class GenerateRecords(beam.DoFn):
EMIT_TIMER = TimerSpec('emit_timer', TimeDomain.REAL_TIME)
COUNT_STATE = CombiningValueStateSpec('count_state', VarIntCoder(),
CountCombineFn())
def __init__(self, frequency, total_records):
self.total_records = total_records
self.frequency = frequency
def process(self,
element,
emit_timer=beam.DoFn.TimerParam(EMIT_TIMER)):
# Processing time timers should be set on ABSOLUTE TIME.
emit_timer.set(self.frequency)
yield element[1]
@on_timer(EMIT_TIMER)
def emit_values(self,
emit_timer=beam.DoFn.TimerParam(EMIT_TIMER),
count_state=beam.DoFn.StateParam(COUNT_STATE)):
count = count_state.read() or 0
if self.total_records == count:
return
count_state.add(1)
# Processing time timers should be set on ABSOLUTE TIME.
emit_timer.set(count + 1 + self.frequency)
yield 'value'
def _pardo_group_into_batches(batch_size, input_coder):
ELEMENT_STATE = BagStateSpec('values', input_coder)
COUNT_STATE = CombiningValueStateSpec('count', input_coder, CountCombineFn())
EXPIRY_TIMER = TimerSpec('expiry', TimeDomain.WATERMARK)
class _GroupIntoBatchesDoFn(DoFn):
def process(self, element,
window=DoFn.WindowParam,
element_state=DoFn.StateParam(ELEMENT_STATE),
count_state=DoFn.StateParam(COUNT_STATE),
expiry_timer=DoFn.TimerParam(EXPIRY_TIMER)):
# Allowed lateness not supported in Python SDK
# https://beam.apache.org/documentation/programming-guide/#watermarks-and-late-data
expiry_timer.set(window.end)
element_state.add(element)
count_state.add(1)
count = count_state.read()
if count >= batch_size:
batch = [element for element in element_state.read()]
yield batch
element_state.clear()
count_state.clear()
@on_timer(EXPIRY_TIMER)
def expiry(self, element_state=DoFn.StateParam(ELEMENT_STATE),
count_state=DoFn.StateParam(COUNT_STATE)):
batch = [element for element in element_state.read()]
if batch:
yield batch
element_state.clear()
count_state.clear()
return _GroupIntoBatchesDoFn()
def _pardo_group_into_batches(
input_coder, batch_size, max_buffering_duration_secs, clock=time.time):
ELEMENT_STATE = BagStateSpec('values', input_coder)
COUNT_STATE = CombiningValueStateSpec('count', input_coder, CountCombineFn())
WINDOW_TIMER = TimerSpec('window_end', TimeDomain.WATERMARK)
BUFFERING_TIMER = TimerSpec('buffering_end', TimeDomain.REAL_TIME)
class _GroupIntoBatchesDoFn(DoFn):
def process(
self,
element,
window=DoFn.WindowParam,
element_state=DoFn.StateParam(ELEMENT_STATE),
count_state=DoFn.StateParam(COUNT_STATE),
window_timer=DoFn.TimerParam(WINDOW_TIMER),
buffering_timer=DoFn.TimerParam(BUFFERING_TIMER)):
# Allowed lateness not supported in Python SDK
# https://beam.apache.org/documentation/programming-guide/#watermarks-and-late-data
window_timer.set(window.end)
element_state.add(element)
count_state.add(1)
count = count_state.read()
if count == 1 and max_buffering_duration_secs > 0:
# This is the first element in batch. Start counting buffering time if a
# limit was set.
# pylint: disable=deprecated-method
buffering_timer.set(clock() + max_buffering_duration_secs)
if count >= batch_size:
return self.flush_batch(element_state, count_state, buffering_timer)
@on_timer(WINDOW_TIMER)
def on_window_timer(
self,
element_state=DoFn.StateParam(ELEMENT_STATE),
count_state=DoFn.StateParam(COUNT_STATE),
buffering_timer=DoFn.TimerParam(BUFFERING_TIMER)):
return self.flush_batch(element_state, count_state, buffering_timer)
@on_timer(BUFFERING_TIMER)
def on_buffering_timer(
self,
element_state=DoFn.StateParam(ELEMENT_STATE),
count_state=DoFn.StateParam(COUNT_STATE),
buffering_timer=DoFn.TimerParam(BUFFERING_TIMER)):
return self.flush_batch(element_state, count_state, buffering_timer)
def flush_batch(self, element_state, count_state, buffering_timer):
batch = [element for element in element_state.read()]
if not batch:
return
key, _ = batch[0]
batch_values = [v for (k, v) in batch]
element_state.clear()
count_state.clear()
buffering_timer.clear()
yield key, batch_values
return _GroupIntoBatchesDoFn()
def expand(self, pcoll):
output = (pcoll
| "ParseJson" >> beam.ParDo(JsonToTaxiRide())
| "FilterForPickups" >>
beam.Filter(lambda x: x.ride_status == 'pickup')
| "WindowByMinute" >> beam.WindowInto(
beam.window.FixedWindows(60),
trigger=AfterWatermark(late=AfterCount(1)),
allowed_lateness=60,
accumulation_mode=AccumulationMode.ACCUMULATING)
| "CountPerMinute" >> beam.CombineGlobally(
CountCombineFn()).without_defaults())
return output
class StatefulBufferingFn(DoFn):
BUFFER_STATE = BagStateSpec('buffer', StrUtf8Coder())
COUNT_STATE = userstate.CombiningValueStateSpec(
'count', VarIntCoder(), CountCombineFn())
def process(self,
element,
buffer_state=beam.DoFn.StateParam(BUFFER_STATE),
count_state=beam.DoFn.StateParam(COUNT_STATE)):
key, value = element
try:
index_value = list(buffer_state.read()).index(value)
except:
index_value = -1
if index_value < 0:
buffer_state.add(value)
index_value = count_state.read()
count_state.add(1)
# print(value, list(buffer_state.read()).index(value), list(buffer_state.read()))
yield ('{}_{}'.format(value, index_value), 1)
def run():
# Command line arguments
parser = argparse.ArgumentParser(
description='Load from Json into BigQuery')
parser.add_argument('--project',
required=True,
help='Specify Google Cloud project')
parser.add_argument('--region',
required=True,
help='Specify Google Cloud region')
parser.add_argument('--staging_location',
required=True,
help='Specify Cloud Storage bucket for staging')
parser.add_argument('--temp_location',
required=True,
help='Specify Cloud Storage bucket for temp')
parser.add_argument('--runner',
required=True,
help='Specify Apache Beam Runner')
parser.add_argument('--input_path',
required=True,
help='Path to events.json')
parser.add_argument('--table_name',
required=True,
help='BigQuery table name')
opts = parser.parse_args()
# Setting up the Beam pipeline options
options = PipelineOptions(save_main_session=True)
options.view_as(GoogleCloudOptions).project = opts.project
options.view_as(GoogleCloudOptions).region = opts.region
options.view_as(
GoogleCloudOptions).staging_location = opts.staging_location
options.view_as(GoogleCloudOptions).temp_location = opts.temp_location
options.view_as(GoogleCloudOptions).job_name = '{0}{1}'.format(
'batch-user-traffic-pipeline-', time.time_ns())
options.view_as(StandardOptions).runner = opts.runner
input_path = opts.input_path
table_name = opts.table_name
# Table schema for BigQuery
table_schema = {
"fields": [
{
"name": "user_id",
"type": "STRING"
},
{
"name": "page_views",
"type": "INTEGER"
},
{
"name": "total_bytes",
"type": "INTEGER"
},
{
"name": "max_bytes",
"type": "INTEGER"
},
{
"name": "min_bytes",
"type": "INTEGER"
},
]
}
# Create the pipeline
p = beam.Pipeline(options=options)
(p | 'ReadFromGCS' >> beam.io.ReadFromText(input_path)
| 'ParseJson' >> beam.Map(parse_json).with_output_types(CommonLog)
| 'PerUserAggregations' >> beam.GroupBy('user_id').aggregate_field(
'user_id', CountCombineFn(), 'page_views').aggregate_field(
'num_bytes', sum, 'total_bytes').aggregate_field(
'num_bytes', max, 'max_bytes').aggregate_field(
'num_bytes', min,
'min_bytes').with_output_types(PerUserAggregation)
| 'ToDict' >> beam.Map(to_dict)
| 'WriteToBQ' >> beam.io.WriteToBigQuery(
table_name,
schema=table_schema,
create_disposition=beam.io.BigQueryDisposition.CREATE_IF_NEEDED,
write_disposition=beam.io.BigQueryDisposition.WRITE_TRUNCATE))
logging.getLogger().setLevel(logging.INFO)
logging.info("Building pipeline ...")
p.run()
def run():
# Command line arguments
parser = argparse.ArgumentParser(
description='Load from Json into BigQuery')
parser.add_argument('--project',
required=True,
help='Specify Google Cloud project')
parser.add_argument('--region',
required=True,
help='Specify Google Cloud region')
parser.add_argument('--staging_location',
required=True,
help='Specify Cloud Storage bucket for staging')
parser.add_argument('--temp_location',
required=True,
help='Specify Cloud Storage bucket for temp')
parser.add_argument('--runner',
required=True,
help='Specify Apache Beam Runner')
parser.add_argument('--input_path',
required=True,
help='Path to events.json')
parser.add_argument('--table_name',
required=True,
help='BigQuery table name')
opts = parser.parse_args()
# Setting up the Beam pipeline options
options = PipelineOptions(save_main_session=True)
options.view_as(GoogleCloudOptions).project = opts.project
options.view_as(GoogleCloudOptions).region = opts.region
options.view_as(
GoogleCloudOptions).staging_location = opts.staging_location
options.view_as(GoogleCloudOptions).temp_location = opts.temp_location
options.view_as(GoogleCloudOptions).job_name = '{0}{1}'.format(
'batch-minute-traffic-pipeline-', time.time_ns())
options.view_as(StandardOptions).runner = opts.runner
input_path = opts.input_path
table_name = opts.table_name
# Table schema for BigQuery
table_schema = {
"fields": [
{
"name": "page_views",
"type": "INTEGER"
},
{
"name": "timestamp",
"type": "STRING"
},
]
}
# Create the pipeline
p = beam.Pipeline(options=options)
(p | 'ReadFromGCS' >> beam.io.ReadFromText(input_path)
| 'ParseJson' >> beam.Map(parse_json).with_output_types(CommonLog)
| 'AddEventTimestamp' >> beam.Map(add_timestamp)
| "WindowByMinute" >> beam.WindowInto(beam.window.FixedWindows(60))
| "CountPerMinute" >> beam.CombineGlobally(
CountCombineFn()).without_defaults()
| "AddWindowTimestamp" >> beam.ParDo(GetTimestampFn())
| 'WriteToBQ' >> beam.io.WriteToBigQuery(
table_name,
schema=table_schema,
create_disposition=beam.io.BigQueryDisposition.CREATE_IF_NEEDED,
write_disposition=beam.io.BigQueryDisposition.WRITE_TRUNCATE))
logging.getLogger().setLevel(logging.INFO)
logging.info("Building pipeline ...")
p.run()
def run():
# Command line arguments
parser = argparse.ArgumentParser(
description='Load from PubSub into BigQuery')
parser.add_argument('--project',
required=True,
help='Specify Google Cloud project')
parser.add_argument('--region',
required=True,
help='Specify Google Cloud region')
parser.add_argument('--staging_location',
required=True,
help='Specify Cloud Storage bucket for staging')
parser.add_argument('--temp_location',
required=True,
help='Specify Cloud Storage bucket for temp')
parser.add_argument('--accum_mode',
required=True,
help='Accumulation mode for pipeline')
opts, pipeline_args = parser.parse_known_args()
options = PipelineOptions(pipeline_args, save_main_session=True)
options.view_as(
GoogleCloudOptions).job_name = f"{opts.accum_mode}-{time.time_ns()}"
options.view_as(GoogleCloudOptions).project = opts.project
options.view_as(GoogleCloudOptions).region = opts.region
options.view_as(
GoogleCloudOptions).staging_location = opts.staging_location
options.view_as(GoogleCloudOptions).temp_location = opts.temp_location
options.view_as(StandardOptions).runner = 'DataflowRunner'
table_schema = {
"fields": [
{
"name": "taxi_events",
"type": "INTEGER"
},
{
"name": "timestamp",
"type": "STRING"
},
]
}
input_topic = "projects/pubsub-public-data/topics/taxirides-realtime"
output_table = f"{opts.project}:dataflow_demos.{opts.accum_mode}"
if opts.accum_mode == 'accumulating':
accum_mode = beam.transforms.trigger.AccumulationMode.ACCUMULATING
elif opts.accum_mode == 'discarding':
accum_mode = beam.transforms.trigger.AccumulationMode.DISCARDING
else:
raise ValueError(
'Invalid accumulation mode value. Use \'accumulating\' or \'discarding\' '
)
p = beam.Pipeline(options=options)
(p | 'ReadFromPubSub' >> beam.io.ReadFromPubSub(input_topic)
| 'ParseJson' >> beam.Map(parse_json).with_output_types(TaxiRide)
| 'WindowByMinute' >> beam.WindowInto(
beam.window.FixedWindows(60),
trigger=AfterWatermark(early=AfterProcessingTime(10)),
accumulation_mode=accum_mode)
| "CountPerMinute" >> beam.CombineGlobally(
CountCombineFn()).without_defaults()
| "AddWindowTimestamp" >> beam.ParDo(GetTimestampFn())
| 'WriteAggToBQ' >> beam.io.WriteToBigQuery(
output_table,
schema=table_schema,
create_disposition=beam.io.BigQueryDisposition.CREATE_IF_NEEDED,
write_disposition=beam.io.BigQueryDisposition.WRITE_APPEND))
logging.getLogger().setLevel(logging.INFO)
logging.info("Building pipeline ...")
p.run()
def run():
# Command line arguments
parser = argparse.ArgumentParser(description='Load from Json from Pub/Sub into BigQuery')
parser.add_argument('--project',required=True, help='Specify Google Cloud project')
parser.add_argument('--region', required=True, help='Specify Google Cloud region')
parser.add_argument('--staging_location', required=True, help='Specify Cloud Storage bucket for staging')
parser.add_argument('--temp_location', required=True, help='Specify Cloud Storage bucket for temp')
parser.add_argument('--runner', required=True, help='Specify Apache Beam Runner')
parser.add_argument('--input_topic', required=True, help='Input Pub/Sub Topic')
parser.add_argument('--agg_table_name', required=True, help='BigQuery table name for aggregate results')
parser.add_argument('--raw_table_name', required=True, help='BigQuery table name for raw inputs')
parser.add_argument('--window_duration', required=True, help='Window duration')
opts = parser.parse_args()
# Setting up the Beam pipeline options
options = PipelineOptions(save_main_session=True, streaming=True)
options.view_as(GoogleCloudOptions).project = opts.project
options.view_as(GoogleCloudOptions).region = opts.region
options.view_as(GoogleCloudOptions).staging_location = opts.staging_location
options.view_as(GoogleCloudOptions).temp_location = opts.temp_location
options.view_as(GoogleCloudOptions).job_name = '{0}{1}'.format('streaming-minute-traffic-pipeline-',time.time_ns())
options.view_as(StandardOptions).runner = opts.runner
input_topic = opts.input_topic
raw_table_name = opts.raw_table_name
agg_table_name = opts.agg_table_name
window_duration = opts.window_duration
# Table schema for BigQuery
agg_table_schema = {
"fields": [
{
"name": "page_views",
"type": "INTEGER"
},
{
"name": "timestamp",
"type": "STRING"
},
]
}
raw_table_schema = {
"fields": [
{
"name": "ip",
"type": "STRING"
},
{
"name": "user_id",
"type": "STRING"
},
{
"name": "user_agent",
"type": "STRING"
},
{
"name": "lat",
"type": "FLOAT",
"mode": "NULLABLE"
},
{
"name": "lng",
"type": "FLOAT",
"mode": "NULLABLE"
},
{
"name": "event_timestamp",
"type": "STRING"
},
{
"name": "processing_timestamp",
"type": "STRING"
},
{
"name": "http_request",
"type": "STRING"
},
{
"name": "http_response",
"type": "INTEGER"
},
{
"name": "num_bytes",
"type": "INTEGER"
}
]
}
# Create the pipeline
p = beam.Pipeline(options=options)
parsed_msgs = (p | 'ReadFromPubSub' >> beam.io.ReadFromPubSub(input_topic)
| 'ParseJson' >> beam.Map(parse_json).with_output_types(CommonLog))
(parsed_msgs
| "AddProcessingTimestamp" >> beam.Map(add_processing_timestamp)
| 'WriteRawToBQ' >> beam.io.WriteToBigQuery(
raw_table_name,
schema=raw_table_schema,
create_disposition=beam.io.BigQueryDisposition.CREATE_IF_NEEDED,
write_disposition=beam.io.BigQueryDisposition.WRITE_APPEND
)
)
(parsed_msgs
| "WindowByMinute" >> beam.WindowInto(beam.window.FixedWindows(60))
| "CountPerMinute" >> beam.CombineGlobally(CountCombineFn()).without_defaults()
| "AddWindowTimestamp" >> beam.ParDo(GetTimestampFn())
| 'WriteAggToBQ' >> beam.io.WriteToBigQuery(
agg_table_name,
schema=agg_table_schema,
create_disposition=beam.io.BigQueryDisposition.CREATE_IF_NEEDED,
write_disposition=beam.io.BigQueryDisposition.WRITE_APPEND
)
)
logging.getLogger().setLevel(logging.INFO)
logging.info("Building pipeline ...")
p.run().wait_until_finish()
def _pardo_group_into_batches_with_multi_bags(
input_coder, batch_size, max_buffering_duration_secs, clock=time.time):
ELEMENT_STATE_0 = BagStateSpec('values0', input_coder)
ELEMENT_STATE_1 = BagStateSpec('values1', input_coder)
ELEMENT_STATE_2 = BagStateSpec('values2', input_coder)
ELEMENT_STATE_3 = BagStateSpec('values3', input_coder)
COUNT_STATE = CombiningValueStateSpec('count', input_coder, CountCombineFn())
WINDOW_TIMER = TimerSpec('window_end', TimeDomain.WATERMARK)
BUFFERING_TIMER = TimerSpec('buffering_end', TimeDomain.REAL_TIME)
class _GroupIntoBatchesDoFnWithMultiBags(DoFn):
def process(
self,
element,
window=DoFn.WindowParam,
element_state_0=DoFn.StateParam(ELEMENT_STATE_0),
element_state_1=DoFn.StateParam(ELEMENT_STATE_1),
element_state_2=DoFn.StateParam(ELEMENT_STATE_2),
element_state_3=DoFn.StateParam(ELEMENT_STATE_3),
count_state=DoFn.StateParam(COUNT_STATE),
window_timer=DoFn.TimerParam(WINDOW_TIMER),
buffering_timer=DoFn.TimerParam(BUFFERING_TIMER)):
# Allowed lateness not supported in Python SDK
# https://beam.apache.org/documentation/programming-guide/#watermarks-and-late-data
window_timer.set(window.end)
count_state.add(1)
count = count_state.read()
element_states = [element_state_0, element_state_1, element_state_2, element_state_3]
element_states[count % 4].add(element)
if count == 1 and max_buffering_duration_secs > 0:
# This is the first element in batch. Start counting buffering time if a
# limit was set.
buffering_timer.set(clock() + max_buffering_duration_secs)
if count >= batch_size:
return self.flush_batch(element_states, count_state, buffering_timer)
@on_timer(WINDOW_TIMER)
def on_window_timer(
self,
element_state_0=DoFn.StateParam(ELEMENT_STATE_0),
element_state_1=DoFn.StateParam(ELEMENT_STATE_1),
element_state_2=DoFn.StateParam(ELEMENT_STATE_2),
element_state_3=DoFn.StateParam(ELEMENT_STATE_3),
count_state=DoFn.StateParam(COUNT_STATE),
buffering_timer=DoFn.TimerParam(BUFFERING_TIMER)):
element_states = [element_state_0, element_state_1, element_state_2, element_state_3]
return self.flush_batch(element_states, count_state, buffering_timer)
@on_timer(BUFFERING_TIMER)
def on_buffering_timer(
self,
element_state_0=DoFn.StateParam(ELEMENT_STATE_0),
element_state_1=DoFn.StateParam(ELEMENT_STATE_1),
element_state_2=DoFn.StateParam(ELEMENT_STATE_2),
element_state_3=DoFn.StateParam(ELEMENT_STATE_3),
count_state=DoFn.StateParam(COUNT_STATE),
buffering_timer=DoFn.TimerParam(BUFFERING_TIMER)):
element_states = [element_state_0, element_state_1, element_state_2, element_state_3]
return self.flush_batch(element_states, count_state, buffering_timer)
def flush_batch(self, element_states, count_state, buffering_timer):
batch_values = []
for element_state in element_states:
for k, v in element_state.read():
key = k
batch_values.append(v)
element_state.clear()
count_state.clear()
buffering_timer.clear()
if not batch_values:
return
yield key, batch_values
return _GroupIntoBatchesDoFnWithMultiBags()
