def test_basic():
print('news_fetcher_test: cleaning all queues...')
queue_cleaner.clear_all()
print('flushing all cache in Redis')
redis_client.flushall() # TODO dangerous to run when deployed
scrape_queue_client = AMQPClient(SCRAPE_QUEUE_URL,
SCRAPE_NEWS_TASK_QUEUE_NAME)
scrape_queue_client.connect()
assert scrape_queue_client.is_connected()
print('test_fetcher_basic: adding news onto scrape queue...')
for message in TEST_SCRAPE_TASK:
scrape_queue_client.send_message(message)
print('getting messages from the queue and process...')
news_fetcher.SLEEP_TIME_IN_SECONDS = 1
news_fetcher.run(len(TEST_SCRAPE_TASK))
should_be_empty_msg = scrape_queue_client.get_message()
print('news_fetcher_test(expecting None):', should_be_empty_msg)
assert should_be_empty_msg is None
scrape_queue_client.close()
queue_cleaner.clear_queue(DEDUPE_QUEUE_URL, DEDUPE_NEWS_TASK_QUEUE_NAME)
print('news_fetcher test passed')
def test_basic():
# print(amqp_url + "({})".format(type(amqp_url)))
client = AMQPClient(AMQP_URL, 'my_queue')
client.connect()
assert client.is_connected()
assert client.get_message() is None
client.send_message('hello world')
assert client.get_message() == 'hello world'
obj = {"hello": "world"}
client.send_message(obj)
assert client.get_message() == obj
assert client.get_message() is None
client.cancel_queue()
client.close()
print('[x] cloud amqp_client test passed')
def run():
redis_client = redis.StrictRedis(REDIS_HOST, REDIS_PORT)
amqp_client = AMQPClient(SCRAPE_NEWS_TASK_QUEUE_URL,
SCRAPE_NEWS_TASK_QUEUE_NAME)
amqp_client.connect()
try:
while True:
news_list = news_client.get_news_from_sources(NEWS_SOURCES)
num_news = 0
for news in news_list:
digest = hashlib.md5(news['title'].encode('utf-8')).hexdigest()
if redis_client.get(digest):
continue
num_news += 1
news['digest'] = digest
redis_client.set(digest, True)
redis_client.expire(digest, NEWS_TIME_OUT_IN_SECONDS)
print(news)
amqp_client.send_message(news)
print('News Monitor: fectched {} news'.format(num_news))
amqp_client.sleep(SLEEP_TIME_IN_SECONDS)
except KeyboardInterrupt:
print('keyboard interrupt')
# except SigTerm
finally:
amqp_client.close()
def test_monitor_basic():
news_monitor.NEWS_SOURCES = news_client.MOCK_SOURCES
MOCK_DATA = news_client.MOCK_DATA
print('test_monitor_basic: cleaning queue "{}" first---------'.format(QUEUE_NAME))
clear_queue(QUEUE_URL, QUEUE_NAME)
# TODO: redis server flush all
redis_client = redis.StrictRedis(REDIS_HOST, REDIS_PORT)
redis_client.flushall()
print('test_monitor_basic: adding message to queue "{}"--------'.format(QUEUE_NAME))
amqp_client = AMQPClient(QUEUE_URL, QUEUE_NAME)
amqp_client.connect()
proc = Process(target=news_monitor.run, name='monitor_run',
args=(REDIS_HOST, REDIS_PORT, QUEUE_URL, QUEUE_NAME))
proc.start()
print('test_monitor_basic: executing... (wait for 2 seconds to cut)')
time.sleep(2)
proc.terminate()
for i in range(len(MOCK_DATA)):
message = amqp_client.get_message()
del message['digest']
print(message, MOCK_DATA[i])
assert message == MOCK_DATA[i]
print('test_monitor_basic: [x] test_monitor_basic test passed')
def run(redis_host=REDIS_HOST,
redis_port=REDIS_PORT,
scrape_queue_url=SCRAPE_NEWS_TASK_QUEUE_URL,
scrape_queue_name=SCRAPE_NEWS_TASK_QUEUE_NAME):
redis_client = redis.StrictRedis(redis_host, redis_port)
amqp_client = AMQPClient(scrape_queue_url, scrape_queue_name)
amqp_client.connect()
while True:
logger.debug('News monitor: iter..')
news_list = news_client.get_news_from_sources(NEWS_SOURCES)
num_news = 0
for news in news_list:
digest = hashlib.md5(news['title'].encode('utf-8')).hexdigest()
if redis_client.get(digest):
continue
num_news += 1
news['digest'] = digest
redis_client.set(digest, True)
redis_client.expire(digest, NEWS_TIME_OUT_IN_SECONDS)
logger.debug('News Monitor: got news {}'.format(news))
amqp_client.send_message(news)
logger.info('News Monitor: fectched {} news'.format(num_news))
amqp_client.sleep(SLEEP_TIME_IN_SECONDS)
def run(scrape_queue_url=SCRAPE_QUEUE_URL, scrape_queue_name=SCRAPE_NEWS_TASK_QUEUE_NAME,
dedupe_queue_url=DEDUPE_QUEUE_URL, dedupe_queue_name=DEDUPE_NEWS_TASK_QUEUE_NAME,
times=-1):
scrape_queue_client = AMQPClient(scrape_queue_url, scrape_queue_name)
scrape_queue_client.connect()
dedupe_queue_client = AMQPClient(dedupe_queue_url, dedupe_queue_name)
dedupe_queue_client.connect()
assert scrape_queue_client.is_connected()
assert dedupe_queue_client.is_connected()
while True:
logger.debug('News fetcher: iter..')
msg = scrape_queue_client.get_message()
if msg is not None:
try:
handle_message(msg)
dedupe_queue_client.send_message(msg)
logger.info('News Fetcher: message sent to dedupe queue (url: {})'
.format(msg.get('url')))
except Exception as e:
logger.warning('News fetcher: handling error: {}'.format(e))
# if decreas count here, weird behavior, decreasing happens before processing message
scrape_queue_client.sleep(SLEEP_TIME_IN_SECONDS)
if times > 0: times -= 1
if times == 0: break
def test_basic():
# not a good idea to test, but a way I can think of seperating the queue
# in test and production
rpc_operations.USER_CLICK_QUEUE_URL = USER_CLICK_QUEUE_URL
rpc_operations.USER_CLICK_QUEUE_NAME = USER_CLICK_QUEUE_NAME
rpc_operations.init()
clickhandler.USER_CLICK_QUEUE_URL = USER_CLICK_QUEUE_URL
clickhandler.USER_CLICK_QUEUE_NAME = USER_CLICK_QUEUE_NAME
print('Click Handler test: cleaning up click queue...')
clear_queue(USER_CLICK_QUEUE_URL, USER_CLICK_QUEUE_NAME)
print('Click Handler test: logging clicks...')
for pair in TEST_CLICK_DATA:
rpc_operations.log_click(*pair)
print(print('Click Handler test: handling clicks...'))
clickhandler.run(len(TEST_CLICK_DATA))
queue_client = AMQPClient(USER_CLICK_QUEUE_URL, USER_CLICK_QUEUE_NAME)
queue_client.connect()
should_be_empty = queue_client.get_message()
print('Click Handler Test: {} should be None'.format(should_be_empty))
assert should_be_empty is None
queue_client.close
print('Click Handler: test passed')
def clear_queue(queue_url, queue_name):
count = 0
amqp_client = AMQPClient(queue_url, queue_name)
amqp_client.connect()
print('xxx cleaning queue "{}"'.format(queue_name))
while True:
message = amqp_client.get_message()
if message is None:
break
print(message)
count += 1
amqp_client.close()
print('xxx cleaned {} message on {}'.format(count, queue_name))
def run(times=-1):
click_queue_client = AMQPClient(USER_CLICK_QUEUE_URL,
USER_CLICK_QUEUE_NAME)
click_queue_client.connect()
assert click_queue_client.is_connected()
print('Click Handler: my queue name: {}'.format(click_queue_client))
while True:
message = click_queue_client.get_message()
try:
handle_message(message)
except Exception as e:
raise e
click_queue_client.sleep(SLEEP_TIME_IN_SECONDS)
if times > 0: times -= 1
if times == 0: break
def test_basic():
# print('Click Learner tester: clearing queue {}'.format(CLICK_QUEUE_NAME))
clear_queue(clicklearner.USER_CLICK_QUEUE_URL, CLICK_QUEUE_NAME)
clicklearner.PREF_COLLECTION_NAME = PREF_COLLECTION_NAME
clicklearner.USER_CLICK_QUEUE_NAME = CLICK_QUEUE_NAME
clicklearner.SLEEP_TIME_IN_SECONDS = 1
click_queue_client = AMQPClient(clicklearner.USER_CLICK_QUEUE_URL,
CLICK_QUEUE_NAME)
click_queue_client.connect()
assert click_queue_client.is_connected()
print('Click Learner tester: clearing collection "{}" in db "{}"...'.format(PREF_COLLECTION_NAME,
PREF_DB_NAME))
pref_collection.remove()
print('Click Learner tester: sending click logs')
for click in TEST_SEQUENCE:
click_queue_client.send_message(click)
print('Click Learner tester: start handling clicks')
pref_model_ref = {'userId': TEST_USER_NAME}
for cat in NEWS_CATEGORIES:
pref_model_ref[cat] = 1 / len(NEWS_CATEGORIES)
for click_log in TEST_SEQUENCE:
clicklearner.run(1)
pref_model = pref_collection.find_one({'userId': TEST_USER_NAME})
selected = news_collection.find_one({'digest': (click_log['newsDigest'])})['category']
pref_model_ref[selected] = (1 - ALPHA) * pref_model_ref[selected] + ALPHA
for cat in NEWS_CATEGORIES:
if cat != selected:
pref_model_ref[cat] = (1 - ALPHA) * pref_model_ref[cat]
del pref_model['_id']
print('Click Learner tester: expecting {} == {}'.format(pref_model, pref_model_ref))
assert pref_model == pref_model_ref
print('xx Click Learner test passed')
def run(times=-1):
click_queue_client = AMQPClient(USER_CLICK_QUEUE_URL,
USER_CLICK_QUEUE_NAME)
click_queue_client.connect()
news_collection = mongodb_client.get_news_collection()
pref_collection = (mongodb_client.get_db(PREF_DB_NAME).get_collection(
PREF_COLLECTION_NAME))
assert click_queue_client.is_connected()
while True:
msg = click_queue_client.get_message()
try:
handle_message(msg, news_collection, pref_collection)
except Exception as e:
raise e
click_queue_client.sleep(SLEEP_TIME_IN_SECONDS)
if times > 0: times -= 1
if times == 0: break
logger = logging.getLogger(__name__)
coloredlogs.install(level=os.environ.get('LOGGER_LEVEL', 'INFO'),
logger=logger)
# TODO: this global are bad, and makes it uncovered by tests
config = os.environ
DB_NAME = config['news_db']
COLLECTION_NAME = config['new_collection']
DEDUPE_QUEUE_URL = config['dedupe_task_queue_url']
DEDUPE_QUEUE_NAME = config['dedupe_task_queue_name']
SLEEP_TIME_IN_SECONDS = 5
dedupe_queue_client = AMQPClient(DEDUPE_QUEUE_URL, DEDUPE_QUEUE_NAME)
dedupe_queue_client.connect()
assert dedupe_queue_client.is_connected()
NEWS_SIMILARITY_THRESHOLD = 0.8
class NotContainPublishTimeError(Exception):
def __str__(self):
return 'News not containing publish time!!!'
def handle_message(msg):
logger.debug('dedupter handling message: {}'.format(msg))
if msg is None or not isinstance(msg, dict):
logger.info('News Deduper: message is broken')