def start_output_stream_and_listen(self,
ctd_stream_id,
data_product_stream_ids,
message_count_per_stream=10):
assertions = self.assertTrue
exchange_name = 'workflow_test'
###
### Make a subscriber in the test to listen for transformed data
###
salinity_subscription_id = self.pubsubclient.create_subscription(
name='test workflow transformations',
exchange_name=exchange_name,
stream_ids=data_product_stream_ids)
result = gevent.event.AsyncResult()
results = []
message_count = len(data_product_stream_ids) * message_count_per_stream
def message_received(message, stream_route, stream_id):
# Heads
results.append(message)
if len(
results
) >= message_count: #Only wait for so many messages - per stream
result.set(True)
subscriber = StandaloneStreamSubscriber(exchange_name='workflow_test',
callback=message_received)
subscriber.xn.purge()
self.addCleanup(subscriber.xn.delete)
subscriber.start()
# after the queue has been created it is safe to activate the subscription
self.pubsubclient.activate_subscription(
subscription_id=salinity_subscription_id)
#Start the input stream process
if ctd_stream_id is not None:
ctd_sim_pid = self.start_simple_input_stream_process(ctd_stream_id)
# Assert that we have received data
assertions(result.get(timeout=30))
# stop the flow parse the messages...
if ctd_stream_id is not None:
self.process_dispatcher.cancel_process(
ctd_sim_pid
) # kill the ctd simulator process - that is enough data
self.pubsubclient.deactivate_subscription(
subscription_id=salinity_subscription_id)
subscriber.stop()
return results
def test_full_pubsub(self):
self.sub1_sat = Event()
self.sub2_sat = Event()
def subscriber1(m, r, s):
self.sub1_sat.set()
def subscriber2(m, r, s):
self.sub2_sat.set()
sub1 = StandaloneStreamSubscriber("sub1", subscriber1)
self.queue_cleanup.append(sub1.xn.queue)
sub1.start()
sub2 = StandaloneStreamSubscriber("sub2", subscriber2)
self.queue_cleanup.append(sub2.xn.queue)
sub2.start()
log_topic = self.pubsub_management.create_topic("instrument_logs", exchange_point="instruments")
science_topic = self.pubsub_management.create_topic("science_data", exchange_point="instruments")
events_topic = self.pubsub_management.create_topic("notifications", exchange_point="events")
log_stream, route = self.pubsub_management.create_stream(
"instrument1-logs", topic_ids=[log_topic], exchange_point="instruments"
)
ctd_stream, route = self.pubsub_management.create_stream(
"instrument1-ctd", topic_ids=[science_topic], exchange_point="instruments"
)
event_stream, route = self.pubsub_management.create_stream(
"notifications", topic_ids=[events_topic], exchange_point="events"
)
raw_stream, route = self.pubsub_management.create_stream("temp", exchange_point="global.data")
self.exchange_cleanup.extend(["instruments", "events", "global.data"])
subscription1 = self.pubsub_management.create_subscription(
"subscription1", stream_ids=[log_stream, event_stream], exchange_name="sub1"
)
subscription2 = self.pubsub_management.create_subscription(
"subscription2", exchange_points=["global.data"], stream_ids=[ctd_stream], exchange_name="sub2"
)
self.pubsub_management.activate_subscription(subscription1)
self.pubsub_management.activate_subscription(subscription2)
self.publish_on_stream(log_stream, 1)
self.assertTrue(self.sub1_sat.wait(4))
self.assertFalse(self.sub2_sat.is_set())
self.publish_on_stream(raw_stream, 1)
self.assertTrue(self.sub1_sat.wait(4))
sub1.stop()
sub2.stop()
def test_demux(self):
self.stream0, self.route0 = self.pubsub_client.create_stream(
'stream0', exchange_point='test')
self.stream1, self.route1 = self.pubsub_client.create_stream(
'stream1', exchange_point='main_data')
self.stream2, self.route2 = self.pubsub_client.create_stream(
'stream2', exchange_point='alt_data')
self.r_stream1 = gevent.event.Event()
self.r_stream2 = gevent.event.Event()
def process(msg, stream_route, stream_id):
if stream_id == self.stream1:
self.r_stream1.set()
elif stream_id == self.stream2:
self.r_stream2.set()
self.container.spawn_process(
'demuxer', 'ion.processes.data.transforms.mux', 'DemuxTransform',
{'process': {
'out_streams': [self.stream1, self.stream2]
}}, 'demuxer_pid')
self.queue_cleanup.append('demuxer_pid')
sub1 = StandaloneStreamSubscriber('sub1', process)
sub2 = StandaloneStreamSubscriber('sub2', process)
sub1.xn.bind(self.route1.routing_key,
self.container.ex_manager.create_xp('main_data'))
sub2.xn.bind(self.route2.routing_key,
self.container.ex_manager.create_xp('alt_data'))
sub1.start()
sub2.start()
self.queue_cleanup.append(sub1.xn)
self.queue_cleanup.append(sub2.xn)
xn = self.container.ex_manager.create_xn_queue('demuxer_pid')
xn.bind(
self.route0.routing_key,
self.container.ex_manager.create_xp(self.route0.exchange_point))
domino = StandaloneStreamPublisher(self.stream0, self.route0)
domino.publish('test')
self.assertTrue(self.r_stream1.wait(2))
self.assertTrue(self.r_stream2.wait(2))
self.container.proc_manager.terminate_process('demuxer_pid')
sub1.stop()
sub2.stop()
def start_output_stream_and_listen(self, ctd_stream_id, data_product_stream_ids, message_count_per_stream=10):
assertions = self.assertTrue
exchange_name = 'workflow_test'
###
### Make a subscriber in the test to listen for transformed data
###
salinity_subscription_id = self.pubsubclient.create_subscription(
name = 'test workflow transformations',
exchange_name = exchange_name,
stream_ids = data_product_stream_ids
)
result = gevent.event.AsyncResult()
results = []
message_count = len(data_product_stream_ids) * message_count_per_stream
def message_received(message, stream_route, stream_id):
# Heads
results.append(message)
if len(results) >= message_count: #Only wait for so many messages - per stream
result.set(True)
subscriber = StandaloneStreamSubscriber(exchange_name='workflow_test', callback=message_received)
subscriber.xn.purge()
self.addCleanup(subscriber.xn.delete)
subscriber.start()
# after the queue has been created it is safe to activate the subscription
self.pubsubclient.activate_subscription(subscription_id=salinity_subscription_id)
#Start the input stream process
if ctd_stream_id is not None:
ctd_sim_pid = self.start_simple_input_stream_process(ctd_stream_id)
# Assert that we have received data
assertions(result.get(timeout=30))
# stop the flow parse the messages...
if ctd_stream_id is not None:
self.process_dispatcher.cancel_process(ctd_sim_pid) # kill the ctd simulator process - that is enough data
self.pubsubclient.deactivate_subscription(subscription_id=salinity_subscription_id)
subscriber.stop()
return results
def test_full_pubsub(self):
self.sub1_sat = Event()
self.sub2_sat = Event()
def subscriber1(m,r,s):
self.sub1_sat.set()
def subscriber2(m,r,s):
self.sub2_sat.set()
sub1 = StandaloneStreamSubscriber('sub1', subscriber1)
self.queue_cleanup.append(sub1.xn.queue)
sub1.start()
sub2 = StandaloneStreamSubscriber('sub2', subscriber2)
self.queue_cleanup.append(sub2.xn.queue)
sub2.start()
log_topic = self.pubsub_management.create_topic('instrument_logs', exchange_point='instruments')
science_topic = self.pubsub_management.create_topic('science_data', exchange_point='instruments')
events_topic = self.pubsub_management.create_topic('notifications', exchange_point='events')
log_stream, route = self.pubsub_management.create_stream('instrument1-logs', topic_ids=[log_topic], exchange_point='instruments')
ctd_stream, route = self.pubsub_management.create_stream('instrument1-ctd', topic_ids=[science_topic], exchange_point='instruments')
event_stream, route = self.pubsub_management.create_stream('notifications', topic_ids=[events_topic], exchange_point='events')
raw_stream, route = self.pubsub_management.create_stream('temp', exchange_point='global.data')
self.exchange_cleanup.extend(['instruments','events','global.data'])
subscription1 = self.pubsub_management.create_subscription('subscription1', stream_ids=[log_stream,event_stream], exchange_name='sub1')
subscription2 = self.pubsub_management.create_subscription('subscription2', exchange_points=['global.data'], stream_ids=[ctd_stream], exchange_name='sub2')
self.pubsub_management.activate_subscription(subscription1)
self.pubsub_management.activate_subscription(subscription2)
self.publish_on_stream(log_stream, 1)
self.assertTrue(self.sub1_sat.wait(4))
self.assertFalse(self.sub2_sat.is_set())
self.publish_on_stream(raw_stream,1)
self.assertTrue(self.sub1_sat.wait(4))
sub1.stop()
sub2.stop()
def test_activation_and_deactivation(self):
stream_id, route = self.pubsub_management.create_stream(
'stream1', 'xp1')
subscription_id = self.pubsub_management.create_subscription(
'sub1', stream_ids=[stream_id])
self.check1 = Event()
def verifier(m, r, s):
self.check1.set()
subscriber = StandaloneStreamSubscriber('sub1', verifier)
subscriber.start()
publisher = StandaloneStreamPublisher(stream_id, route)
publisher.publish('should not receive')
self.assertFalse(self.check1.wait(0.25))
self.pubsub_management.activate_subscription(subscription_id)
publisher.publish('should receive')
self.assertTrue(self.check1.wait(2))
self.check1.clear()
self.assertFalse(self.check1.is_set())
self.pubsub_management.deactivate_subscription(subscription_id)
publisher.publish('should not receive')
self.assertFalse(self.check1.wait(0.5))
self.pubsub_management.activate_subscription(subscription_id)
publisher.publish('should receive')
self.assertTrue(self.check1.wait(2))
subscriber.stop()
self.pubsub_management.deactivate_subscription(subscription_id)
self.pubsub_management.delete_subscription(subscription_id)
self.pubsub_management.delete_stream(stream_id)
def test_activation_and_deactivation(self):
stream_id, route = self.pubsub_management.create_stream('stream1','xp1')
subscription_id = self.pubsub_management.create_subscription('sub1', stream_ids=[stream_id])
self.check1 = Event()
def verifier(m,r,s):
self.check1.set()
subscriber = StandaloneStreamSubscriber('sub1',verifier)
subscriber.start()
publisher = StandaloneStreamPublisher(stream_id, route)
publisher.publish('should not receive')
self.assertFalse(self.check1.wait(0.25))
self.pubsub_management.activate_subscription(subscription_id)
publisher.publish('should receive')
self.assertTrue(self.check1.wait(2))
self.check1.clear()
self.assertFalse(self.check1.is_set())
self.pubsub_management.deactivate_subscription(subscription_id)
publisher.publish('should not receive')
self.assertFalse(self.check1.wait(0.5))
self.pubsub_management.activate_subscription(subscription_id)
publisher.publish('should receive')
self.assertTrue(self.check1.wait(2))
subscriber.stop()
self.pubsub_management.deactivate_subscription(subscription_id)
self.pubsub_management.delete_subscription(subscription_id)
self.pubsub_management.delete_stream(stream_id)
def test_demux(self):
self.stream0, self.route0 = self.pubsub_client.create_stream('stream0', exchange_point='test')
self.stream1, self.route1 = self.pubsub_client.create_stream('stream1', exchange_point='main_data')
self.stream2, self.route2 = self.pubsub_client.create_stream('stream2', exchange_point='alt_data')
self.r_stream1 = gevent.event.Event()
self.r_stream2 = gevent.event.Event()
def process(msg, stream_route, stream_id):
if stream_id == self.stream1:
self.r_stream1.set()
elif stream_id == self.stream2:
self.r_stream2.set()
self.container.spawn_process('demuxer', 'ion.processes.data.transforms.mux', 'DemuxTransform', {'process':{'out_streams':[self.stream1, self.stream2]}}, 'demuxer_pid')
self.queue_cleanup.append('demuxer_pid')
sub1 = StandaloneStreamSubscriber('sub1', process)
sub2 = StandaloneStreamSubscriber('sub2', process)
sub1.xn.bind(self.route1.routing_key, self.container.ex_manager.create_xp('main_data'))
sub2.xn.bind(self.route2.routing_key, self.container.ex_manager.create_xp('alt_data'))
sub1.start()
sub2.start()
self.queue_cleanup.append(sub1.xn)
self.queue_cleanup.append(sub2.xn)
xn = self.container.ex_manager.create_xn_queue('demuxer_pid')
xn.bind(self.route0.routing_key, self.container.ex_manager.create_xp(self.route0.exchange_point))
domino = StandaloneStreamPublisher(self.stream0, self.route0)
domino.publish('test')
self.assertTrue(self.r_stream1.wait(2))
self.assertTrue(self.r_stream2.wait(2))
self.container.proc_manager.terminate_process('demuxer_pid')
sub1.stop()
sub2.stop()
def test_replay_pause(self):
# Get a precompiled parameter dictionary with basic ctd fields
pdict_id = self.dataset_management.read_parameter_dictionary_by_name('ctd_parsed_param_dict',id_only=True)
context_ids = self.dataset_management.read_parameter_contexts(pdict_id, id_only=True)
# Add a field that supports binary data input.
bin_context = ParameterContext('binary', param_type=ArrayType())
context_ids.append(self.dataset_management.create_parameter_context('binary', bin_context.dump()))
# Add another field that supports dictionary elements.
rec_context = ParameterContext('records', param_type=RecordType())
context_ids.append(self.dataset_management.create_parameter_context('records', rec_context.dump()))
pdict_id = self.dataset_management.create_parameter_dictionary('replay_pdict', parameter_context_ids=context_ids, temporal_context='time')
stream_def_id = self.pubsub_management.create_stream_definition('replay_stream', parameter_dictionary_id=pdict_id)
replay_stream, replay_route = self.pubsub_management.create_stream('replay', 'xp1', stream_definition_id=stream_def_id)
dataset_id = self.create_dataset(pdict_id)
scov = DatasetManagementService._get_simplex_coverage(dataset_id)
bb = CoverageCraft(scov)
bb.rdt['time'] = np.arange(100)
bb.rdt['temp'] = np.random.random(100) + 30
bb.sync_with_granule()
DatasetManagementService._persist_coverage(dataset_id, bb.coverage) # This invalidates it for multi-host configurations
# Set up the subscriber to verify the data
subscriber = StandaloneStreamSubscriber(self.exchange_space_name, self.validate_granule_subscription)
xp = self.container.ex_manager.create_xp('xp1')
self.queue_buffer.append(self.exchange_space_name)
subscriber.start()
subscriber.xn.bind(replay_route.routing_key, xp)
# Set up the replay agent and the client wrapper
# 1) Define the Replay (dataset and stream to publish on)
self.replay_id, process_id = self.data_retriever.define_replay(dataset_id=dataset_id, stream_id=replay_stream)
# 2) Make a client to the interact with the process (optionall provide it a process to bind with)
replay_client = ReplayClient(process_id)
# 3) Start the agent (launch the process)
self.data_retriever.start_replay_agent(self.replay_id)
# 4) Start replaying...
replay_client.start_replay()
# Wait till we get some granules
self.assertTrue(self.event.wait(5))
# We got granules, pause the replay, clear the queue and allow the process to finish consuming
replay_client.pause_replay()
gevent.sleep(1)
subscriber.xn.purge()
self.event.clear()
# Make sure there's no remaining messages being consumed
self.assertFalse(self.event.wait(1))
# Resume the replay and wait until we start getting granules again
replay_client.resume_replay()
self.assertTrue(self.event.wait(5))
# Stop the replay, clear the queues
replay_client.stop_replay()
gevent.sleep(1)
subscriber.xn.purge()
self.event.clear()
# Make sure that it did indeed stop
self.assertFalse(self.event.wait(1))
subscriber.stop()
def test_dm_end_2_end(self):
#--------------------------------------------------------------------------------
# Set up a stream and have a mock instrument (producer) send data
#--------------------------------------------------------------------------------
self.event.clear()
# Get a precompiled parameter dictionary with basic ctd fields
pdict_id = self.dataset_management.read_parameter_dictionary_by_name('ctd_parsed_param_dict',id_only=True)
context_ids = self.dataset_management.read_parameter_contexts(pdict_id, id_only=True)
# Add a field that supports binary data input.
bin_context = ParameterContext('binary', param_type=ArrayType())
context_ids.append(self.dataset_management.create_parameter_context('binary', bin_context.dump()))
# Add another field that supports dictionary elements.
rec_context = ParameterContext('records', param_type=RecordType())
context_ids.append(self.dataset_management.create_parameter_context('records', rec_context.dump()))
pdict_id = self.dataset_management.create_parameter_dictionary('replay_pdict', parameter_context_ids=context_ids, temporal_context='time')
stream_definition = self.pubsub_management.create_stream_definition('ctd data', parameter_dictionary_id=pdict_id)
stream_id, route = self.pubsub_management.create_stream('producer', exchange_point=self.exchange_point_name, stream_definition_id=stream_definition)
#--------------------------------------------------------------------------------
# Start persisting the data on the stream
# - Get the ingestion configuration from the resource registry
# - Create the dataset
# - call persist_data_stream to setup the subscription for the ingestion workers
# on the stream that you specify which causes the data to be persisted
#--------------------------------------------------------------------------------
ingest_config_id = self.get_ingestion_config()
dataset_id = self.create_dataset(pdict_id)
self.ingestion_management.persist_data_stream(stream_id=stream_id, ingestion_configuration_id=ingest_config_id, dataset_id=dataset_id)
#--------------------------------------------------------------------------------
# Now the granules are ingesting and persisted
#--------------------------------------------------------------------------------
self.launch_producer(stream_id)
self.wait_until_we_have_enough_granules(dataset_id,40)
#--------------------------------------------------------------------------------
# Now get the data in one chunk using an RPC Call to start_retreive
#--------------------------------------------------------------------------------
replay_data = self.data_retriever.retrieve(dataset_id)
self.assertIsInstance(replay_data, Granule)
rdt = RecordDictionaryTool.load_from_granule(replay_data)
self.assertTrue((rdt['time'][:10] == np.arange(10)).all(),'%s' % rdt['time'][:])
self.assertTrue((rdt['binary'][:10] == np.array(['hi']*10, dtype='object')).all())
#--------------------------------------------------------------------------------
# Now to try the streamed approach
#--------------------------------------------------------------------------------
replay_stream_id, replay_route = self.pubsub_management.create_stream('replay_out', exchange_point=self.exchange_point_name, stream_definition_id=stream_definition)
self.replay_id, process_id = self.data_retriever.define_replay(dataset_id=dataset_id, stream_id=replay_stream_id)
log.info('Process ID: %s', process_id)
replay_client = ReplayClient(process_id)
#--------------------------------------------------------------------------------
# Create the listening endpoint for the the retriever to talk to
#--------------------------------------------------------------------------------
xp = self.container.ex_manager.create_xp(self.exchange_point_name)
subscriber = StandaloneStreamSubscriber(self.exchange_space_name, self.validate_granule_subscription)
self.queue_buffer.append(self.exchange_space_name)
subscriber.start()
subscriber.xn.bind(replay_route.routing_key, xp)
self.data_retriever.start_replay_agent(self.replay_id)
self.assertTrue(replay_client.await_agent_ready(5), 'The process never launched')
replay_client.start_replay()
self.assertTrue(self.event.wait(10))
subscriber.stop()
self.data_retriever.cancel_replay_agent(self.replay_id)
#--------------------------------------------------------------------------------
# Test the slicing capabilities
#--------------------------------------------------------------------------------
granule = self.data_retriever.retrieve(dataset_id=dataset_id, query={'tdoa':slice(0,5)})
rdt = RecordDictionaryTool.load_from_granule(granule)
b = rdt['time'] == np.arange(5)
self.assertTrue(b.all() if not isinstance(b,bool) else b)
self.streams.append(stream_id)
self.stop_ingestion(stream_id)
def test_topic_craziness(self):
self.msg_queue = Queue()
def subscriber1(m, r, s):
self.msg_queue.put(m)
sub1 = StandaloneStreamSubscriber('sub1', subscriber1)
self.queue_cleanup.append(sub1.xn.queue)
sub1.start()
topic1 = self.pubsub_management.create_topic('topic1',
exchange_point='xp1')
topic2 = self.pubsub_management.create_topic('topic2',
exchange_point='xp1',
parent_topic_id=topic1)
topic3 = self.pubsub_management.create_topic('topic3',
exchange_point='xp1',
parent_topic_id=topic1)
topic4 = self.pubsub_management.create_topic('topic4',
exchange_point='xp1',
parent_topic_id=topic2)
topic5 = self.pubsub_management.create_topic('topic5',
exchange_point='xp1',
parent_topic_id=topic2)
topic6 = self.pubsub_management.create_topic('topic6',
exchange_point='xp1',
parent_topic_id=topic3)
topic7 = self.pubsub_management.create_topic('topic7',
exchange_point='xp1',
parent_topic_id=topic3)
# Tree 2
topic8 = self.pubsub_management.create_topic('topic8',
exchange_point='xp2')
topic9 = self.pubsub_management.create_topic('topic9',
exchange_point='xp2',
parent_topic_id=topic8)
topic10 = self.pubsub_management.create_topic('topic10',
exchange_point='xp2',
parent_topic_id=topic9)
topic11 = self.pubsub_management.create_topic('topic11',
exchange_point='xp2',
parent_topic_id=topic9)
topic12 = self.pubsub_management.create_topic('topic12',
exchange_point='xp2',
parent_topic_id=topic11)
topic13 = self.pubsub_management.create_topic('topic13',
exchange_point='xp2',
parent_topic_id=topic11)
self.exchange_cleanup.extend(['xp1', 'xp2'])
stream1_id, route = self.pubsub_management.create_stream(
'stream1',
topic_ids=[topic7, topic4, topic5],
exchange_point='xp1')
stream2_id, route = self.pubsub_management.create_stream(
'stream2', topic_ids=[topic8], exchange_point='xp2')
stream3_id, route = self.pubsub_management.create_stream(
'stream3', topic_ids=[topic10, topic13], exchange_point='xp2')
stream4_id, route = self.pubsub_management.create_stream(
'stream4', topic_ids=[topic9], exchange_point='xp2')
stream5_id, route = self.pubsub_management.create_stream(
'stream5', topic_ids=[topic11], exchange_point='xp2')
subscription1 = self.pubsub_management.create_subscription(
'sub1', topic_ids=[topic1])
subscription2 = self.pubsub_management.create_subscription(
'sub2', topic_ids=[topic8], exchange_name='sub1')
subscription3 = self.pubsub_management.create_subscription(
'sub3', topic_ids=[topic9], exchange_name='sub1')
subscription4 = self.pubsub_management.create_subscription(
'sub4',
topic_ids=[topic10, topic13, topic11],
exchange_name='sub1')
#--------------------------------------------------------------------------------
self.pubsub_management.activate_subscription(subscription1)
self.publish_on_stream(stream1_id, 1)
self.assertEquals(self.msg_queue.get(timeout=10), 1)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.1)
self.pubsub_management.deactivate_subscription(subscription1)
self.pubsub_management.delete_subscription(subscription1)
#--------------------------------------------------------------------------------
self.pubsub_management.activate_subscription(subscription2)
self.publish_on_stream(stream2_id, 2)
self.assertEquals(self.msg_queue.get(timeout=10), 2)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.1)
self.pubsub_management.deactivate_subscription(subscription2)
self.pubsub_management.delete_subscription(subscription2)
#--------------------------------------------------------------------------------
self.pubsub_management.activate_subscription(subscription3)
self.publish_on_stream(stream2_id, 3)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.3)
self.publish_on_stream(stream3_id, 4)
self.assertEquals(self.msg_queue.get(timeout=10), 4)
self.pubsub_management.deactivate_subscription(subscription3)
self.pubsub_management.delete_subscription(subscription3)
#--------------------------------------------------------------------------------
self.pubsub_management.activate_subscription(subscription4)
self.publish_on_stream(stream4_id, 5)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.3)
self.publish_on_stream(stream5_id, 6)
self.assertEquals(self.msg_queue.get(timeout=10), 6)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.3)
self.pubsub_management.deactivate_subscription(subscription4)
self.pubsub_management.delete_subscription(subscription4)
#--------------------------------------------------------------------------------
sub1.stop()
self.pubsub_management.delete_topic(topic13)
self.pubsub_management.delete_topic(topic12)
self.pubsub_management.delete_topic(topic11)
self.pubsub_management.delete_topic(topic10)
self.pubsub_management.delete_topic(topic9)
self.pubsub_management.delete_topic(topic8)
self.pubsub_management.delete_topic(topic7)
self.pubsub_management.delete_topic(topic6)
self.pubsub_management.delete_topic(topic5)
self.pubsub_management.delete_topic(topic4)
self.pubsub_management.delete_topic(topic3)
self.pubsub_management.delete_topic(topic2)
self.pubsub_management.delete_topic(topic1)
self.pubsub_management.delete_stream(stream1_id)
self.pubsub_management.delete_stream(stream2_id)
self.pubsub_management.delete_stream(stream3_id)
self.pubsub_management.delete_stream(stream4_id)
self.pubsub_management.delete_stream(stream5_id)
def test_full_pubsub(self):
self.sub1_sat = Event()
self.sub2_sat = Event()
def subscriber1(m, r, s):
self.sub1_sat.set()
def subscriber2(m, r, s):
self.sub2_sat.set()
sub1 = StandaloneStreamSubscriber('sub1', subscriber1)
self.queue_cleanup.append(sub1.xn.queue)
sub1.start()
sub2 = StandaloneStreamSubscriber('sub2', subscriber2)
self.queue_cleanup.append(sub2.xn.queue)
sub2.start()
log_topic = self.pubsub_management.create_topic(
'instrument_logs', exchange_point='instruments')
science_topic = self.pubsub_management.create_topic(
'science_data', exchange_point='instruments')
events_topic = self.pubsub_management.create_topic(
'notifications', exchange_point='events')
log_stream, route = self.pubsub_management.create_stream(
'instrument1-logs',
topic_ids=[log_topic],
exchange_point='instruments')
ctd_stream, route = self.pubsub_management.create_stream(
'instrument1-ctd',
topic_ids=[science_topic],
exchange_point='instruments')
event_stream, route = self.pubsub_management.create_stream(
'notifications', topic_ids=[events_topic], exchange_point='events')
raw_stream, route = self.pubsub_management.create_stream(
'temp', exchange_point='global.data')
self.exchange_cleanup.extend(['instruments', 'events', 'global.data'])
subscription1 = self.pubsub_management.create_subscription(
'subscription1',
stream_ids=[log_stream, event_stream],
exchange_name='sub1')
subscription2 = self.pubsub_management.create_subscription(
'subscription2',
exchange_points=['global.data'],
stream_ids=[ctd_stream],
exchange_name='sub2')
self.pubsub_management.activate_subscription(subscription1)
self.pubsub_management.activate_subscription(subscription2)
self.publish_on_stream(log_stream, 1)
self.assertTrue(self.sub1_sat.wait(4))
self.assertFalse(self.sub2_sat.is_set())
self.publish_on_stream(raw_stream, 1)
self.assertTrue(self.sub1_sat.wait(4))
sub1.stop()
sub2.stop()
def test_topic_craziness(self):
self.msg_queue = Queue()
def subscriber1(m, r, s):
self.msg_queue.put(m)
sub1 = StandaloneStreamSubscriber("sub1", subscriber1)
self.queue_cleanup.append(sub1.xn.queue)
sub1.start()
topic1 = self.pubsub_management.create_topic("topic1", exchange_point="xp1")
topic2 = self.pubsub_management.create_topic("topic2", exchange_point="xp1", parent_topic_id=topic1)
topic3 = self.pubsub_management.create_topic("topic3", exchange_point="xp1", parent_topic_id=topic1)
topic4 = self.pubsub_management.create_topic("topic4", exchange_point="xp1", parent_topic_id=topic2)
topic5 = self.pubsub_management.create_topic("topic5", exchange_point="xp1", parent_topic_id=topic2)
topic6 = self.pubsub_management.create_topic("topic6", exchange_point="xp1", parent_topic_id=topic3)
topic7 = self.pubsub_management.create_topic("topic7", exchange_point="xp1", parent_topic_id=topic3)
# Tree 2
topic8 = self.pubsub_management.create_topic("topic8", exchange_point="xp2")
topic9 = self.pubsub_management.create_topic("topic9", exchange_point="xp2", parent_topic_id=topic8)
topic10 = self.pubsub_management.create_topic("topic10", exchange_point="xp2", parent_topic_id=topic9)
topic11 = self.pubsub_management.create_topic("topic11", exchange_point="xp2", parent_topic_id=topic9)
topic12 = self.pubsub_management.create_topic("topic12", exchange_point="xp2", parent_topic_id=topic11)
topic13 = self.pubsub_management.create_topic("topic13", exchange_point="xp2", parent_topic_id=topic11)
self.exchange_cleanup.extend(["xp1", "xp2"])
stream1_id, route = self.pubsub_management.create_stream(
"stream1", topic_ids=[topic7, topic4, topic5], exchange_point="xp1"
)
stream2_id, route = self.pubsub_management.create_stream("stream2", topic_ids=[topic8], exchange_point="xp2")
stream3_id, route = self.pubsub_management.create_stream(
"stream3", topic_ids=[topic10, topic13], exchange_point="xp2"
)
stream4_id, route = self.pubsub_management.create_stream("stream4", topic_ids=[topic9], exchange_point="xp2")
stream5_id, route = self.pubsub_management.create_stream("stream5", topic_ids=[topic11], exchange_point="xp2")
subscription1 = self.pubsub_management.create_subscription("sub1", topic_ids=[topic1])
subscription2 = self.pubsub_management.create_subscription("sub2", topic_ids=[topic8], exchange_name="sub1")
subscription3 = self.pubsub_management.create_subscription("sub3", topic_ids=[topic9], exchange_name="sub1")
subscription4 = self.pubsub_management.create_subscription(
"sub4", topic_ids=[topic10, topic13, topic11], exchange_name="sub1"
)
# --------------------------------------------------------------------------------
self.pubsub_management.activate_subscription(subscription1)
self.publish_on_stream(stream1_id, 1)
self.assertEquals(self.msg_queue.get(timeout=10), 1)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.1)
self.pubsub_management.deactivate_subscription(subscription1)
self.pubsub_management.delete_subscription(subscription1)
# --------------------------------------------------------------------------------
self.pubsub_management.activate_subscription(subscription2)
self.publish_on_stream(stream2_id, 2)
self.assertEquals(self.msg_queue.get(timeout=10), 2)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.1)
self.pubsub_management.deactivate_subscription(subscription2)
self.pubsub_management.delete_subscription(subscription2)
# --------------------------------------------------------------------------------
self.pubsub_management.activate_subscription(subscription3)
self.publish_on_stream(stream2_id, 3)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.3)
self.publish_on_stream(stream3_id, 4)
self.assertEquals(self.msg_queue.get(timeout=10), 4)
self.pubsub_management.deactivate_subscription(subscription3)
self.pubsub_management.delete_subscription(subscription3)
# --------------------------------------------------------------------------------
self.pubsub_management.activate_subscription(subscription4)
self.publish_on_stream(stream4_id, 5)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.3)
self.publish_on_stream(stream5_id, 6)
self.assertEquals(self.msg_queue.get(timeout=10), 6)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.3)
self.pubsub_management.deactivate_subscription(subscription4)
self.pubsub_management.delete_subscription(subscription4)
# --------------------------------------------------------------------------------
sub1.stop()
self.pubsub_management.delete_topic(topic13)
self.pubsub_management.delete_topic(topic12)
self.pubsub_management.delete_topic(topic11)
self.pubsub_management.delete_topic(topic10)
self.pubsub_management.delete_topic(topic9)
self.pubsub_management.delete_topic(topic8)
self.pubsub_management.delete_topic(topic7)
self.pubsub_management.delete_topic(topic6)
self.pubsub_management.delete_topic(topic5)
self.pubsub_management.delete_topic(topic4)
self.pubsub_management.delete_topic(topic3)
self.pubsub_management.delete_topic(topic2)
self.pubsub_management.delete_topic(topic1)
self.pubsub_management.delete_stream(stream1_id)
self.pubsub_management.delete_stream(stream2_id)
self.pubsub_management.delete_stream(stream3_id)
self.pubsub_management.delete_stream(stream4_id)
self.pubsub_management.delete_stream(stream5_id)
def test_replay_pause(self):
# Get a precompiled parameter dictionary with basic ctd fields
pdict_id = self.dataset_management.read_parameter_dictionary_by_name(
'ctd_parsed_param_dict', id_only=True)
context_ids = self.dataset_management.read_parameter_contexts(
pdict_id, id_only=True)
# Add a field that supports binary data input.
bin_context = ParameterContext('binary', param_type=ArrayType())
context_ids.append(
self.dataset_management.create_parameter_context(
'binary', bin_context.dump()))
# Add another field that supports dictionary elements.
rec_context = ParameterContext('records', param_type=RecordType())
context_ids.append(
self.dataset_management.create_parameter_context(
'records', rec_context.dump()))
pdict_id = self.dataset_management.create_parameter_dictionary(
'replay_pdict',
parameter_context_ids=context_ids,
temporal_context='time')
stream_def_id = self.pubsub_management.create_stream_definition(
'replay_stream', parameter_dictionary_id=pdict_id)
replay_stream, replay_route = self.pubsub_management.create_stream(
'replay', 'xp1', stream_definition_id=stream_def_id)
dataset_id = self.create_dataset(pdict_id)
scov = DatasetManagementService._get_coverage(dataset_id)
bb = CoverageCraft(scov)
bb.rdt['time'] = np.arange(100)
bb.rdt['temp'] = np.random.random(100) + 30
bb.sync_with_granule()
DatasetManagementService._persist_coverage(
dataset_id,
bb.coverage) # This invalidates it for multi-host configurations
# Set up the subscriber to verify the data
subscriber = StandaloneStreamSubscriber(
self.exchange_space_name, self.validate_granule_subscription)
xp = self.container.ex_manager.create_xp('xp1')
self.queue_buffer.append(self.exchange_space_name)
subscriber.start()
subscriber.xn.bind(replay_route.routing_key, xp)
# Set up the replay agent and the client wrapper
# 1) Define the Replay (dataset and stream to publish on)
self.replay_id, process_id = self.data_retriever.define_replay(
dataset_id=dataset_id, stream_id=replay_stream)
# 2) Make a client to the interact with the process (optionall provide it a process to bind with)
replay_client = ReplayClient(process_id)
# 3) Start the agent (launch the process)
self.data_retriever.start_replay_agent(self.replay_id)
# 4) Start replaying...
replay_client.start_replay()
# Wait till we get some granules
self.assertTrue(self.event.wait(5))
# We got granules, pause the replay, clear the queue and allow the process to finish consuming
replay_client.pause_replay()
gevent.sleep(1)
subscriber.xn.purge()
self.event.clear()
# Make sure there's no remaining messages being consumed
self.assertFalse(self.event.wait(1))
# Resume the replay and wait until we start getting granules again
replay_client.resume_replay()
self.assertTrue(self.event.wait(5))
# Stop the replay, clear the queues
replay_client.stop_replay()
gevent.sleep(1)
subscriber.xn.purge()
self.event.clear()
# Make sure that it did indeed stop
self.assertFalse(self.event.wait(1))
subscriber.stop()
def test_dm_end_2_end(self):
#--------------------------------------------------------------------------------
# Set up a stream and have a mock instrument (producer) send data
#--------------------------------------------------------------------------------
self.event.clear()
# Get a precompiled parameter dictionary with basic ctd fields
pdict_id = self.dataset_management.read_parameter_dictionary_by_name(
'ctd_parsed_param_dict', id_only=True)
context_ids = self.dataset_management.read_parameter_contexts(
pdict_id, id_only=True)
# Add a field that supports binary data input.
bin_context = ParameterContext('binary', param_type=ArrayType())
context_ids.append(
self.dataset_management.create_parameter_context(
'binary', bin_context.dump()))
# Add another field that supports dictionary elements.
rec_context = ParameterContext('records', param_type=RecordType())
context_ids.append(
self.dataset_management.create_parameter_context(
'records', rec_context.dump()))
pdict_id = self.dataset_management.create_parameter_dictionary(
'replay_pdict',
parameter_context_ids=context_ids,
temporal_context='time')
stream_definition = self.pubsub_management.create_stream_definition(
'ctd data', parameter_dictionary_id=pdict_id)
stream_id, route = self.pubsub_management.create_stream(
'producer',
exchange_point=self.exchange_point_name,
stream_definition_id=stream_definition)
#--------------------------------------------------------------------------------
# Start persisting the data on the stream
# - Get the ingestion configuration from the resource registry
# - Create the dataset
# - call persist_data_stream to setup the subscription for the ingestion workers
# on the stream that you specify which causes the data to be persisted
#--------------------------------------------------------------------------------
ingest_config_id = self.get_ingestion_config()
dataset_id = self.create_dataset(pdict_id)
self.ingestion_management.persist_data_stream(
stream_id=stream_id,
ingestion_configuration_id=ingest_config_id,
dataset_id=dataset_id)
#--------------------------------------------------------------------------------
# Now the granules are ingesting and persisted
#--------------------------------------------------------------------------------
self.launch_producer(stream_id)
self.wait_until_we_have_enough_granules(dataset_id, 40)
#--------------------------------------------------------------------------------
# Now get the data in one chunk using an RPC Call to start_retreive
#--------------------------------------------------------------------------------
replay_data = self.data_retriever.retrieve(dataset_id)
self.assertIsInstance(replay_data, Granule)
rdt = RecordDictionaryTool.load_from_granule(replay_data)
self.assertTrue((rdt['time'][:10] == np.arange(10)).all(),
'%s' % rdt['time'][:])
self.assertTrue((rdt['binary'][:10] == np.array(['hi'] * 10,
dtype='object')).all())
#--------------------------------------------------------------------------------
# Now to try the streamed approach
#--------------------------------------------------------------------------------
replay_stream_id, replay_route = self.pubsub_management.create_stream(
'replay_out',
exchange_point=self.exchange_point_name,
stream_definition_id=stream_definition)
self.replay_id, process_id = self.data_retriever.define_replay(
dataset_id=dataset_id, stream_id=replay_stream_id)
log.info('Process ID: %s', process_id)
replay_client = ReplayClient(process_id)
#--------------------------------------------------------------------------------
# Create the listening endpoint for the the retriever to talk to
#--------------------------------------------------------------------------------
xp = self.container.ex_manager.create_xp(self.exchange_point_name)
subscriber = StandaloneStreamSubscriber(
self.exchange_space_name, self.validate_granule_subscription)
self.queue_buffer.append(self.exchange_space_name)
subscriber.start()
subscriber.xn.bind(replay_route.routing_key, xp)
self.data_retriever.start_replay_agent(self.replay_id)
self.assertTrue(replay_client.await_agent_ready(5),
'The process never launched')
replay_client.start_replay()
self.assertTrue(self.event.wait(10))
subscriber.stop()
self.data_retriever.cancel_replay_agent(self.replay_id)
#--------------------------------------------------------------------------------
# Test the slicing capabilities
#--------------------------------------------------------------------------------
granule = self.data_retriever.retrieve(dataset_id=dataset_id,
query={'tdoa': slice(0, 5)})
rdt = RecordDictionaryTool.load_from_granule(granule)
b = rdt['time'] == np.arange(5)
self.assertTrue(b.all() if not isinstance(b, bool) else b)
self.streams.append(stream_id)
self.stop_ingestion(stream_id)
def test_topic_craziness(self):
self.msg_queue = Queue()
def subscriber1(m,r,s):
self.msg_queue.put(m)
sub1 = StandaloneStreamSubscriber('sub1', subscriber1)
self.queue_cleanup.append(sub1.xn.queue)
sub1.start()
topic1 = self.pubsub_management.create_topic('topic1', exchange_point='xp1')
topic2 = self.pubsub_management.create_topic('topic2', exchange_point='xp1', parent_topic_id=topic1)
topic3 = self.pubsub_management.create_topic('topic3', exchange_point='xp1', parent_topic_id=topic1)
topic4 = self.pubsub_management.create_topic('topic4', exchange_point='xp1', parent_topic_id=topic2)
topic5 = self.pubsub_management.create_topic('topic5', exchange_point='xp1', parent_topic_id=topic2)
topic6 = self.pubsub_management.create_topic('topic6', exchange_point='xp1', parent_topic_id=topic3)
topic7 = self.pubsub_management.create_topic('topic7', exchange_point='xp1', parent_topic_id=topic3)
# Tree 2
topic8 = self.pubsub_management.create_topic('topic8', exchange_point='xp2')
topic9 = self.pubsub_management.create_topic('topic9', exchange_point='xp2', parent_topic_id=topic8)
topic10 = self.pubsub_management.create_topic('topic10', exchange_point='xp2', parent_topic_id=topic9)
topic11 = self.pubsub_management.create_topic('topic11', exchange_point='xp2', parent_topic_id=topic9)
topic12 = self.pubsub_management.create_topic('topic12', exchange_point='xp2', parent_topic_id=topic11)
topic13 = self.pubsub_management.create_topic('topic13', exchange_point='xp2', parent_topic_id=topic11)
self.exchange_cleanup.extend(['xp1','xp2'])
stream1_id, route = self.pubsub_management.create_stream('stream1', topic_ids=[topic7, topic4, topic5], exchange_point='xp1')
stream2_id, route = self.pubsub_management.create_stream('stream2', topic_ids=[topic8], exchange_point='xp2')
stream3_id, route = self.pubsub_management.create_stream('stream3', topic_ids=[topic10,topic13], exchange_point='xp2')
stream4_id, route = self.pubsub_management.create_stream('stream4', topic_ids=[topic9], exchange_point='xp2')
stream5_id, route = self.pubsub_management.create_stream('stream5', topic_ids=[topic11], exchange_point='xp2')
subscription1 = self.pubsub_management.create_subscription('sub1', topic_ids=[topic1])
subscription2 = self.pubsub_management.create_subscription('sub2', topic_ids=[topic8], exchange_name='sub1')
subscription3 = self.pubsub_management.create_subscription('sub3', topic_ids=[topic9], exchange_name='sub1')
subscription4 = self.pubsub_management.create_subscription('sub4', topic_ids=[topic10,topic13, topic11], exchange_name='sub1')
#--------------------------------------------------------------------------------
self.pubsub_management.activate_subscription(subscription1)
self.publish_on_stream(stream1_id,1)
self.assertEquals(self.msg_queue.get(timeout=10), 1)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.1)
self.pubsub_management.deactivate_subscription(subscription1)
self.pubsub_management.delete_subscription(subscription1)
#--------------------------------------------------------------------------------
self.pubsub_management.activate_subscription(subscription2)
self.publish_on_stream(stream2_id,2)
self.assertEquals(self.msg_queue.get(timeout=10), 2)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.1)
self.pubsub_management.deactivate_subscription(subscription2)
self.pubsub_management.delete_subscription(subscription2)
#--------------------------------------------------------------------------------
self.pubsub_management.activate_subscription(subscription3)
self.publish_on_stream(stream2_id, 3)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.3)
self.publish_on_stream(stream3_id, 4)
self.assertEquals(self.msg_queue.get(timeout=10),4)
self.pubsub_management.deactivate_subscription(subscription3)
self.pubsub_management.delete_subscription(subscription3)
#--------------------------------------------------------------------------------
self.pubsub_management.activate_subscription(subscription4)
self.publish_on_stream(stream4_id, 5)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.3)
self.publish_on_stream(stream5_id, 6)
self.assertEquals(self.msg_queue.get(timeout=10),6)
with self.assertRaises(Empty):
self.msg_queue.get(timeout=0.3)
self.pubsub_management.deactivate_subscription(subscription4)
self.pubsub_management.delete_subscription(subscription4)
#--------------------------------------------------------------------------------
sub1.stop()
self.pubsub_management.delete_topic(topic13)
self.pubsub_management.delete_topic(topic12)
self.pubsub_management.delete_topic(topic11)
self.pubsub_management.delete_topic(topic10)
self.pubsub_management.delete_topic(topic9)
self.pubsub_management.delete_topic(topic8)
self.pubsub_management.delete_topic(topic7)
self.pubsub_management.delete_topic(topic6)
self.pubsub_management.delete_topic(topic5)
self.pubsub_management.delete_topic(topic4)
self.pubsub_management.delete_topic(topic3)
self.pubsub_management.delete_topic(topic2)
self.pubsub_management.delete_topic(topic1)
self.pubsub_management.delete_stream(stream1_id)
self.pubsub_management.delete_stream(stream2_id)
self.pubsub_management.delete_stream(stream3_id)
self.pubsub_management.delete_stream(stream4_id)
self.pubsub_management.delete_stream(stream5_id)