def setUp(self):
super().setUp()
# Create retrieval manager
self.retrieval_manager = PeriodicRetrievalManager(RETRIEVAL_PERIOD, self.update_mapper, self.logger)
def run(config_location):
# Load config
config = load_config(os.path.join(config_location, "setup.conf"))
# Setup measurement logging
logging_buffer_latency = timedelta(seconds=config.influxdb.buffer_latency)
influxdb_config = InfluxDBConnectionConfig(config.influxdb.host, config.influxdb.port, config.influxdb.username,
config.influxdb.password, config.influxdb.database)
logger = InfluxDBLogger(influxdb_config, buffer_latency=logging_buffer_latency)
# Set HTTP(S) connection pool size (for CouchDB)
# NOTE This is taken from an environment variable, as it's not
# something that would probably need tweaking that much:
pool_size = int(os.environ.get('COOKIEMONSTER_POOL_SIZE', 16))
patch_connection_pools(maxsize=pool_size)
# Setup cookie jar
cookie_jar = RateLimitedBiscuitTin(config.cookie_jar.max_requests_per_second,
config.cookie_jar.url,
config.cookie_jar.database,
config.cookie_jar.buffer_capacity,
config.cookie_jar.buffer_latency,
verify=config.cookie_jar.cacert)
add_cookie_jar_logging(cookie_jar, logger)
add_couchdb_logging(cookie_jar, logger)
# Setup data retrieval manager
update_mapper = BatonUpdateMapper(config.baton.binaries_location, zone=config.baton.zone)
retrieval_manager = PeriodicRetrievalManager(config.retrieval.period, update_mapper, logger)
# # Setup basic Slack client
# slack = BasicSlackClient(config.slack.token, config.slack.default_channel, config.slack.default_username)
slack = None
# Setup rule output log file writer
rule_log_writer = RuleOutputWriter(config.output.log_file)
# # Setup basic message queue (e.g. RabbitMQ) client
# message_queue = BasicMessageQueue(config.message_queue.host, config.message_queue.port,
# config.message_queue.username, config.message_queue.password)
message_queue = None
# Define the context that rules and enrichment loaders has access to
context = HgiContext(cookie_jar, config, rule_log_writer, slack, message_queue)
# Setup rules source
rules_source = RuleSource(config.processing.rules_location, context)
rules_source.start()
# Setup enrichment loader source
enrichment_loader_source = EnrichmentLoaderSource(config.processing.enrichment_loaders_location, context)
enrichment_loader_source.start()
# Setup the data processor manager
processor_manager = BasicProcessorManager(cookie_jar, rules_source, enrichment_loader_source,
config.processing.max_threads, logger)
# Connect components to the cookie jar
_connect_retrieval_manager_to_cookie_jar(retrieval_manager, cookie_jar, config.cookie_jar.max_requests_per_second,
logger)
_connect_retrieval_manager_to_since_file(retrieval_manager, config_location)
_connect_processor_manager_to_cookie_jar(processor_manager, cookie_jar)
# Setup the HTTP API
api = HTTP_API()
api.inject(APIDependency.CookieJar, cookie_jar)
api.inject(APIDependency.System, None)
api.listen(config.api.port)
# Start the retrieval manager from the last known successful
# retrieval time (or invocation time, otherwise)
try:
with open(os.path.join(config_location, "since"), "r") as f:
since_time = datetime.fromtimestamp(int(f.read()))
except:
since_time = datetime.now()
retrieval_manager.start(since_time)
# Start processing of any unprocessed cookies
processor_manager.process_any_cookies()
# Setup monitors
ThreadsMonitor(logger, logging_buffer_latency).start()
CookieJarMonitor(logger, logging_buffer_latency, cookie_jar).start()
class TestPeriodicRetrievalManager(_BaseRetrievalManagerTest):
"""
Test cases for `PeriodicRetrievalManager`.
"""
def setUp(self):
super().setUp()
# Create retrieval manager
self.retrieval_manager = PeriodicRetrievalManager(RETRIEVAL_PERIOD, self.update_mapper, self.logger)
@unittest.skip("Flaky test")
def test_run(self):
cycles = 10
listener = MagicMock()
self.logger.add = MagicMock()
self.retrieval_manager.add_listener(listener)
self._setup_to_do_n_cycles(cycles, self.updates)
self.assertEqual(self.logger.record.call_count, cycles)
listener.assert_has_calls([call(self.updates) for _ in range(cycles)])
def test_run_if_running(self):
Thread(target=self.retrieval_manager.run).start()
self.assertRaises(RuntimeError, self.retrieval_manager.run)
def test_stop_and_then_restart(self):
self.retrieval_manager.start()
self.retrieval_manager.stop()
self.retrieval_manager.start()
def _setup_to_do_n_cycles(self, number_of_cycles: int, updates_each_cycle: UpdateCollection=None):
"""
Sets up the test so that the retriever will only do n cycles.
:param number_of_cycles: the number of cycles to do
"""
if updates_each_cycle is None:
updates_each_cycle = UpdateCollection([])
semaphore = Semaphore(0)
lock_until_counted = Lock()
lock_until_counted.acquire()
def increase_counter(*args) -> UpdateCollection:
semaphore.release()
lock_until_counted.acquire()
return updates_each_cycle
self.retrieval_manager.update_mapper.get_all_since.side_effect = increase_counter
self.retrieval_manager.start()
run_counter = 0
while run_counter < number_of_cycles:
semaphore.acquire()
run_counter += 1
lock_until_counted.release()
if run_counter == number_of_cycles:
self.retrieval_manager.stop()
self.retrieval_manager.update_mapper.get_all_since.side_effect = None
def tearDown(self):
self.retrieval_manager.stop()
def run(config_location):
# Load config
config = load_config(config_location)
# Setup measurement logging
logging_buffer_latency = timedelta(seconds=config.influxdb.buffer_latency)
influxdb_config = InfluxDBConnectionConfig(config.influxdb.host, config.influxdb.port, config.influxdb.username,
config.influxdb.password, config.influxdb.database)
logger = InfluxDBLogger(influxdb_config, buffer_latency=logging_buffer_latency)
# Set HTTP connection pool size (for CouchDB)
# NOTE This is taken from an environment variable, as it's not
# something that would probably need tweaking that much:
pool_size = int(environ.get('COOKIEMONSTER_HTTP_POOL_SIZE', 16))
patch_http_connection_pool(maxsize=pool_size)
# Setup cookie jar
cookie_jar = RateLimitedBiscuitTin(config.cookie_jar.max_requests_per_second, config.cookie_jar.url,
config.cookie_jar.database)
add_cookie_jar_logging(cookie_jar, logger)
# Setup data retrieval manager
update_mapper = BatonUpdateMapper(config.baton.binaries_location, zone=config.baton.zone)
retrieval_manager = PeriodicRetrievalManager(config.retrieval.period, update_mapper, logger)
# # Setup basic Slack client
# slack = BasicSlackClient(config.slack.token, config.slack.default_channel, config.slack.default_username)
#
# # Setup basic message queue (e.g. RabbitMQ) client
# message_queue = BasicMessageQueue(config.message_queue.host, config.message_queue.port,
# config.message_queue.username, config.message_queue.password)
slack = message_queue = None
# Define the context that rules and enrichment loaders has access to
context = HgiContext(cookie_jar, config, slack, message_queue)
# Setup rules source
rules_source = RuleSource(config.processing.rules_location, context)
rules_source.start()
# Setup enrichment loader source
enrichment_loader_source = EnrichmentLoaderSource(config.processing.enrichment_loaders_location, context)
enrichment_loader_source.start()
# Setup the data processor manager
processor_manager = BasicProcessorManager(cookie_jar, rules_source, enrichment_loader_source,
config.processing.max_threads, logger)
# Connect components to the cookie jar
_connect_retrieval_manager_to_cookie_jar(retrieval_manager, cookie_jar, config.cookie_jar.max_requests_per_second,
logger)
_connect_processor_manager_to_cookie_jar(processor_manager, cookie_jar)
# Setup the HTTP API
api = HTTP_API()
api.inject(APIDependency.CookieJar, cookie_jar)
api.listen(config.api.port)
# Start the retrieval manager
retrieval_manager.start(config.retrieval.since)
# Start processing of any unprocessed cookies
processor_manager.process_any_cookies()
# Setup monitors
ThreadsMonitor(logger, logging_buffer_latency).start()
CookieJarMonitor(logger, logging_buffer_latency, cookie_jar).start()
