def expand_gbk(stages):
"""Transforms each GBK into a write followed by a read.
"""
good_coder_urns = set(beam.coders.Coder._known_urns.keys()) - set([
urns.PICKLED_CODER])
coders = pipeline_components.coders
for coder_id, coder_proto in coders.items():
if coder_proto.spec.spec.urn == urns.BYTES_CODER:
bytes_coder_id = coder_id
break
else:
bytes_coder_id = unique_name(coders, 'bytes_coder')
pipeline_components.coders[bytes_coder_id].CopyFrom(
beam.coders.BytesCoder().to_runner_api(None))
coder_substitutions = {}
def wrap_unknown_coders(coder_id, with_bytes):
if (coder_id, with_bytes) not in coder_substitutions:
wrapped_coder_id = None
coder_proto = coders[coder_id]
if coder_proto.spec.spec.urn == urns.LENGTH_PREFIX_CODER:
coder_substitutions[coder_id, with_bytes] = (
bytes_coder_id if with_bytes else coder_id)
elif coder_proto.spec.spec.urn in good_coder_urns:
wrapped_components = [wrap_unknown_coders(c, with_bytes)
for c in coder_proto.component_coder_ids]
if wrapped_components == list(coder_proto.component_coder_ids):
# Use as is.
coder_substitutions[coder_id, with_bytes] = coder_id
else:
wrapped_coder_id = unique_name(
coders,
coder_id + ("_bytes" if with_bytes else "_len_prefix"))
coders[wrapped_coder_id].CopyFrom(coder_proto)
coders[wrapped_coder_id].component_coder_ids[:] = [
wrap_unknown_coders(c, with_bytes)
for c in coder_proto.component_coder_ids]
coder_substitutions[coder_id, with_bytes] = wrapped_coder_id
else:
# Not a known coder.
if with_bytes:
coder_substitutions[coder_id, with_bytes] = bytes_coder_id
else:
wrapped_coder_id = unique_name(coders, coder_id + "_len_prefix")
len_prefix_coder_proto = beam_runner_api_pb2.Coder(
spec=beam_runner_api_pb2.SdkFunctionSpec(
spec=beam_runner_api_pb2.FunctionSpec(
urn=urns.LENGTH_PREFIX_CODER)),
component_coder_ids=[coder_id])
coders[wrapped_coder_id].CopyFrom(len_prefix_coder_proto)
coder_substitutions[coder_id, with_bytes] = wrapped_coder_id
# This operation is idempotent.
if wrapped_coder_id:
coder_substitutions[wrapped_coder_id, with_bytes] = wrapped_coder_id
return coder_substitutions[coder_id, with_bytes]
def fix_pcoll_coder(pcoll):
new_coder_id = wrap_unknown_coders(pcoll.coder_id, False)
safe_coders[new_coder_id] = wrap_unknown_coders(pcoll.coder_id, True)
pcoll.coder_id = new_coder_id
for stage in stages:
assert len(stage.transforms) == 1
transform = stage.transforms[0]
if transform.spec.urn == urns.GROUP_BY_KEY_TRANSFORM:
for pcoll_id in transform.inputs.values():
fix_pcoll_coder(pipeline_components.pcollections[pcoll_id])
for pcoll_id in transform.outputs.values():
fix_pcoll_coder(pipeline_components.pcollections[pcoll_id])
# This is used later to correlate the read and write.
param = str("group:%s" % stage.name)
gbk_write = Stage(
transform.unique_name + '/Write',
[beam_runner_api_pb2.PTransform(
unique_name=transform.unique_name + '/Write',
inputs=transform.inputs,
spec=beam_runner_api_pb2.FunctionSpec(
urn=bundle_processor.DATA_OUTPUT_URN,
any_param=proto_utils.pack_Any(
wrappers_pb2.BytesValue(value=param)),
payload=param))],
downstream_side_inputs=frozenset(),
must_follow=stage.must_follow)
yield gbk_write
yield Stage(
transform.unique_name + '/Read',
[beam_runner_api_pb2.PTransform(
unique_name=transform.unique_name + '/Read',
outputs=transform.outputs,
spec=beam_runner_api_pb2.FunctionSpec(
urn=bundle_processor.DATA_INPUT_URN,
any_param=proto_utils.pack_Any(
wrappers_pb2.BytesValue(value=param)),
payload=param))],
downstream_side_inputs=frozenset(),
must_follow=union(frozenset([gbk_write]), stage.must_follow))
else:
yield stage
def sink_flattens(stages):
"""Sink flattens and remove them from the graph.
A flatten that cannot be sunk/fused away becomes multiple writes (to the
same logical sink) followed by a read.
"""
# TODO(robertwb): Actually attempt to sink rather than always materialize.
# TODO(robertwb): Possibly fuse this into one of the stages.
pcollections = pipeline_components.pcollections
for stage in stages:
assert len(stage.transforms) == 1
transform = stage.transforms[0]
if transform.spec.urn == urns.FLATTEN_TRANSFORM:
# This is used later to correlate the read and writes.
param = str("materialize:%s" % transform.unique_name)
output_pcoll_id, = transform.outputs.values()
output_coder_id = pcollections[output_pcoll_id].coder_id
flatten_writes = []
for local_in, pcoll_in in transform.inputs.items():
if pcollections[pcoll_in].coder_id != output_coder_id:
# Flatten inputs must all be written with the same coder as is
# used to read them.
pcollections[pcoll_in].coder_id = output_coder_id
transcoded_pcollection = (
transform.unique_name + '/Transcode/' + local_in + '/out')
yield Stage(
transform.unique_name + '/Transcode/' + local_in,
[beam_runner_api_pb2.PTransform(
unique_name=
transform.unique_name + '/Transcode/' + local_in,
inputs={local_in: pcoll_in},
outputs={'out': transcoded_pcollection},
spec=beam_runner_api_pb2.FunctionSpec(
urn=bundle_processor.IDENTITY_DOFN_URN))],
downstream_side_inputs=frozenset(),
must_follow=stage.must_follow)
pcollections[transcoded_pcollection].CopyFrom(
pcollections[pcoll_in])
pcollections[transcoded_pcollection].coder_id = output_coder_id
else:
transcoded_pcollection = pcoll_in
flatten_write = Stage(
transform.unique_name + '/Write/' + local_in,
[beam_runner_api_pb2.PTransform(
unique_name=transform.unique_name + '/Write/' + local_in,
inputs={local_in: transcoded_pcollection},
spec=beam_runner_api_pb2.FunctionSpec(
urn=bundle_processor.DATA_OUTPUT_URN,
any_param=proto_utils.pack_Any(
wrappers_pb2.BytesValue(
value=param)),
payload=param))],
downstream_side_inputs=frozenset(),
must_follow=stage.must_follow)
flatten_writes.append(flatten_write)
yield flatten_write
yield Stage(
transform.unique_name + '/Read',
[beam_runner_api_pb2.PTransform(
unique_name=transform.unique_name + '/Read',
outputs=transform.outputs,
spec=beam_runner_api_pb2.FunctionSpec(
urn=bundle_processor.DATA_INPUT_URN,
any_param=proto_utils.pack_Any(
wrappers_pb2.BytesValue(
value=param)),
payload=param))],
downstream_side_inputs=frozenset(),
must_follow=union(frozenset(flatten_writes), stage.must_follow))
else:
yield stage
def to_runner_api_parameter(self, context):
return (_StreamingGroupAlsoByWindow.urn,
wrappers_pb2.BytesValue(
value=context.windowing_strategies.get_id(self.windowing)))
def to_runner_api_parameter(self, context):
return (python_urns.PICKLED_TRANSFORM,
wrappers_pb2.BytesValue(value=pickler.dumps(self)))
def to_runner_api_parameter(self, context):
return (urns.PICKLED_WINDOW_FN,
wrappers_pb2.BytesValue(value=pickler.dumps(self)))
def _map_task_to_protos(self, map_task, data_operation_spec):
input_data = {}
side_input_data = {}
runner_sinks = {}
context = pipeline_context.PipelineContext()
transform_protos = {}
used_pcollections = {}
def uniquify(*names):
# An injective mapping from string* to string.
return ':'.join("%s:%d" % (name, len(name)) for name in names)
def pcollection_id(op_ix, out_ix):
if (op_ix, out_ix) not in used_pcollections:
used_pcollections[op_ix,
out_ix] = uniquify(map_task[op_ix][0], 'out',
str(out_ix))
return used_pcollections[op_ix, out_ix]
def get_inputs(op):
if hasattr(op, 'inputs'):
inputs = op.inputs
elif hasattr(op, 'input'):
inputs = [op.input]
else:
inputs = []
return {
'in%s' % ix: pcollection_id(*input)
for ix, input in enumerate(inputs)
}
def get_outputs(op_ix):
op = map_task[op_ix][1]
return {
tag: pcollection_id(op_ix, out_ix)
for out_ix, tag in enumerate(
getattr(op, 'output_tags', ['out']))
}
def only_element(iterable):
element, = iterable
return element
for op_ix, (stage_name, operation) in enumerate(map_task):
transform_id = uniquify(stage_name)
if isinstance(operation, operation_specs.WorkerInMemoryWrite):
# Write this data back to the runner.
target_name = only_element(get_inputs(operation).keys())
runner_sinks[(transform_id, target_name)] = operation
transform_spec = beam_runner_api_pb2.FunctionSpec(
urn=sdk_worker.DATA_OUTPUT_URN,
parameter=proto_utils.pack_Any(data_operation_spec))
elif isinstance(operation, operation_specs.WorkerRead):
# A Read from an in-memory source is done over the data plane.
if (isinstance(operation.source.source,
maptask_executor_runner.InMemorySource)
and isinstance(
operation.source.source.default_output_coder(),
WindowedValueCoder)):
target_name = only_element(get_outputs(op_ix).keys())
input_data[(
transform_id, target_name)] = self._reencode_elements(
operation.source.source.read(None),
operation.source.source.default_output_coder())
transform_spec = beam_runner_api_pb2.FunctionSpec(
urn=sdk_worker.DATA_INPUT_URN,
parameter=proto_utils.pack_Any(data_operation_spec))
else:
# Otherwise serialize the source and execute it there.
# TODO: Use SDFs with an initial impulse.
# The Dataflow runner harness strips the base64 encoding. do the same
# here until we get the same thing back that we sent in.
transform_spec = beam_runner_api_pb2.FunctionSpec(
urn=sdk_worker.PYTHON_SOURCE_URN,
parameter=proto_utils.pack_Any(
wrappers_pb2.BytesValue(value=base64.b64decode(
pickler.dumps(operation.source.source)))))
elif isinstance(operation, operation_specs.WorkerDoFn):
# Record the contents of each side input for access via the state api.
side_input_extras = []
for si in operation.side_inputs:
assert isinstance(si.source, iobase.BoundedSource)
element_coder = si.source.default_output_coder()
# TODO(robertwb): Actually flesh out the ViewFn API.
side_input_extras.append((si.tag, element_coder))
side_input_data[sdk_worker.side_input_tag(
transform_id, si.tag)] = (self._reencode_elements(
si.source.read(
si.source.get_range_tracker(None, None)),
element_coder))
augmented_serialized_fn = pickler.dumps(
(operation.serialized_fn, side_input_extras))
transform_spec = beam_runner_api_pb2.FunctionSpec(
urn=sdk_worker.PYTHON_DOFN_URN,
parameter=proto_utils.pack_Any(
wrappers_pb2.BytesValue(
value=augmented_serialized_fn)))
elif isinstance(operation, operation_specs.WorkerFlatten):
# Flatten is nice and simple.
transform_spec = beam_runner_api_pb2.FunctionSpec(
urn=sdk_worker.IDENTITY_DOFN_URN)
else:
raise NotImplementedError(operation)
transform_protos[transform_id] = beam_runner_api_pb2.PTransform(
unique_name=stage_name,
spec=transform_spec,
inputs=get_inputs(operation),
outputs=get_outputs(op_ix))
pcollection_protos = {
name: beam_runner_api_pb2.PCollection(
unique_name=name,
coder_id=context.coders.get_id(
map_task[op_id][1].output_coders[out_id]))
for (op_id, out_id), name in used_pcollections.items()
}
# Must follow creation of pcollection_protos to capture used coders.
context_proto = context.to_runner_api()
process_bundle_descriptor = beam_fn_api_pb2.ProcessBundleDescriptor(
id=self._next_uid(),
transforms=transform_protos,
pcollections=pcollection_protos,
codersyyy=dict(context_proto.coders.items()),
windowing_strategies=dict(
context_proto.windowing_strategies.items()),
environments=dict(context_proto.environments.items()))
return input_data, side_input_data, runner_sinks, process_bundle_descriptor
def to_runner_api_parameter(self, context):
return (python_urns.PICKLED_CODER,
wrappers_pb2.BytesValue(value=serialize_coder(self)), ())
def unpack_function_spec_data(function_spec):
"""Returns unpacked data from function spec proto."""
data = wrappers_pb2.BytesValue()
function_spec.data.Unpack(data)
return data.value
def greedily_fuse(stages):
"""Places transforms sharing an edge in the same stage, whenever possible.
"""
producers_by_pcoll = {}
consumers_by_pcoll = collections.defaultdict(list)
# Used to always reference the correct stage as the producer and
# consumer maps are not updated when stages are fused away.
replacements = {}
def replacement(s):
old_ss = []
while s in replacements:
old_ss.append(s)
s = replacements[s]
for old_s in old_ss[:-1]:
replacements[old_s] = s
return s
def fuse(producer, consumer):
fused = producer.fuse(consumer)
replacements[producer] = fused
replacements[consumer] = fused
# First record the producers and consumers of each PCollection.
for stage in stages:
for transform in stage.transforms:
for input in transform.inputs.values():
consumers_by_pcoll[input].append(stage)
for output in transform.outputs.values():
producers_by_pcoll[output] = stage
logging.debug('consumers\n%s', consumers_by_pcoll)
logging.debug('producers\n%s', producers_by_pcoll)
# Now try to fuse away all pcollections.
for pcoll, producer in producers_by_pcoll.items():
pcoll_as_param = proto_utils.pack_Any(
wrappers_pb2.BytesValue(value=str("materialize:%s" %
pcoll)))
write_pcoll = None
for consumer in consumers_by_pcoll[pcoll]:
producer = replacement(producer)
consumer = replacement(consumer)
# Update consumer.must_follow set, as it's used in can_fuse.
consumer.must_follow = set(
replacement(s) for s in consumer.must_follow)
if producer.can_fuse(consumer):
fuse(producer, consumer)
else:
# If we can't fuse, do a read + write.
if write_pcoll is None:
write_pcoll = Stage(pcoll + '/Write', [
beam_runner_api_pb2.PTransform(
unique_name=pcoll + '/Write',
inputs={'in': pcoll},
spec=beam_runner_api_pb2.FunctionSpec(
urn=bundle_processor.DATA_OUTPUT_URN,
parameter=pcoll_as_param))
])
fuse(producer, write_pcoll)
if consumer.has_as_main_input(pcoll):
read_pcoll = Stage(pcoll + '/Read', [
beam_runner_api_pb2.PTransform(
unique_name=pcoll + '/Read',
outputs={'out': pcoll},
spec=beam_runner_api_pb2.FunctionSpec(
urn=bundle_processor.DATA_INPUT_URN,
parameter=pcoll_as_param))
],
must_follow={write_pcoll})
fuse(read_pcoll, consumer)
# Everything that was originally a stage or a replacement, but wasn't
# replaced, should be in the final graph.
final_stages = frozenset(stages).union(
replacements.values()).difference(replacements.keys())
for stage in final_stages:
# Update all references to their final values before throwing
# the replacement data away.
stage.must_follow = frozenset(
replacement(s) for s in stage.must_follow)
# Two reads of the same stage may have been fused. This is unneeded.
stage.deduplicate_read()
return final_stages
