def get_last_values(cls, dataset_id):
coverage = DatasetManagementService._get_coverage(dataset_id)
black_box = CoverageCraft(coverage)
black_box.sync_rdt_with_coverage(tdoa=slice(-1,None))
granule = black_box.to_granule()
return granule
def get_last_granule(cls, container, dataset_id):
dsm_cli = DatasetManagementServiceClient()
dataset = dsm_cli.read_dataset(dataset_id)
cc = container
datastore_name = dataset.datastore_name
view_name = dataset.view_name
datastore = cc.datastore_manager.get_datastore(datastore_name, DataStore.DS_PROFILE.SCIDATA)
opts = dict(
start_key = [dataset_id, {}],
end_key = [dataset_id, 0],
descending = True,
limit = 1,
include_docs = True
)
results = datastore.query_view(view_name,opts=opts)
if not results:
raise NotFound('A granule could not be located.')
if results[0] is None:
raise NotFound('A granule could not be located.')
doc = results[0].get('doc')
if doc is None:
return None
ts = float(doc.get('ts_create',0))
coverage = DatasetManagementService._get_coverage(dataset_id)
black_box = CoverageCraft(coverage)
black_box.sync_rdt_with_coverage(start_time=ts,end_time=None)
granule = black_box.to_granule()
return granule
def publish_hifi(self, stream_id, offset=0):
pub = SimpleStreamPublisher.new_publisher(self.container, self.exchange_point_name, stream_id)
black_box = CoverageCraft()
black_box.rdt["time"] = np.arange(10) + (offset * 10)
black_box.rdt["temp"] = (np.arange(10) + (offset * 10)) * 2
granule = black_box.to_granule()
pub.publish(granule)
def test_to_coverage(self):
granule = self.sample_granule()
crafter = CoverageCraft()
crafter.sync_with_granule(granule)
coverage = crafter.coverage
time_vals = coverage.get_time_values()
comp = time_vals == np.arange(20)
self.assertTrue(comp.all())
def execute_retrieve(self):
'''
execute_retrieve Executes a retrieval and returns the result
as a value in lieu of publishing it on a stream
'''
coverage = DatasetManagementService._get_coverage(self.dataset_id)
crafter = CoverageCraft(coverage)
#@todo: add bounds checking to ensure the dataset being retrieved is not too large
crafter.sync_rdt_with_coverage(start_time=self.start_time,end_time=self.end_time,parameters=self.parameters)
granule = crafter.to_granule()
return granule
def sample_granule(self):
pdict = CoverageCraft.create_parameters()
rdt = RecordDictionaryTool(param_dictionary=pdict)
rdt['time'] = np.arange(20)
rdt['temp'] = np.array([5] * 20)
rdt['conductivity'] = np.array([10] * 20)
rdt['lat'] = np.array([0] * 20)
rdt['lon'] = np.array([0] * 20)
rdt['depth'] = np.array([0] * 20)
rdt['data'] = np.array([0x01] * 20)
return build_granule('sample', param_dictionary=pdict, record_dictionary=rdt)
def add_granule(self,stream_id, granule):
'''
Appends the granule's data to the coverage and persists it.
'''
#--------------------------------------------------------------------------------
# Coverage determiniation and appending
#--------------------------------------------------------------------------------
dataset_id = self.get_dataset(stream_id)
if not dataset_id:
log.error('No dataset could be determined on this stream: %s', stream_id)
return
coverage = self.get_coverage(stream_id)
if not coverage:
log.error('Could not persist coverage from granule, coverage is None')
return
#--------------------------------------------------------------------------------
# Actual persistence
#--------------------------------------------------------------------------------
covcraft = CoverageCraft(coverage)
covcraft.sync_with_granule(granule)
DatasetManagementService._persist_coverage(dataset_id,coverage)
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 _trigger_func(self, stream_id):
t_i = 0
while not self.finished.is_set():
length = 10
black_box = CoverageCraft()
black_box.rdt['time'] = numpy.arange(10) + t_i*10
black_box.rdt['temp'] = numpy.random.random(10) * 10
black_box.rdt['lat'] = numpy.array([0] * 10)
black_box.rdt['lon'] = numpy.array([0] * 10)
black_box.rdt['depth'] = numpy.array([0] * 10)
black_box.rdt['conductivity'] = numpy.random.random(10) * 10
black_box.rdt['data'] = numpy.random.randint(0,255,10) # Simulates random bytes
black_box.sync_with_granule()
granule = black_box.to_granule()
self.publish(granule)
gevent.sleep(self.interval)
t_i += 1
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_coverage_ingest(self):
stream_id = self.pubsub_management.create_stream()
dataset_id = self.create_dataset()
# I freaking hate this bug
self.get_datastore(dataset_id)
ingestion_config_id = self.get_ingestion_config()
self.ingestion_management.persist_data_stream(
stream_id=stream_id, ingestion_configuration_id=ingestion_config_id, dataset_id=dataset_id
)
black_box = CoverageCraft()
black_box.rdt["time"] = np.arange(20)
black_box.rdt["temp"] = np.random.random(20) * 10
black_box.sync_with_granule()
granule = black_box.to_granule()
publisher = SimpleStreamPublisher.new_publisher(self.container, self.exchange_point_name, stream_id)
publisher.publish(granule)
self.wait_until_we_have_enough_granules(dataset_id, 1)
coverage = DatasetManagementService._get_coverage(dataset_id)
black_box = CoverageCraft(coverage)
black_box.sync_rdt_with_coverage()
comp = black_box.rdt["time"] == np.arange(20)
self.assertTrue(comp.all())
black_box = CoverageCraft()
black_box.rdt["time"] = np.arange(20) + 20
black_box.rdt["temp"] = np.random.random(20) * 10
black_box.sync_with_granule()
granule = black_box.to_granule()
publisher.publish(granule)
self.wait_until_we_have_enough_granules(dataset_id, 2)
coverage = DatasetManagementService._get_coverage(dataset_id)
black_box = CoverageCraft(coverage)
black_box.sync_rdt_with_coverage()
comp = black_box.rdt["time"] == np.arange(40)
self.assertTrue(comp.all())
granule = self.data_retriever.retrieve(dataset_id)
black_box = CoverageCraft()
black_box.sync_rdt_with_granule(granule)
comp = black_box.rdt["time"] == np.arange(40)
self.assertTrue(comp.all())
