info("rabbitmq_host: " + self.config["rabbitmq_host"])
self.logger.info("rabbitmq_user: "******"rabbitmq_user"])
self.logger.info("rabbitmq_password: "******"rabbitmq_password"])
self.logger.info("rabbitmq_vhost: " +
self.config["rabbitmq_vhost"])
""""""
schedule_exchange = \
Exchange("airtime-pypo", "direct",
durable=True, auto_delete=True)
schedule_queue = \
Queue("pypo-fetch", exchange=schedule_exchange, key="foo")
connection = BrokerConnection(self.config["rabbitmq_host"], \
self.config["rabbitmq_user"], \
self.config["rabbitmq_password"], \
self.config["rabbitmq_vhost"])
channel = connection.channel()
self.simple_queue = SimpleQueue(channel, schedule_queue)
"""
connection = Connection('amqp://*****:*****@172.16.82.1:5672//pypox')
self.simple_queue = connection.SimpleQueue('pypo-fetch')
#message = simple_queue.get(block=True, timeout=1)
"""
except Exception, e:
self.logger.error(e)
return False
class InceptionerService:
"""Service endpoint for Inceptioner"""
name = "inceptioner_service"
test_exchange = Exchange('nameko_test_exchange', type='direct')
test_queue = Queue('nameko_test_queue', exchange=test_exchange)
@http('GET', '/get/<int:value>')
def get_method_for_test(self, request, value):
return json.dumps({'value': value})
@http('POST', '/recognize/base64')
def do_post(self, request):
logging.info('Received Request on recognition from base64')
request_data = request.data
logging.debug('Data Received: {}'.format(request_data))
res = process_request(request.data)
print(res)
return str(res)
@consume(test_queue)
def handle_event(self, payload):
logging.info('Received request on recognition on the queue')
logging.debug('Data received: {}'.format(payload))
res = process_request(payload)
print(res)
return res
def __init__(self, hostname="127.0.0.1", userid="guest", password="******",
virtual_host="/", port=5672, name="", routing_key=""):
""" Setup a connection to the AMQP server, get a channel
and create an exchange.
A specific service listener implementation overrides the name
and routing_key
"""
if name == "":
raise Exception("Name must be non-empty string")
self.name = name
self.routing_key = routing_key
if routing_key == "":
exchange_type = "fanout"
elif "*" in routing_key or "#" in routing_key:
exchange_type = "topic"
else :
exchange_type = "direct"
self.connection = BrokerConnection(hostname=hostname,
userid=userid, password=password,
virtual_host=virtual_host, port=443,
insist=False, ssl=False)
self.channel = self.connection.channel()
self.exchange = Exchange(name=self.name, type=exchange_type, durable=False,
channel=self.channel)
self.connection.connect()
return
class ConsumerService:
"""
Microservice responsible for consume data notifications from Service4 and
dispatching those data to the Client by saving those data to Redis database.
Attributes:
name (str): The microservice name.
_publication (Exchange): Messagin exchange object.
_queue (Queue): Messaging publications queue to be consumed.
_redis (Redis): Nameko Redis connector object.
"""
name = 'consumer'
_publication = Exchange('new_publication', type='direct')
_queue = Queue('publication_queue', exchange=_publication)
_redis = Redis('my_redis')
@consume(_queue)
def receive_new_publication(self, payload: str):
"""
Responsible for consuming incoming data received from service4 by
saving data to Redis Queue.
Args:
payload (str): Data to be consumed.
"""
try:
self._redis.rpush('publication_queue', payload)
except Exception as e:
print('Ooops!', e)
def init_rabbit_mq(self):
"""
This function will attempt to connect to RabbitMQ Server and if successful
return 'True'. Returns 'False' otherwise.
"""
self.logger.info("Initializing RabbitMQ stuff")
try:
schedule_exchange = Exchange("airtime-media-monitor",
"direct",
durable=True,
auto_delete=True)
schedule_queue = Queue("media-monitor",
exchange=schedule_exchange,
key="filesystem")
self.connection = BrokerConnection(
self.config.cfg["rabbitmq_host"],
self.config.cfg["rabbitmq_user"],
self.config.cfg["rabbitmq_password"],
self.config.cfg["rabbitmq_vhost"])
channel = self.connection.channel()
consumer = Consumer(channel, schedule_queue)
consumer.register_callback(self.handle_message)
consumer.consume()
except Exception, e:
self.logger.error(e)
return False
class Consumer:
name = 'worker_consumer'
test = Exchange('test', type='direct')
tq = Queue('q1', exchange=test)
@consume(tq)
def handle_consume(self, body):
print("Received message: {0}".format(body))
def setup_rabbit_mq_channel(self):
service_exchange = Exchange(self.acord_control_exchange, "topic", durable=False)
# connections/channels
connection = BrokerConnection(self.rabbit_host, self.rabbit_user, self.rabbit_password)
logging.info("Connection to RabbitMQ server successful")
channel = connection.channel()
# produce
self.producer = Producer(channel, exchange=service_exchange, routing_key='notifications.info')
self.publish = connection.ensure(self.producer, self.producer.publish, errback=self.errback, max_retries=3)
def get_exchanges(self):
"""
Define one exchange only for all messages and log. Routing
will be done only at the routing key level.
"""
# todo: use topic routing ?
# http://packages.python.org/kombu/reference/kombu.entity.html?#kombu.entity.Exchange.type
return {'psms': Exchange("psms", "direct", durable=self.persistent)}
class WorkerSubscriber:
name = 'worker_subscriber'
test = Exchange('test', type='direct')
publish = Publisher(exchange=test)
@event_handler("api", "say_hello")
def handle_event(self, payload):
print("{0} said hello!".format(payload))
self.publish("Goodbye {0}".format(payload))
def init_rabbit_mq(self):
self.logger.info("Initializing RabbitMQ stuff")
try:
schedule_exchange = Exchange("airtime-pypo", "direct", durable=True, auto_delete=True)
schedule_queue = Queue("pypo-fetch", exchange=schedule_exchange, key="foo")
connection = BrokerConnection(config["rabbitmq_host"], config["rabbitmq_user"], config["rabbitmq_password"], config["rabbitmq_vhost"])
channel = connection.channel()
self.simple_queue = SimpleQueue(channel, schedule_queue)
except Exception, e:
self.logger.error(e)
return False
def setup_rabbit_mq_channel(self):
service_exchange = Exchange(cfg.CONF.udpservice.acord_control_exchange, "topic", durable=False)
rabbit_host = cfg.CONF.udpservice.rabbit_hosts
rabbit_user = cfg.CONF.udpservice.rabbit_userid
rabbit_password = cfg.CONF.udpservice.rabbit_password
# connections/channels
connection = BrokerConnection(rabbit_host, rabbit_user, rabbit_password)
print 'Connection to RabbitMQ server successful'
channel = connection.channel()
# produce
self.producer = Producer(channel, exchange=service_exchange, routing_key='notifications.info')
class Sender():
""" Simple class to wrap the operations needed for an AMQP listener """
def __init__(self, hostname="127.0.0.1", userid="guest", password="******",
virtual_host="/", port=5672, name="", routing_key=""):
""" Setup a connection to the AMQP server, get a channel
and create an exchange.
A specific service listener implementation overrides the name
and routing_key
"""
if name == "":
raise Exception("Name must be non-empty string")
self.name = name
self.routing_key = routing_key
if routing_key == "":
exchange_type = "fanout"
elif "*" in routing_key or "#" in routing_key:
exchange_type = "topic"
else :
exchange_type = "direct"
self.connection = BrokerConnection(hostname=hostname,
userid=userid, password=password,
virtual_host=virtual_host, port=443,
insist=False, ssl=False)
self.channel = self.connection.channel()
self.exchange = Exchange(name=self.name, type=exchange_type, durable=False,
channel=self.channel)
self.connection.connect()
return
def send(self, msg):
""" Publishes a message to the AMQP server
on the initialized exchange
msg is a string, usually a JSON dump
"""
self.exchange.publish(self.exchange.Message(msg), routing_key=self.routing_key)
return
def setup_rabbit_mq_channel():
global producer
global rabbit_user, rabbit_password, rabbit_host, vcpeservice_rabbit_exchange,cpe_publisher_id
vcpeservice_exchange = Exchange(vcpeservice_rabbit_exchange, "topic", durable=False)
# connections/channels
connection = BrokerConnection(rabbit_host, rabbit_user, rabbit_password)
logger.info('Connection to RabbitMQ server successful')
channel = connection.channel()
# produce
producer = Producer(channel, exchange=vcpeservice_exchange, routing_key='notifications.info')
p = subprocess.Popen('hostname', shell=True, stdout=subprocess.PIPE)
(hostname, error) = p.communicate()
cpe_publisher_id = cpe_publisher_id + '_on_' + hostname
logger.info('cpe_publisher_id=%s',cpe_publisher_id)
def main():
cfg = {
'hostname': 'localhost',
'userid': 'guest',
'password': '******',
'virtual_host': '/',
'port': 5672
}
transport = 'pika'
#transport = 'librabbitmq'
connection = BrokerConnection(transport=transport, **cfg)
connection.connect()
cfg = {
'name': 'simple-test-1',
'auto_delete': True,
'durable': False,
'delivery_mode': 'transient'
}
channel = connection.channel()
exchange = Exchange(channel=channel, **cfg)
#exchange = exchange_def(channel)
routing_key = 'simple-test-1-route'
queue = Queue(exchange=exchange, routing_key=routing_key, **cfg)
channel = connection.channel()
producer = Producer(channel=channel,
exchange=exchange,
routing_key=routing_key)
channel = connection.channel()
consumer = Consumer(channel=channel, queues=[queue], callbacks=[receive])
consumer.consume()
def serve_forever():
while True:
#print 'drain'
#gevent.sleep(0.0001)
connection.drain_events(timeout=1)
def publish_forever():
while True:
producer.publish(loremIpsum)
gevent.sleep(0.0001)
#g1, g2 = gevent.spawn(publish_forever), gevent.spawn(serve_forever)
g2 = gevent.spawn(serve_forever)
g1 = gevent.spawn(publish_forever)
gevent.joinall([g1, g2])
class Service4:
"""
Microservice responsible for receiving data notifications from Service3 and
dispatching those data to the Client.
Attributes:
name (str): The microservice name.
_redis (Redis): Nameko Redis connector object.
_publication (Exchange): Messagin exchange object.
_publish (Publisher): Messaging publisher object.
"""
name = 'service4'
_redis = Redis('my_redis')
_publication = Exchange('new_publication', type='direct')
_publish = Publisher(exchange=_publication)
@event_handler('service3', 'number_published')
def receive_publication(self, payload: str):
"""
Event handler function receiver published number from service3
Args:
payload (str): A new number published according service3 rules
"""
self.dispatch_publication(payload)
@rpc
def dispatch_publication(self, payload: str):
"""
Notify an event with the passed payload
:param payload: A published number to be notify to the client
"""
self._publish(payload)
@rpc
def get_history(self) -> List[str]:
"""
Get the last 100 publications from Redis Database
Returns:
List[str]: Last publications
"""
if self._redis.llen('published_numbers') > 100:
history = self._redis.lrange('published_numbers', -100, -1)
else:
history = self._redis.lrange('published_numbers', 0, -1)
return history
def setup_rabbit_mq_channel(self):
ceilometer_exchange = Exchange(self.rabbit_exchange,
"topic",
durable=False)
# connections/channels
connection = BrokerConnection(self.rabbit_host, self.rabbit_user,
self.rabbit_password)
LOG.info(
"BroadViewPublisher: Connection to RabbitMQ server successful")
channel = connection.channel()
# produce
self._producer = Producer(channel,
exchange=ceilometer_exchange,
routing_key='notifications.info')
self._publish = connection.ensure(self._producer,
self._producer.publish,
errback=self.errback,
max_retries=3)
def init_rabbit_mq(self):
self.logger.info("Initializing RabbitMQ stuff")
try:
schedule_exchange = Exchange("airtime-pypo",
"direct",
durable=True,
auto_delete=True)
schedule_queue = Queue("pypo-fetch",
exchange=schedule_exchange,
key="foo")
with Connection(self.config["host"], \
self.config["user"], \
self.config["password"], \
self.config["vhost"], \
heartbeat = 5) as connection:
rabbit = RabbitConsumer(connection, [schedule_queue], self)
rabbit.run()
except Exception as e:
self.logger.error(e)
def init_rabbit_mq(self):
logger.info("Initializing RabbitMQ stuff")
try:
schedule_exchange = Exchange("airtime-pypo",
"direct",
durable=True,
auto_delete=True)
schedule_queue = Queue("pypo-fetch",
exchange=schedule_exchange,
key="foo")
with Connection(
f"amqp://{self.config.user}:{self.config.password}"
f"@{self.config.host}:{self.config.port}"
f"/{self.config.vhost}",
heartbeat=5,
) as connection:
rabbit = RabbitConsumer(connection, [schedule_queue], self)
rabbit.run()
except Exception as e:
logger.error(e)
def init_rabbit_mq(self):
self.logger.info("Initializing RabbitMQ stuff")
try:
schedule_exchange = Exchange("airtime-media-monitor",
"direct",
durable=True,
auto_delete=True)
schedule_queue = Queue("media-monitor",
exchange=schedule_exchange,
key="filesystem")
self.connection = BrokerConnection(
self.config.cfg["rabbitmq_host"],
self.config.cfg["rabbitmq_user"],
self.config.cfg["rabbitmq_password"], "/")
channel = self.connection.channel()
consumer = Consumer(channel, schedule_queue)
consumer.register_callback(self.handle_message)
consumer.consume()
except Exception, e:
self.logger.error(e)
return False
def init_rabbit_mq(self):
logger = logging.getLogger('fetch')
logger.info("Initializing RabbitMQ stuff")
try:
schedule_exchange = Exchange("airtime-schedule",
"direct",
durable=True,
auto_delete=True)
schedule_queue = Queue("pypo-fetch",
exchange=schedule_exchange,
key="foo")
self.connection = BrokerConnection(config["rabbitmq_host"],
config["rabbitmq_user"],
config["rabbitmq_password"],
"/")
channel = self.connection.channel()
consumer = Consumer(channel, schedule_queue)
consumer.register_callback(handle_message)
consumer.consume()
except Exception, e:
logger.error(e)
return False
def handle(self):
log.debug("Start consuming")
exchange = Exchange('amqp.topic', type='direct', durable=True)
self._connection = BrokerConnection(*CONNECTION)
channel = self._connection.channel()
for entry in self.queues:
log.debug("prepare to consume %s" % entry['routing_key'])
queue = Queue(entry['name'],
exchange=exchange,
routing_key=entry['routing_key'])
consumer = Consumer(channel, queue)
consumer.register_callback(getattr(self, entry['handler']))
consumer.consume()
log.debug("start consuming...")
while True:
try:
self._connection.drain_events()
except socket.timeout:
log.debug("nothing to consume...")
break
self._connection.close()
def deadlettering_exchange(self, rabbit_config, exchange, queue):
conn = Connection(rabbit_config[AMQP_URI_CONFIG_KEY])
with connections[conn].acquire(block=True) as connection:
deadletter_exchange = Exchange(name="deadletter", type="topic")
deadletter_exchange.maybe_bind(connection)
deadletter_exchange.declare()
deadletter_queue = Queue(
name="deadletter",
exchange=deadletter_exchange,
routing_key="#",
queue_arguments={'x-dead-letter-exchange': exchange.name})
deadletter_queue.maybe_bind(connection)
deadletter_queue.declare()
return deadletter_exchange
def _amqp_exchange():
return Exchange("rpc",
type="topic",
delivery_mode=TRANSIENT_DELIVERY_MODE,
durable=False)
class NotifierService(object):
name = 'notifier'
misaki = web.Slack('misaki')
error = ErrorHandler()
@property
def channel(self):
return f'#{os.getenv("NOTIFICATION_CHANNEL", "notifications")}'
@staticmethod
def _format_notification(input_):
keys = ('id', 'source', 'type', 'content')
if not all(k in input_.keys() for k in keys):
raise NotifierServiceError(
'Some keys are missing in the input dict')
blocks = [{
'type': 'section',
'text': {
'type': 'mrkdwn',
'text': f'*{input_["content"]}*',
}
}, {
'type':
'context',
'elements': [{
'type': 'mrkdwn',
'text': f'{k}: {input_[k]}'
} for k in input_ if k != 'content']
}]
return blocks
@consume(queue=Queue(name='evt_all_notifications',
exchange=Exchange(name='all_notifications',
type='topic',
auto_delete=True)))
def handle_all_notifications(self, payload):
_log.info(f'Received {payload}')
input_ = bson.json_util.loads(payload)
self.misaki.api_call('chat.postMessage',
channel=self.channel,
blocks=self._format_notification(input_),
text=input_['content'])
@rpc
def send_to_slack(self, channel, msg, image_url=None, context=None):
_log.info(f'Sending message {msg} to slack channel {channel} ...')
slack_msg = [{
'type': 'section',
'text': {
'type': 'mrkdwn',
'text': f'*{msg}*'
}
}]
if image_url:
slack_msg.extend([{
'type': 'section',
'text': {
'type':
'mrkdwn',
'text':
f'Please find your image at the following <{image_url}|link>'
}
}, {
'type': 'image',
'image_url': image_url,
'alt_text': 'Can not be displayed here'
}])
if context:
slack_msg.append({
'type': 'context',
'elements': [{
'type': 'mrkdwn',
'text': context
}]
})
self.misaki.api_call('chat.postMessage',
channel=channel,
blocks=slack_msg,
text=msg)
@rtm.handle_message
def handle_any_event(self, event, message):
_log.info(event)
_log.info(message)
class ElectionCollectorService(object):
name = 'election_collector'
database = MongoDatabase(result_backend=False)
election = Election()
error = ErrorHandler()
pub_input = Publisher(exchange=Exchange(name='all_inputs',
type='topic',
durable=True,
auto_delete=True,
delivery_mode=PERSISTENT))
pub_notif = Publisher(exchange=Exchange(name='all_notifications',
type='topic',
durable=True,
auto_delete=True,
delivery_mode=PERSISTENT))
def add_election(self, election_id):
self.database['elections'].update_one({'id': election_id},
{'$set': {
'id': election_id
}},
upsert=True)
@staticmethod
def handle_missing_number(doc, key):
if key not in doc:
return None
d = doc[key]
if 'Nombre' in doc[key]:
d = doc[key]['Nombre']
try:
return int(d)
except ValueError:
return None
@staticmethod
def to_boolean(doc, key):
if key not in doc:
return None
if doc[key] == 'O':
return True
return False
@staticmethod
def extract_scrutin(doc):
_log.info('Handling scrutin informations ...')
return {
'scrutin_type': doc['Type'],
'scrutin_annee': int(doc['Annee'])
}
@staticmethod
def extract_commune(doc):
_log.info('Handling commune informations ...')
return {
'commune_code': doc['CodSubCom'],
'commune_lib': doc['LibSubCom'],
'circonscription_code': doc.get('CodCirLg', None),
'circonscription_lib': doc.get('LibFraSubCom', None),
'mode_scrutin': doc.get('ModeScrutin', None)
}
@staticmethod
def extract_tour(doc):
_log.info('Handling tour informations ...')
return {'num_tour': int(doc['NumTour'])}
@staticmethod
def extract_mention(doc):
_log.info('Handling mention informations ...')
return {
'inscrits':
int(doc['Inscrits']['Nombre']),
'abstentions':
int(doc['Abstentions']['Nombre']),
'votants':
int(doc['Votants']['Nombre']),
'blancs':
ElectionCollectorService.handle_missing_number(doc, 'Blancs'),
'nuls':
ElectionCollectorService.handle_missing_number(doc, 'Nuls'),
'blancs_nuls':
ElectionCollectorService.handle_missing_number(
doc, 'BlancsOuNuls'),
'exprimes':
int(doc['Exprimes']['Nombre'])
}
@staticmethod
def extract_resultats(doc):
_log.info('Handling resultats informations ...')
return {
'nb_sap':
ElectionCollectorService.handle_missing_number(doc, 'NbSap'),
'nb_sp':
ElectionCollectorService.handle_missing_number(
doc, 'NbSiePourvus')
}
@staticmethod
def extract_departement(doc):
_log.info('Handling departement information ...')
return {
'departement_code': doc['CodDpt'],
'departement_lib': doc['LibDpt']
}
@staticmethod
def extract_nuance(doc):
_log.info('Handling nuance information ...')
return {
'nuance_code':
doc['CodNua'],
'nuance_lib':
doc['LibNua'],
'nb_voix':
int(doc['NbVoix']),
'nuance_nb_siege':
ElectionCollectorService.handle_missing_number(doc, 'NbSieges')
}
@staticmethod
def extract_candidat(doc):
_log.info('Handling candidat information ...')
return {
'candidat_numero':
ElectionCollectorService.handle_missing_number(
doc, 'NumPanneauCand'),
'candidat_nom':
doc['NomPsn'],
'candidat_prenom':
doc['PrenomPsn'],
'candidat_civilite':
doc['CivilitePsn'],
'candidat_ordre':
ElectionCollectorService.handle_missing_number(
doc, 'NumeroOrdCand'),
'candidat_elu':
ElectionCollectorService.to_boolean(doc, 'Elu'),
'nuance_code':
doc.get('CodNua', None),
'nuance_lib':
doc.get('LibNua', None),
'nb_voix':
int(doc['NbVoix'])
}
@staticmethod
def extract_liste(doc):
_log.info('Handling liste information ...')
return {
'liste_code':
doc['CodSeqLisCand'],
'liste_lib':
doc['NomListe'],
'liste_tete_nom':
doc.get('NomTeteListe', None),
'liste_tete_prenom':
doc.get('PrenomTeteListe', None),
'liste_tete_civilite':
doc.get('CiviliteTeteListe', None),
'liste_nb_elus':
ElectionCollectorService.handle_missing_number(doc, 'NbSieges'),
'nb_voix':
int(doc['NbVoix'])
}
@staticmethod
def complete_records(records, extract_func, rec_type, prev):
def create_record(r):
c = extract_func(r)
c.update(prev)
c['type'] = rec_type
for m in [meta[0] for meta in META]:
if m not in c:
c[m] = None
return c
return [create_record(r) for r in records]
@staticmethod
def build_records(doc, er):
_log.info(f'Building data from {er.url}')
election = doc['Election']
root = ElectionCollectorService.extract_scrutin(election['Scrutin'])
root['election_id'] = er.election_id
root['feed_id'] = er.feed_id
def ensure_list(d):
if isinstance(d, list):
return d
return [d]
def handle_tour(t, prev):
t_lvl = ElectionCollectorService.extract_tour(t)
t_lvl.update(prev)
t_lvl.update(
ElectionCollectorService.extract_mention(t['Mentions']))
res = t['Resultats']
t_lvl.update(ElectionCollectorService.extract_resultats(res))
if 'Nuances' in res:
return ElectionCollectorService.complete_records(
res['Nuances']['Nuance'],
ElectionCollectorService.extract_nuance, 'N', t_lvl)
elif 'Listes' in res:
return ElectionCollectorService.complete_records(
res['Listes']['Liste'],
ElectionCollectorService.extract_liste, 'L', t_lvl)
elif 'Candidats' in res:
return ElectionCollectorService.complete_records(
res['Candidats']['Candidat'],
ElectionCollectorService.extract_candidat, 'C', t_lvl)
else:
raise ElectionCollectorError(
'Cannot find neither Nuances, Listes nor Candidats under Resultats'
)
def handle_commune(c, prev):
c_lvl = ElectionCollectorService.extract_commune(c)
c_lvl.update(prev)
return list(
itertools.chain.from_iterable([
handle_tour(t, c_lvl)
for t in ensure_list(c['Tours']['Tour'])
]))
if 'Departement' in election:
root.update(
ElectionCollectorService.extract_departement(
election['Departement']))
return list(
itertools.chain.from_iterable([
handle_commune(c, root) for c in ensure_list(
election['Departement']['Communes']['Commune'])
]))
return list(
itertools.chain.from_iterable([
handle_tour(t, root)
for t in ensure_list(election['Tours']['Tour'])
]))
def update_checksum(self, id_, checksum):
self.database['elections'].update_one({'id': id_},
{'$set': {
'checksum': checksum
}})
@rpc
def publish(self, election_id):
_log.info(f'Publishing election {election_id} ...')
for r in self.election.results(election_id):
_log.info(f'Getting {r.url} ...')
doc = r.call()
try:
records = ElectionCollectorService.build_records(doc, r)
except ElectionCollectorError as e:
_log.error(f'Error on {r.url}: {str(e)}')
continue
data = {
'referential': {},
'datastore': [{
'write_policy': 'delete_bulk_insert',
'meta': META,
'target_table': 'french_election',
'delete_keys': {
'feed_id': r.feed_id
},
'records': records,
'chunk_size': 100
}],
'id':
r.feed_id,
'status':
'CREATED',
'checksum':
None,
'meta': {
'type': 'election',
'source': 'interieur',
'content_id': r.feed_id
}
}
self.pub_input(bson.json_util.dumps(data))
@event_handler('loader',
'input_loaded',
handler_type=BROADCAST,
reliable_delivery=False)
def ack(self, payload):
msg = bson.json_util.loads(payload)
meta = msg.get('meta', None)
if not meta:
return
if 'type' not in meta or 'source' not in meta or meta[
'source'] != 'interieur':
return
self.pub_notif(
bson.json_util.dumps({
'id': msg['id'],
'source': meta['source'],
'type': meta['type'],
'content': 'French election loaded!'
}))
@staticmethod
def is_meta_valid(msg):
if 'meta' not in msg:
return False
if 'type' not in msg['meta'] or 'source' not in msg['meta']:
return False
if msg['meta']['type'] != 'election' or msg['meta'][
'source'] != 'interieur':
return False
if 'config' not in msg:
_log.warning('Missing config within the message. Ignoring ...')
return False
if 'election' not in msg['config']:
_log.error('Missing mandatory field: election')
return False
return True
@event_handler('api_service',
'input_config',
handler_type=BROADCAST,
reliable_delivery=False)
def handle_input_config(self, payload):
msg = bson.json_util.loads(payload)
if not ElectionCollectorService.is_meta_valid(msg):
return
election_id = msg['config']['election']
_log.info('Received a related input config ...')
self.add_election(election_id)
self.publish(election_id)
def run(self):
print 'pypo Pusher'
if self.action == 'update_schedule':
print 'update_schedule!!'
credentials = pika.PlainCredentials(MQ_USER, MQ_PASS)
connection = pika.BlockingConnection(pika.ConnectionParameters(MQ_HOST,
5672,
'/airtime',
credentials))
channel = connection.channel()
channel.queue_declare(queue='pypo-fetch', durable=True)
message = {
'schedule': {
'media': {}
},
'event_type': 'update_schedule'
}
import json
message = json.dumps(message)
message = 'hallo'
channel.basic_publish(exchange='airtime-pypo',
routing_key='pypo-fetch',
body=message)
channel.close()
connection.close()
if self.action == 'update_schedule_kombu':
print 'update_schedule!!'
exchange = Exchange("airtime-pypo", "direct", durable=True, auto_delete=True)
queue = Queue("pypo-fetch", exchange=exchange, key="foo", durable=True)
connection = BrokerConnection(MQ_HOST, MQ_USER, MQ_PASS, MQ_VHOST)
channel = connection.channel()
simple_queue = SimpleQueue(channel, queue)
message = {
'schedule': {
'media': {}
},
'event_type': 'update_schedule'
}
print simple_queue.qsize()
print 'do:'
producer = Producer(channel, exchange=exchange, routing_key=None, serializer="json")
producer.publish(message, routing_key='pypo-fetch')
print simple_queue.qsize()
channel.close()
class LoaderService(object):
name = 'loader'
metadata = RpcProxy('metadata')
datastore = RpcProxy('datastore')
referential = RpcProxy('referential')
dispatch = EventDispatcher()
error = ErrorHandler()
@rpc
def write(self, write_policy, meta, target_table, records, upsert_key=None, delete_keys=None, chunk_size=None):
_log.info(
f'Writing in {target_table} using {write_policy} strategy ...')
if write_policy not in ('insert', 'upsert', 'bulk_insert', 'delete_insert', 'delete_bulk_insert',
'truncate_insert', 'truncate_bulk_insert'):
_log.error(f'{write_policy} not supported')
raise LoaderServiceError('Wrong value for parameter write_policy')
if write_policy in ('bulk_insert', 'delete_bulk_insert', 'truncate_bulk_insert') and not chunk_size:
_log.error('chunk_size missing')
raise LoaderServiceError(
'Bulk loading strategy requires a chunk size')
try:
meta = list(map(tuple, meta))
except:
_log.error('Bad formated meta')
raise LoaderServiceError('Bad formated meta')
if write_policy == 'insert':
self.datastore.insert(
target_table, bson.json_util.dumps(records), meta)
elif write_policy == 'upsert':
self.datastore.upsert(target_table, upsert_key,
bson.json_util.dumps(records), meta)
elif write_policy == 'bulk_insert':
self.datastore.bulk_insert(
target_table, bson.json_util.dumps(records), meta, chunk_size=chunk_size)
elif write_policy == 'delete_insert':
self.datastore.delete(target_table, delete_keys)
self.datastore.insert(
target_table, bson.json_util.dumps(records), meta)
elif write_policy == 'delete_bulk_insert':
self.datastore.delete(target_table, delete_keys)
self.datastore.bulk_insert(
target_table, bson.json_util.dumps(records), meta, chunk_size=chunk_size)
elif write_policy == 'truncate_insert':
self.datastore.truncate(target_table)
self.datastore.insert(
target_table, bson.json_util.dumps(records), meta)
else:
self.datastore.truncate(target_table)
self.datastore.bulk_insert(
target_table, bson.json_util.dumps(records), meta, chunk_size=chunk_size)
_log.info('Datastore microservice wrote all the records !')
return {'target_table': target_table, 'count': len(records)}
def _compute_transformation(self, t, param_value=None):
_log.info(f"Computing transformation {t['id']}")
try:
self.datastore.create_or_replace_python_function(
t['function_name'], t['function'])
except:
_log.error(
'Something went wrong while creating the underlying python function')
raise LoaderServiceError(
'An error occured while creating python function in transformation {}'.format(t['id']))
if t['type'] == 'fit' and t['process_date'] is None:
_log.info(
'Transformation has been set as \"fit\" kind. This must be processed !')
try:
last_entry = bson.json_util.loads(
self.datareader.select(t['output']))
if last_entry and len(last_entry) > 0:
_log.info('Deleting the previous result ...')
self.datastore.delete(t['target_table'], {
'id': last_entry[0]['id']})
_log.info('Computing current result ...')
self.datastore.insert_from_select(
t['target_table'], t['output'], None)
except:
_log.error(
'Something went wrong while deleting and inserting the result')
raise LoaderServiceError(
'An error occured while fitting transformation {}'.format(t['id']))
_log.info('Updating process date in metadata ...')
self.metadata.update_process_date(t['id'])
elif t['type'] in ('transform', 'predict',) and t['materialized'] is True:
_log.info(
'Transformation has been set as materialized \"transform\" or \"predict\" kind. This must be processed !')
if t['parameters'] is None:
_log.info('No parameters truncating the table ...')
self.datastore.truncate(t['target_table'])
_log.info('Inserting the result ...')
self.datastore.insert_from_select(
t['target_table'], t['output'], None)
else:
if len(t['parameters']) > 1:
raise LoaderServiceError(
'Does not support transformation with multiple parameters')
param_name = t['parameters'][0]
if param_value is None:
raise LoaderServiceError(
'Transformation requires a parameter')
_log.info(
'We will delete the previous result according to the provided parameter')
self.datastore.delete(t['target_table'], {
param_name: param_value})
_log.info('Inserting the result ...')
self.datastore.insert_from_select(
t['target_table'], t['output'], [param_value])
_log.info('Updating process date in metadata ...')
self.metadata.update_process_date(t['id'])
def update_transformations(self, trigger_table, param_value=None):
_log.info(f'Updating transformation related to {trigger_table} ...')
meta = self.metadata.get_update_pipeline(trigger_table)
if not meta:
_log.info('Nothing to do ...')
return {'trigger_table': trigger_table}
pipeline = bson.json_util.loads(meta)
for job in pipeline:
for t in job['transformations']:
self._compute_transformation(t, param_value)
return {'trigger_table': trigger_table}
def apply_transformation(self, transformation_id, param_value=None):
result = self.metadata.get_transformation(transformation_id)
transformation = bson.json_util.loads(result)
self._compute_transformation(transformation, param_value)
return {'id': transformation_id}
def update_entry_ngrams(self, entry_id):
return self.referential.update_entry_ngrams(entry_id)
def add_entity(self, data):
self.referential.add_entity(**data)
return {'id': data['id']}
def add_event(self, data):
data = self.referential.add_event(**data)
return {'id': data['id']}
def add_informations_to_entity(self, data):
data = self.referential.add_informations_to_entity(data['id'], data)
return {'id': data['id']}
@consume(queue=Queue(name='evt_all_inputs',
exchange=Exchange(name='all_inputs', type='topic', auto_delete=True)))
def handle_all_inputs(self, payload):
input_ = bson.json_util.loads(payload)
_log.info(f'Handling input {input_["id"]}')
if input_['status'] == 'UNCHANGED':
_log.info('Received an unchanged input ... ignoring !')
return
ref = input_['referential']
if ref.get('entities', None):
_log.info('Handling entities ...')
for e in ref['entities']:
ent = self.add_entity(e)
self.update_entry_ngrams(ent['id'])
if ref.get('events', None):
_log.info('Handling events ...')
for e in ref['events']:
evt = self.add_event(e)
self.update_entry_ngrams(evt['id'])
if ref.get('informations', None):
_log.info('Handling informations ...')
for e in ref['informations']:
self.add_informations_to_entity(e)
datastore = input_['datastore']
for d in datastore:
res = self.write(**d)
d_keys = d.get('delete_keys', None)
param_value = list(d_keys.values())[0] if d_keys else None
self.update_transformations(
res['target_table'], param_value=param_value)
ack = bson.json_util.dumps({
'id': input_['id'],
'checksum': input_.get('checksum', None),
'meta': input_.get('meta', None)})
self.dispatch('input_loaded', ack)
class SDMXCollectorService(object):
name = 'sdmx_collector'
database = MongoDatabase(result_backend=False)
sdmx = SDMX()
error = ErrorHandler()
pub_input = Publisher(exchange=Exchange(name='all_inputs',
type='topic',
durable=True,
auto_delete=True,
delivery_mode=PERSISTENT))
pub_notif = Publisher(exchange=Exchange(name='all_notifications',
type='topic',
durable=True,
auto_delete=True,
delivery_mode=PERSISTENT))
def add_dataflow(self, root_url, agency_id, resource_id, version, kind,
keys):
try:
self.sdmx.initialize(root_url, agency_id, resource_id, version,
kind, keys)
except Exception as e:
raise SDMXCollectorError(str(e))
self.database['dataset'].create_index([('agency', pymongo.ASCENDING),
('resource', pymongo.ASCENDING)
])
self.database['dataset'].create_index('id')
_id = SDMXCollectorService.table_name(agency_id, resource_id)
doc = {
'agency': agency_id,
'resource': resource_id,
'id': _id,
'version': version,
'kind': kind,
'keys': keys or {},
'root_url': root_url
}
self.database['dataset'].update_one(
{
'agency': agency_id,
'resource': resource_id
}, {'$set': doc},
upsert=True)
return _id
def get_dataflows(self):
return self.database['dataset'].find({})
@staticmethod
def clean(l):
return re.sub(r'[^0-9a-zA-Z_]+', '_', l)
@staticmethod
def table_name(provider, dataflow):
return f'{SDMXCollectorService.clean(provider.lower())}_{SDMXCollectorService.clean(dataflow.lower())}'
@staticmethod
def to_table_meta(meta, provider, dataflow):
table_name = SDMXCollectorService.table_name(provider, dataflow)
codelist = meta['codelist']
def handle_dim_att(d, is_dim=True):
name, code = d
cl = [c for c in codelist if c[0] == code]
if not cl:
return (SDMXCollectorService.clean(name), 'TEXT')
if is_dim:
size = functools.reduce(
lambda x, y: len(y[1]) if len(y[1]) > x else x, cl, 1)
return (SDMXCollectorService.clean(name), f'VARCHAR({size})')
size = functools.reduce(lambda x, y: x + len(y[1]), cl, 0)
return (SDMXCollectorService.clean(name), f'VARCHAR({size})')
table_meta = [handle_dim_att(d) for d in meta['dimensions'] if d[1]]
table_meta = table_meta + [
handle_dim_att(d, is_dim=False) for d in meta['attributes'] if d[1]
]
table_meta.append((SDMXCollectorService.clean(meta['time_dimension']),
'VARCHAR(20)'))
table_meta.append(
(SDMXCollectorService.clean(meta['primary_measure']), 'FLOAT'))
table_meta.append(('query', 'VARCHAR(250)'))
return {
'write_policy': 'delete_bulk_insert',
'meta': table_meta,
'target_table': table_name,
'chunk_size': 500,
'delete_keys': {
'query': meta['query']
}
}
@staticmethod
def codelist_table_meta(agency):
return {
'write_policy':
'upsert',
'meta': [('code_id', 'VARCHAR(250)'), ('id', 'VARCHAR(250)'),
('code_name', 'TEXT'), ('name', 'TEXT'),
('ref', 'VARCHAR(32)')],
'target_table':
f'{SDMXCollectorService.clean(agency).lower()}_codelist',
'upsert_key':
'ref'
}
@staticmethod
def checksum(data):
return hashlib.md5(''.join([str(r) for r in data
]).encode('utf-8')).hexdigest()
def get_status(self, provider, dataflow, checksum):
old = self.database['dataset'].find_one({
'agency': provider,
'resource': dataflow
})
if not old or 'checksum' not in old:
return 'CREATED'
if old['checksum'] == checksum:
return 'UNCHANGED'
return 'UPDATED'
@staticmethod
def dataflow_to_entity(df):
return {
'id': df['id'],
'common_name': df['name'],
'provider': df['structure']['agency_id'],
'type': 'dataflow',
'informations': df
}
def get_dataset(self, root_url, agency, resource, version, kind, keys):
self.sdmx.initialize(root_url, agency, resource, version, kind, keys)
meta = {
'name': self.sdmx.name(),
'codelist': self.sdmx.codelist(),
'dimensions': self.sdmx.dimensions(),
'attributes': self.sdmx.attributes(),
'primary_measure': self.sdmx.primary_measure(),
'time_dimension': self.sdmx.time_dimension(),
'query': self.sdmx.query()
}
table_meta = SDMXCollectorService.to_table_meta(meta, agency, resource)
def handle_number(m, v):
if m.lower() in ('float', 'double'):
try:
d = float(v)
if math.isnan(d):
return None
return d
except:
return None
return v
data = [{
k[0]: handle_number(
k[1],
r.get(k[0], None) if k[0] != 'query' else meta['query'])
for k in table_meta['meta']
} for r in self.sdmx.data()]
codelist_meta = SDMXCollectorService.codelist_table_meta(agency)
def hash_row(row):
h = hashlib.md5(str(row).encode('utf-8'))
return h.hexdigest()
codelist = [
dict(
zip([m[0] for m in codelist_meta['meta']],
r + (hash_row(r[0:2]), ))) for r in meta['codelist']
]
checksum = SDMXCollectorService.checksum(data)
return {
'referential': {
'entities': [{
'id': table_meta['target_table'],
'common_name': meta['name'],
'provider': 'internal',
'type': 'dataset',
'informations': {
'id': table_meta['target_table'],
'name': meta['name'],
'table': table_meta['target_table']
}
}, *[
SDMXCollectorService.dataflow_to_entity(d)
for d in self.sdmx.agency_dataflows
]]
},
'datastore': [{
**table_meta, 'records': data
}, {
**codelist_meta, 'records': codelist
}],
'checksum':
checksum,
'id':
table_meta['target_table'],
'status':
self.get_status(agency, resource, checksum),
'meta': {
'type': SDMXCollectorService.clean(agency).lower(),
'source': 'sdmx'
}
}
def update_checksum(self, id_, checksum):
self.database['dataset'].update_one({'id': id_},
{'$set': {
'checksum': checksum
}})
@timer(interval=24 * 60 * 60)
@rpc
def publish(self):
for f in self.get_dataflows():
agency = f['agency']
resource = f['resource']
root_url = f['root_url']
version = f['version']
kind = f['kind']
keys = f['keys']
_log.info(
f'Downloading dataset {resource} provided by {agency} ...')
try:
dataset = self.get_dataset(root_url, agency, resource, version,
kind, keys)
except Exception as e:
_log.error(
f'Can not handle dataset {resource} provided by {agency}: {str(e)}'
)
continue
_log.info('Publishing ...')
self.pub_input(bson.json_util.dumps(dataset))
@event_handler('loader',
'input_loaded',
handler_type=BROADCAST,
reliable_delivery=False)
def ack(self, payload):
msg = bson.json_util.loads(payload)
meta = msg.get('meta', None)
if not meta:
return
checksum = msg.get('checksum', None)
if not checksum:
return
if 'source' not in meta or meta['source'] != 'sdmx':
return
t = meta['type']
_log.info(f'Acknowledging {t} file: {msg["id"]}')
self.update_checksum(msg['id'], checksum)
_log.info(f'Publishing notification for {msg["id"]}')
self.pub_notif(
bson.json_util.dumps({
'id': msg['id'],
'source': 'sdmx',
'type': t,
'content': msg["id"]
}))
@event_handler('api_service',
'input_config',
handler_type=BROADCAST,
reliable_delivery=False)
def handle_input_config(self, payload):
msg = bson.json_util.loads(payload)
if 'meta' not in msg or 'source' not in msg[
'meta'] or msg['meta']['source'] != 'sdmx':
return
_log.info('Received a related input config ...')
if 'config' not in msg:
_log.warning('No config within the message. Ignoring ...')
return
config = msg['config']
if 'agency' not in config or 'resource' not in config or\
'version' not in config or 'kind' not in config or 'keys' not in config\
or 'root_url' not in config:
_log.error(
'Missing at least one of these mandatory fields: root_url, agency, resource, version, kind or keys'
)
return
id_ = self.add_dataflow(config['root_url'], config['agency'],
config['resource'], config['version'],
config['kind'], config['keys'])
self.pub_notif(
bson.json_util.dumps({
'id': id_,
'source': msg['meta']['source'],
'type': '',
'content': 'A new SDMX feed has been added.'
}))
self.publish()
def exchange(self, amqp_uri):
exchange = Exchange(name="exchange")
return exchange
def exchange(self):
backoff_exchange = Exchange(
type="headers",
name="backoff"
)
return backoff_exchange
def exchange():
return Exchange("messages")
