def test_multiple_consumers(redis: StrictRedis, data):
out = Stream().pluck(2)
S = set()
out.pluck("i").sink(S.add)
stream, group = uuid(2)
sources = set()
for _ in range(3):
con = uuid()
source = Stream.from_redis_consumer_group(
stream,
group,
con,
count=1,
timeout=0.1,
)
source.connect(out)
source.start()
sources.add(source)
for x in data:
redis.xadd(stream, x)
wait_for(lambda: len(S) == 50, 1)
assert S == set(x["i"] for x in data)
for s in sources:
s.stop()
def test_heartbeats(redis: StrictRedis):
stream, group = uuid(2)
redis.xgroup_create(stream, group, mkstream=True)
interval = 0.1
timeout = 0.5
hearts = []
for _ in range(5):
heart = Heart(stream,
group,
uuid(),
interval=interval,
timeout=timeout)
hearts.append(heart)
heart.start()
S = set()
sub = redis.pubsub()
sub.subscribe(group)
def predicate():
m = sub.get_message()
if m is not None:
S.add(m["data"])
return len(S) == 5
wait_for(predicate, 5, period=0.01)
for h in hearts:
h.stop()
def run_and_fail():
name = uuid()
source = Stream.from_redis_consumer_group(
stream,
group,
name,
count=1,
timeout=0.1,
)
buffer = source.buffer(10)
buffer.rate_limit(0.1).pluck(1).sink_to_redis_list(target)
source.start()
wait_for(lambda: buffer.queue.qsize() == 10, 3)
buffer.queue = Queue(10) # lose data in the buffer, won't be ACKed
source.stop()
def pending_10():
cons = convert_bytes(redis.xpending(stream, group))["consumers"]
for con in cons:
if con["name"] == name and con["pending"] == 10:
return True
return False
wait_for(pending_10, 1, period=0.1)
def test_backpressure_connect_empty_stream():
@Stream.register_api()
class from_list(Stream):
def __init__(self, source, **kwargs):
self.source = source
super().__init__(ensure_io_loop=True, **kwargs)
def start(self):
self.stopped = False
self.loop.add_callback(self.run)
@gen.coroutine
def run(self):
while not self.stopped and len(self.source) > 0:
yield self._emit(self.source.pop(0))
source_list = [0, 1, 2, 3, 4]
source = Stream.from_list(source_list)
sout = Stream()
L = sout.rate_limit(1).sink_to_list()
source.connect(sout)
source.start()
wait_for(lambda: L == [0], 0.01)
assert len(source_list) > 0
def test_simple():
cat = intake.open_catalog(catfile)
s = cat.simple.read()
l = s.sink_to_list()
assert not l
s.start()
wait_for(lambda: l == [1, 2, 3], timeout=1)
def test_from_kafka():
j = random.randint(0, 10000)
ARGS = {
'bootstrap.servers': 'localhost:9092',
'group.id': 'streamz-test%i' % j
}
with kafka_service() as kafka:
kafka, TOPIC = kafka
stream = Stream.from_kafka([TOPIC], ARGS, asynchronous=True)
out = stream.sink_to_list()
stream.start()
yield gen.sleep(0.1) # for loop to run
for i in range(10):
yield gen.sleep(0.2)
kafka.produce(TOPIC, b'value-%d' % i)
kafka.flush()
# it takes some time for messages to come back out of kafka
wait_for(lambda: len(out) == 10, 10, period=0.1)
assert out[-1] == b'value-9'
kafka.produce(TOPIC, b'final message')
kafka.flush()
wait_for(lambda: out[-1] == b'final message', 10, period=0.1)
stream._close_consumer()
kafka.produce(TOPIC, b'lost message')
kafka.flush()
# absolute sleep here, since we expect output list *not* to change
yield gen.sleep(1)
assert out[-1] == b'final message'
stream._close_consumer()
def launch_kafka():
stop_docker(let_fail=True)
subprocess.call(shlex.split("docker pull spotify/kafka"))
cmd = ("docker run -d -p 2181:2181 -p 9092:9092 --env "
"ADVERTISED_HOST=127.0.0.1 --env ADVERTISED_PORT=9092 "
"--name streamz-kafka spotify/kafka")
print(cmd)
cid = subprocess.check_output(shlex.split(cmd)).decode()[:-1]
def end():
if cid:
stop_docker(cid=cid)
atexit.register(end)
def predicate():
try:
out = subprocess.check_output(['docker', 'logs', cid],
stderr=subprocess.STDOUT)
return b'kafka entered RUNNING state' in out
except subprocess.CalledProcessError:
pass
wait_for(predicate, 10, period=0.1)
return cid
def test_from_kafka_thread():
j = random.randint(0, 10000)
ARGS = {
'bootstrap.servers': 'localhost:9092',
'group.id': 'streamz-test%i' % j
}
with kafka_service() as kafka:
stream = Stream.from_kafka([TOPIC], ARGS)
out = stream.sink_to_list()
stream.start()
for i in range(10):
kafka.produce(TOPIC, b'value-%d' % i)
kafka.flush()
# it takes some time for messages to come back out of kafka
startlen = len([o for o in out if o])
wait_for(lambda: len([o for o in out if o]) == startlen + 10,
10,
period=0.1)
assert out[-1] == b'value-9'
kafka.produce(TOPIC, b'final message')
kafka.flush()
wait_for(lambda: out[-1] == b'final message', 10, period=0.1)
stream._close_consumer()
kafka.produce(TOPIC, b'lost message')
kafka.flush()
# absolute sleep here, since we expect output list *not* to change
sleep(1)
assert out[-1] == b'final message'
def test_tcp():
port = 9876
s = Source.from_tcp(port)
out = s.sink_to_list()
s.start()
wait_for(lambda: s.server is not None, 2, period=0.02)
try:
sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
sock.connect(("localhost", port))
sock.send(b'data\n')
sock.close()
sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
sock.connect(("localhost", port))
sock.send(b'data\n')
sock2 = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
sock2.connect(("localhost", port))
sock2.send(b'data2\n')
wait_for(lambda: out == [b'data\n', b'data\n', b'data2\n'],
2,
period=0.01)
finally:
s.stop()
sock.close()
sock2.close()
def test_kafka_batch():
j = random.randint(0, 10000)
ARGS = {
'bootstrap.servers': 'localhost:9092',
'group.id': 'streamz-test%i' % j,
'auto.offset.reset': 'latest'
}
with kafka_service() as kafka:
kafka, TOPIC = kafka
# These messages aren't read since Stream starts reading from latest offsets
for i in range(10):
kafka.produce(TOPIC, b'value-%d' % i, b'%d' % i)
kafka.flush()
stream = Stream.from_kafka_batched(TOPIC,
ARGS,
max_batch_size=4,
keys=True)
out = stream.sink_to_list()
stream.start()
wait_for(lambda: stream.upstream.started, 10, 0.1)
for i in range(10):
kafka.produce(TOPIC, b'value-%d' % i, b'%d' % i)
kafka.flush()
# out may still be empty or first item of out may be []
wait_for(lambda: any(out) and out[-1][-1]['value'] == b'value-9',
10,
period=0.2)
assert out[-1][-1]['key'] == b'9'
# max_batch_size checks
assert len(out[0]) == len(out[1]) == 4 and len(out) == 3
stream.upstream.stopped = True
def test_from_iterable_backpressure():
it = iter(range(5))
source = Source.from_iterable(it)
L = source.rate_limit(0.1).sink_to_list()
source.start()
wait_for(lambda: L == [0], 1, period=0.01)
assert next(it) == 2 # 1 is in blocked _emit
def test_periodic():
s = Source.from_periodic(lambda: True)
l = s.sink_to_list()
assert s.stopped
s.start()
wait_for(lambda: l, 0.3, period=0.01)
wait_for(lambda: len(l) > 1, 0.3, period=0.01)
assert all(l)
def test_df():
dataframe = pytest.importorskip("streamz.dataframe")
cat = intake.open_catalog(catfile)
s = cat.df.read()
assert isinstance(s, dataframe.DataFrame)
s.start()
wait_for(lambda: s.current_value is not None, timeout=1)
cv = s.current_value
wait_for(lambda: not s.current_value.equals(cv), timeout=1)
def test_from_redis_lists(redis: StrictRedis):
name = uuid()
source = Stream.from_redis_lists(name, timeout=0.1)
L = source.pluck(1).map(int).sink_to_list()
source.start()
redis.rpush(name, *list(range(3)))
wait_for(lambda: L == [0, 1, 2], 3)
source.stop()
def test_dask():
distr = pytest.importorskip("dask.distributed")
import streamz.dask
with distr.Client(processes=False):
cat = intake.open_catalog(catfile)
s = cat.simple.to_dask()
assert isinstance(s, streamz.dask.DaskStream)
l = s.gather().sink_to_list()
s.start()
wait_for(lambda: l == [1, 2, 3], timeout=1)
def test_claim(redis: StrictRedis, data):
stream, group, target = uuid(3)
for x in data:
redis.xadd(stream, x)
def run_and_fail():
name = uuid()
source = Stream.from_redis_consumer_group(
stream,
group,
name,
count=1,
timeout=0.1,
)
buffer = source.buffer(10)
buffer.rate_limit(0.1).pluck(1).sink_to_redis_list(target)
source.start()
wait_for(lambda: buffer.queue.qsize() == 10, 3)
buffer.queue = Queue(10) # lose data in the buffer, won't be ACKed
source.stop()
def pending_10():
cons = convert_bytes(redis.xpending(stream, group))["consumers"]
for con in cons:
if con["name"] == name and con["pending"] == 10:
return True
return False
wait_for(pending_10, 1, period=0.1)
for _ in range(10):
run_and_fail()
source = Stream.from_redis_consumer_group(
stream,
group,
uuid(),
heartbeat_interval=0.1,
claim_timeout=1,
count=10,
timeout=0.1,
)
source.pluck(1).sink_to_redis_list(target)
source.start()
wait_for(
lambda: redis.llen(target) == 500,
15,
lambda: print(redis.llen(target)),
period=0.1,
)
source.stop()
def test_basic(redis: StrictRedis, data):
stream = uuid()
source = Stream.from_redis_streams(stream, timeout=0.1, default_start_id=0)
L = source.sink_to_list()
source.start()
for x in data:
redis.xadd(stream, x)
wait_for(lambda: len(L) == 3, 2)
assert [x[2] for x in L] == data
source.stop()
def test_multiple(redis: StrictRedis):
l1, l2 = uuid(2)
source = Stream.from_redis_lists([l1, l2], timeout=0.1)
L = source.pluck(1).map(int).sink_to_list()
source.start()
redis.rpush(l1, *list(range(3)))
redis.rpush(l2, *list(range(3)))
wait_for(lambda: len(L) == 6, 2)
source.stop()
def test_index(stream):
df = pd.DataFrame({'x': [1, 2, 3], 'y': [4, 5, 6]})
a = DataFrame(example=df, stream=stream)
b = a.index + 5
L = b.stream.gather().sink_to_list()
a.emit(df)
a.emit(df)
wait_for(lambda: len(L) > 1, timeout=2, period=0.05)
assert_eq(L[0], df.index + 5)
assert_eq(L[1], df.index + 5)
def test_ws_roundtrip():
pytest.importorskip("websockets")
s0 = Stream.from_websocket("localhost", 8989, start=True)
l = s0.sink_to_list()
data = [b'0123'] * 4
s = Stream.from_iterable(data)
s.to_websocket("ws://localhost:8989")
s.start()
wait_for(lambda: data == l, timeout=1)
s.stop()
s0.stop()
def test_from_iterable_stop():
from _pytest.outcomes import Failed
source = Source.from_iterable(range(5))
L = source.rate_limit(0.01).sink_to_list()
source.start()
wait_for(lambda: L == [0], 1)
source.stop()
assert source.stopped
with pytest.raises(Failed):
wait_for(lambda: L == [0, 1, 2], 0.1)
def test_mqtt_roundtrip():
pytest.importorskip("paho.mqtt.client")
s0 = Stream.from_mqtt("mqtt.eclipseprojects.io", 1883,
"streamz/sensor/temperature")
l = s0.map(lambda msg: msg.payload).sink_to_list()
s0.start()
data = [b'0123'] * 4
s = Stream.from_iterable(data)
s.to_mqtt("mqtt.eclipseprojects.io", 1883, "streamz/sensor/temperature")
s.start()
wait_for(lambda: data == l, timeout=1)
s.stop()
s0.stop()
def test_increment_restart(pg):
table = "inc_re"
src = Stream.from_postgres_increment(table,
pg,
initial_value=30,
polling_interval=1,
limit=10)
L = src.sink_to_list()
w = Writer(src.strategy.loader.connection, table)
w.create_table()
w.insert(50)
src.start()
wait_for(lambda: len(L) == 20, 1, period=0.1)
def test_to_kafka():
ARGS = {'bootstrap.servers': 'localhost:9092'}
with kafka_service() as kafka:
_, TOPIC = kafka
source = Stream()
kafka = source.to_kafka(TOPIC, ARGS)
out = kafka.sink_to_list()
for i in range(10):
yield source.emit(b'value-%d' % i)
source.emit('final message')
kafka.flush()
wait_for(lambda: len(out) == 11, 10, period=0.1)
assert out[-1] == b'final message'
def test_source(clients, info):
topic, subscription = info
source = Stream.from_gcp_pubsub(subscription, timeout=1)
L = source.map(lambda x: int(x.data)).sink_to_list()
source.start()
pub, _ = clients
for i in range(10):
pub.publish(topic, str(i).encode())
wait_for(lambda: list(range(10)) == L, 1)
with pytest.raises(Failed):
wait_for(lambda: len(L) > 10, 3) # test ack
source.stop()
def test_ack(redis: StrictRedis, data):
stream, group, con = uuid(3)
source = Stream.from_redis_consumer_group(stream, group, con, timeout=0.1)
L = source.sink_to_list()
for x in data:
redis.xadd(stream, x)
source.start()
wait_for(lambda: len(L) == 3, 3, lambda: print(L))
sleep(0.05) # wait a bit for the last ack
for _, messages in redis.xreadgroup(group, con, {stream: 0}):
assert messages == []
source.stop()
def redis(name="test-streamz-redis"):
cleanup(name=name)
run_cmd = shlex.split(f"docker run -d -p 6379:6379 --name {name} redis")
subprocess.check_call(run_cmd)
def predicate():
cmd = shlex.split(f"docker logs {name}")
logs = subprocess.check_output(cmd, stderr=subprocess.STDOUT)
return b"Ready to accept connections" in logs
wait_for(predicate, 10, period=0.1)
try:
with StrictRedis() as client:
yield client
finally:
cleanup(name=name, fail=True)
def test_multiple(redis: StrictRedis, data):
stream1, stream2 = uuid(2)
source = Stream.from_redis_streams({stream1: 0, stream2: 0}, timeout=0.1)
L1 = source.pluck(0).filter(lambda x: x == stream1).sink_to_list()
L2 = source.pluck(0).filter(lambda x: x == stream2).sink_to_list()
source.start()
for x in data:
redis.xadd(stream1, x)
redis.xadd(stream2, x)
wait_for(lambda: len(L1) == 3, 3)
wait_for(lambda: len(L2) == 3, 3)
assert L1 == [stream1] * 3
assert L2 == [stream2] * 3
source.stop()
def test_kafka_batch_checkpointing_sync_nodes():
'''
Streams 1 and 3 have different consumer groups, while Stream 2
has the same group as 1. Hence, Stream 2 does not re-read the
data that had been finished processing by Stream 1, i.e. it
picks up from where Stream 1 had left off.
'''
j1 = random.randint(0, 10000)
ARGS1 = {
'bootstrap.servers': 'localhost:9092',
'group.id': 'streamz-test%i' % j1,
'enable.auto.commit': False,
'auto.offset.reset': 'earliest'
}
j2 = j1 + 1
ARGS2 = {
'bootstrap.servers': 'localhost:9092',
'group.id': 'streamz-test%i' % j2,
'enable.auto.commit': False,
'auto.offset.reset': 'earliest'
}
with kafka_service() as kafka:
kafka, TOPIC = kafka
for i in range(10):
kafka.produce(TOPIC, b'value-%d' % i)
kafka.flush()
stream1 = Stream.from_kafka_batched(TOPIC, ARGS1)
out1 = stream1.map(split).filter(
lambda x: x[-1] % 2 == 1).sink_to_list()
stream1.start()
wait_for(lambda: any(out1) and out1[-1][-1] == 9, 10, period=0.2)
stream1.upstream.stopped = True
stream2 = Stream.from_kafka_batched(TOPIC, ARGS1)
out2 = stream2.map(split).filter(
lambda x: x[-1] % 2 == 1).sink_to_list()
stream2.start()
time.sleep(5)
assert len(out2) == 0
stream2.upstream.stopped = True
stream3 = Stream.from_kafka_batched(TOPIC, ARGS2)
out3 = stream3.map(split).filter(
lambda x: x[-1] % 2 == 1).sink_to_list()
stream3.start()
wait_for(lambda: any(out3) and out3[-1][-1] == 9, 10, period=0.2)
stream3.upstream.stopped = True
def test_kafka_batch():
j = random.randint(0, 10000)
ARGS = {
'bootstrap.servers': 'localhost:9092',
'group.id': 'streamz-test%i' % j
}
with kafka_service() as kafka:
stream = Stream.from_kafka_batched(TOPIC, ARGS)
out = stream.sink_to_list()
stream.start()
for i in range(10):
kafka.produce(TOPIC, b'value-%d' % i)
kafka.flush()
# out may still be empty or first item of out may be []
wait_for(lambda: any(out) and out[-1][-1] == b'value-9',
10,
period=0.2)
stream.stopped = True
