def main(brokers, topic):
consumer = KafkaConsumer(bootstrap_servers=brokers)
print(f"Topics = {consumer.topics()}")
tp = TopicPartition(topic, 0)
consumer.assign([tp])
# Move to one from the end
consumer.seek_to_end(tp)
end = consumer.position(tp)
consumer.seek(tp, end - 1)
while True:
data = []
while not data:
data = consumer.poll(5)
for message in data[tp]:
print("%s %s:%d:%d: key=%s value=%s" % (
message.timestamp,
message.topic,
message.partition,
message.offset,
message.key,
message.value[0:20],
))
ans = deserialise_hs00(message.value)
print(f"\nHistogram data:\n{ans}")
print(f"Total events: {ans['data'].sum()}")
def test_if_histogram_has_id_then_that_is_added_to_the_info_field(self):
histogrammer = create_histogrammer(self.hist_sink, START_CONFIG)
histogrammer.add_data(EVENT_DATA)
histogrammer.publish_histograms()
data = deserialise_hs00(self.spy_producer.messages[0][1])
info = json.loads(data["info"])
assert info["id"] == "abcdef"
def test_published_histogram_has_non_default_timestamp_set(self):
histogrammer = create_histogrammer(self.hist_sink, STOP_CONFIG)
histogrammer.add_data(EVENT_DATA)
timestamp = 1234567890
histogrammer.publish_histograms(timestamp)
data = deserialise_hs00(self.spy_producer.messages[0][1])
assert data["timestamp"] == timestamp
def test_before_counting_published_histogram_is_labelled_to_indicate_not_started(
self):
histogrammer = create_histogrammer(self.hist_sink, START_CONFIG)
histogrammer.publish_histograms()
data = deserialise_hs00(self.spy_producer.messages[0][1])
info = json.loads(data["info"])
assert info["state"] == HISTOGRAM_STATES["INITIALISED"]
def test_serialises_hs00_message_with_info_field_filled_out_correctly(
self):
"""
Sanity check: checks the combination of libraries work as expected.
"""
info_message = "info_message"
buf = serialise_hs00(self.hist_1d, info_message=info_message)
hist = deserialise_hs00(buf)
assert hist["info"] == info_message
def test_after_stop_published_histogram_is_labelled_to_indicate_finished(
self):
histogrammer = create_histogrammer(self.hist_sink, STOP_CONFIG)
histogrammer.add_data(EVENT_DATA)
histogrammer.publish_histograms()
data = deserialise_hs00(self.spy_producer.messages[0][1])
info = json.loads(data["info"])
assert info["state"] == HISTOGRAM_STATES["FINISHED"]
def test_while_counting_published_histogram_is_labelled_to_indicate_counting(
self):
histogrammer = create_histogrammer(self.hist_sink, START_CONFIG)
histogrammer.add_data(EVENT_DATA)
histogrammer.publish_histograms()
data = deserialise_hs00(self.spy_producer.messages[0][1])
info = json.loads(data["info"])
assert info["state"] == HISTOGRAM_STATES["COUNTING"]
def get_hist_data_from_kafka(self):
data = []
# Move it to one from the end so we can read the final histogram
self.consumer.seek_to_end(self.topic_part)
end_pos = self.consumer.position(self.topic_part)
self.consumer.seek(self.topic_part, end_pos - 1)
while not data:
data = self.consumer.poll(5)
msg = data[self.topic_part][-1]
return deserialise_hs00(msg.value)
def test_if_timestamp_not_supplied_then_it_is_zero(self):
"""
Sanity check: checks the combination of libraries work as expected.
"""
buf = serialise_hs00(self.hist_1d)
hist = deserialise_hs00(buf)
assert hist["source"] == "just-bin-it"
assert hist["timestamp"] == 0
assert hist["current_shape"] == [self.hist_1d.num_bins]
assert np.array_equal(hist["dim_metadata"][0]["bin_boundaries"],
self.hist_1d.x_edges.tolist())
assert hist["dim_metadata"][0]["length"] == self.hist_1d.num_bins
assert np.array_equal(hist["data"], self.hist_1d.data)
def test_serialises_hs00_message_correctly_for_1d(self):
"""
Sanity check: checks the combination of libraries work as expected.
"""
timestamp = 1234567890
buf = serialise_hs00(self.hist_1d, timestamp)
hist = deserialise_hs00(buf)
assert hist["source"] == "just-bin-it"
assert hist["timestamp"] == timestamp
assert hist["current_shape"] == [self.hist_1d.num_bins]
assert np.array_equal(hist["dim_metadata"][0]["bin_boundaries"],
self.hist_1d.x_edges.tolist())
assert hist["dim_metadata"][0]["length"] == self.hist_1d.num_bins
assert np.array_equal(hist["data"], self.hist_1d.data)
def test_deserialises_hs00_message_correctly(self):
"""
Sanity check: checks the combination of libraries work as expected.
"""
data = deserialise_hs00(self.buf)
assert data["source"] == "just-bin-it"
assert data["timestamp"] == 987_654_321
assert data["current_shape"] == [50]
assert len(data["data"]) == 50
assert len(data["dim_metadata"]) == 1
assert data["info"] == "hello"
assert data["dim_metadata"][0]["length"] == 50
assert len(data["dim_metadata"][0]["bin_boundaries"]) == 51
assert data["dim_metadata"][0]["bin_boundaries"][0] == 0.0
assert data["dim_metadata"][0]["bin_boundaries"][50] == 100_000_000.0
def _process_record(self, record):
try:
return deserialise_hs00(record)
except Exception as error:
raise SourceException(error)
def test_if_schema_is_incorrect_then_throws(self):
new_buf = self.buf[:4] + b"na12" + self.buf[8:]
with pytest.raises(JustBinItException):
deserialise_hs00(new_buf)
