def TopicType(value):
"""Converts [NAME:]TOPIC format string to TopicInfo object.
Args:
value: a topic string, provided in [NAME:]TOPIC format.
Returns:
TopicInfo message (name=NAME, topic=TOPIC)
Raises:
TopicTypeError: when topic format is invalid
"""
topic_name_re = re.compile(r'^[a-zA-Z0-9-_]*$')
# topic should not start with '$' and should not contain any of '+#,:'
topic_re = re.compile(r'^[^$+#,:][^+#,:]*$')
topic_parts = value.split(':')
if len(topic_parts) > 2:
raise TopicTypeError()
if len(topic_parts) == 2:
name, topic = topic_parts
if len(topic_parts) == 1:
name = ''
topic = topic_parts[0]
if not topic_name_re.match(name) or not topic_re.match(topic):
raise TopicTypeError()
messages = util.GetMessagesModule()
return messages.TopicInfo(id=name, topic=topic)
def ProcessModelHook(ref, args, req):
"""Analyzes given model, and converts model if necessary.
Args:
ref: A resource ref to the parsed Edge ML Model resource,
unused in this hook
args: The parsed args namespace from CLI
req: Created request for the API call
Returns:
req, with new model URI, input/out tensor information, accelerator type
if applicable.
Raises:
InvalidFrameworkException: if framework is FRAMEWORK_UNSPECIFIED.
This should not happen.
"""
del ref # Unused.
edgeml_messages = edgeml_util.GetMessagesModule()
model_types = edgeml_messages.AnalyzeModelResponse.ModelTypeValueValuesEnum
tf_model_types = (
model_types.TENSORFLOW_LITE,
model_types.TENSORFLOW_LITE_EDGE_TPU_OPTIMIZED,
model_types.TENSORFLOW_SAVED_MODEL,
)
edge_messages = edge_util.GetMessagesModule()
framework_types = edge_messages.MlModel.FrameworkValueValuesEnum
patch_req_type = (
edge_messages.EdgeProjectsLocationsRegistriesDevicesMlModelsPatchRequest)
analyze_result = edgeml.EdgeMlClient().Analyze(req.mlModel.modelUri)
if req.mlModel.framework == framework_types.TFLITE:
if analyze_result.modelType not in tf_model_types:
raise exceptions.InvalidArgumentException(
'--framework', 'tflite provided for non-Tensorflow model.')
_ProcessTensorflowModel(req.mlModel, args, analyze_result)
if isinstance(req, patch_req_type):
# updateMask should have some pre-filled values.
update_fields = set(req.updateMask.split(','))
update_fields.update({
'modelUri', 'acceleratorType', 'inputTensors', 'outputTensors'
})
req.updateMask = ','.join(sorted(update_fields))
# Try to deploy as a scikit-learn model if it's not a TF model.
elif req.mlModel.framework == framework_types.SCIKIT_LEARN:
if analyze_result.modelType in tf_model_types:
raise exceptions.InvalidArgumentException(
'--framework', 'scikit-learn provided for Tensorflow model.')
else:
raise InvalidFrameworkException() # FRAMEWORK_UNSPECIFIED is not allowed.
return req
def VolumeBindingType(value):
"""Verifies volume binding flag format, and returns VolumeBinding messages.
Args:
value: a volume binding string parsed by ArgList from CLI flag.
Returns:
VolumeBinding message
Raises:
VolumeBindingTypeError: when the format is invalid.
"""
binding_parts = value.split(':')
if len(binding_parts) > 3:
raise VolumeBindingTypeError()
if len(binding_parts) == 3:
source, destination, read_only = binding_parts
if len(binding_parts) == 2:
if binding_parts[1] in ['ro', 'rw']:
source = destination = binding_parts[0]
read_only = binding_parts[1]
else:
source, destination = binding_parts
read_only = 'rw'
if len(binding_parts) == 1:
source = destination = binding_parts[0]
read_only = 'rw'
if not destination.startswith('/'):
raise VolumeBindingTypeError(
'DESTINATION {0} is not a valid absolute path.'.format(
destination))
if source and not source.startswith('/'):
raise VolumeBindingTypeError(
'SOURCE {0} is not a valid absolute path.'.format(source))
if read_only not in ['ro', 'rw']:
raise VolumeBindingTypeError(
'The last value should be "ro" for read-only volume, and'
' "rw" for writable volume.')
messages = util.GetMessagesModule()
return messages.VolumeBinding(
source=source or destination,
destination=destination,
# read_only is one of 'ro' or 'rw'
readOnly=(read_only == 'ro'))
def _ConvertTensorRef(edgeml_tensor_refs):
"""Converts edgeml.TensorRef[] to edge.TensorInfo[]."""
edge_messages = edge_util.GetMessagesModule()
inference_type = edge_messages.TensorInfo.InferenceTypeValueValuesEnum
edge_tensor_infos = []
for tensor_ref in edgeml_tensor_refs:
tensor_info = edge_messages.TensorInfo(
index=tensor_ref.index,
dimensions=tensor_ref.tensorInfo.dimensions,
tensorName=tensor_ref.tensorInfo.tensorName,
inferenceType=inference_type(tensor_ref.tensorInfo.inferenceType.name))
edge_tensor_infos.append(tensor_info)
return edge_tensor_infos
def DeviceBindingType(value):
"""Verifies device binding flag format, and returns device binding list.
Args:
value: a device binding string parsed by ArgList from CLI flag.
Returns:
DeviceBinding message
Raises:
DeviceBindingTypeError: when the format is invalid.
"""
cgroup_perms_re = re.compile(r'^r?w?m?$')
binding_parts = value.split(':')
if len(binding_parts) > 3:
raise DeviceBindingTypeError()
if len(binding_parts) == 3:
source, destination, cgroup_permissions = binding_parts
if len(binding_parts) == 2:
if cgroup_perms_re.match(binding_parts[1]):
source = destination = binding_parts[0]
cgroup_permissions = binding_parts[1]
else:
source, destination = binding_parts
cgroup_permissions = 'rwm'
if len(binding_parts) == 1:
source = destination = binding_parts[0]
cgroup_permissions = 'rwm'
if not source.startswith('/'):
raise DeviceBindingTypeError(
'SOURCE {0} is not a valid absolute path.'.format(source))
if destination and not destination.startswith('/'):
raise DeviceBindingTypeError(
'DESTINATION {0} is not a valid absolute path.'.format(
destination))
if not cgroup_perms_re.match(cgroup_permissions):
raise DeviceBindingTypeError(
'CGROUP_PERMS should be a combination of the following flags'
' in order: "r/w/m."')
messages = util.GetMessagesModule()
return messages.DeviceBinding(source=source,
destination=destination or source,
cgroupPermissions=cgroup_permissions)
def _ProcessTensorflowModel(model, args, analyze_result):
"""Processes Tensorflow (Lite) model according to analyze result.
Args:
model: edge.MlModel message from request
args: The parsed args namespace from CLI
analyze_result: edgeml.AnalyzeModelResponse from Analyze method call.
Raises:
UncompilableModelException: if given model cannot be optimized for Edge TPU.
"""
client = edgeml.EdgeMlClient()
edgeml_messages = edgeml_util.GetMessagesModule()
edge_messages = edge_util.GetMessagesModule()
model_types = edgeml_messages.AnalyzeModelResponse.ModelTypeValueValuesEnum
accelerator_types = edge_messages.MlModel.AcceleratorTypeValueValuesEnum
model_type = analyze_result.modelType
model_signature = analyze_result.modelSignature
edgetpu_compiliability = analyze_result.edgeTpuCompilability
# Convert method converts TF SavedModel to TF Lite model.
if model_type == model_types.TENSORFLOW_SAVED_MODEL:
convert_result, model.modelUri = client.Convert(model.modelUri)
model_signature = convert_result.modelSignature
edgetpu_compiliability = convert_result.edgeTpuCompilability
model_type = model_types.TENSORFLOW_LITE
if model_type == model_types.TENSORFLOW_LITE:
# Always use accelerator value from command line, and ignore previous
# acceleratorType of the model.
if args.accelerator == 'tpu':
if edgetpu_compiliability.uncompilableReason:
raise UncompilableModelException(
edgetpu_compiliability.uncompilableReason)
compile_result, model.modelUri = client.Compile(model.modelUri)
model_signature = compile_result.modelSignature
model_type = model_types.TENSORFLOW_LITE_EDGE_TPU_OPTIMIZED
if model_type == model_types.TENSORFLOW_LITE_EDGE_TPU_OPTIMIZED:
if args.IsSpecified('accelerator') and args.accelerator != 'tpu':
raise exceptions.InvalidArgumentException(
'--accelerator',
'TPU should be provided for Edge TPU optimized model.')
if not args.IsSpecified('accelerator'):
log.info('Setting accelerator to TPU for Edge TPU model.')
model.acceleratorType = accelerator_types.TPU
_FillModelSignature(model, model_signature)
def ParseSamplingInfo(path):
messages = edge_util.GetMessagesModule()
sampling_info = cloudbuild_util.LoadMessageFromPath(
path, messages.MlSamplingInfo, 'Edge ML sampling info')
return sampling_info
def ParseTopicBridgingTable(unused_ref, args, req):
messages = util.GetMessagesModule()
parsed_table = cloudbuild_util.LoadMessageFromPath(
args.rule_file, messages.TopicBridgingTable, 'topic bridging table')
req.rules = parsed_table.rules
return req
