def _event_parse(event_str, latest_file_name):
"""
Transform `Event` data to tensor_event and update it to EventsData.
This method is static to avoid sending unnecessary objects to other processes.
Args:
event_str (str): Message event string in summary proto, data read from file handler.
latest_file_name (str): Latest file name.
"""
plugins = {
'scalar_value': PluginNameEnum.SCALAR,
'image': PluginNameEnum.IMAGE,
'histogram': PluginNameEnum.HISTOGRAM,
'tensor': PluginNameEnum.TENSOR
}
logger.debug("Start to parse event string. Event string len: %s.",
len(event_str))
event = summary_pb2.Event.FromString(event_str)
logger.debug("Deserialize event string completed.")
ret_tensor_events = []
if event.HasField('summary'):
for value in event.summary.value:
for plugin in plugins:
if not value.HasField(plugin):
continue
plugin_name_enum = plugins[plugin]
logger.debug("Processing plugin value: %s.",
plugin_name_enum)
tensor_event_value = _SummaryParser._parse_summary_value(
value, plugin)
if tensor_event_value is None:
continue
tensor_event = TensorEvent(
wall_time=event.wall_time,
step=event.step,
tag='{}/{}'.format(value.tag, plugin_name_enum.value),
plugin_name=plugin_name_enum.value,
value=tensor_event_value,
filename=latest_file_name)
logger.debug(
"Tensor event generated, plugin is %s, tag is %s, step is %s.",
plugin_name_enum, value.tag, event.step)
ret_tensor_events.append(tensor_event)
elif event.HasField('graph_def'):
graph = MSGraph()
graph.build_graph(event.graph_def)
tensor_event = TensorEvent(wall_time=event.wall_time,
step=event.step,
tag=latest_file_name,
plugin_name=PluginNameEnum.GRAPH.value,
value=graph,
filename=latest_file_name)
ret_tensor_events.append(tensor_event)
return ret_tensor_events
def _event_parse(self, event):
"""
Transform `Event` data to tensor_event and update it to EventsData.
Args:
event (Event): Message event in summary proto, data read from file handler.
"""
if event.HasField('summary'):
for value in event.summary.value:
if value.HasField('scalar_value'):
tag = '{}/{}'.format(value.tag,
PluginNameEnum.SCALAR.value)
tensor_event = TensorEvent(
wall_time=event.wall_time,
step=event.step,
tag=tag,
plugin_name=PluginNameEnum.SCALAR.value,
value=value.scalar_value)
self._events_data.add_tensor_event(tensor_event)
if value.HasField('image'):
tag = '{}/{}'.format(value.tag, PluginNameEnum.IMAGE.value)
tensor_event = TensorEvent(
wall_time=event.wall_time,
step=event.step,
tag=tag,
plugin_name=PluginNameEnum.IMAGE.value,
value=value.image)
self._events_data.add_tensor_event(tensor_event)
if event.HasField('graph_def'):
graph_proto = event.graph_def
graph = MSGraph()
graph.build_graph(graph_proto)
tensor_event = TensorEvent(wall_time=event.wall_time,
step=event.step,
tag=self._latest_summary_filename,
plugin_name=PluginNameEnum.GRAPH.value,
value=graph)
try:
graph_tags = self._events_data.list_tags_by_plugin(
PluginNameEnum.GRAPH.value)
except KeyError:
graph_tags = []
summary_tags = self._filter_summary_files(graph_tags)
for tag in summary_tags:
self._events_data.delete_tensor_event(tag)
self._events_data.add_tensor_event(tensor_event)
def _parse_pb_file(self, filename):
"""
Parse pb file and write content to `EventsData`.
Args:
filename (str): The file path of pb file.
"""
file_path = FileHandler.join(self._summary_dir, filename)
logger.info("Start to load graph from pb file, file path: %s.",
file_path)
filehandler = FileHandler(file_path)
model_proto = anf_ir_pb2.ModelProto()
try:
model_proto.ParseFromString(filehandler.read())
except ParseError:
logger.warning(
"The given file is not a valid pb file, file path: %s.",
file_path)
return
graph = MSGraph()
graph.build_graph(model_proto.graph)
tensor_event = TensorEvent(wall_time=FileHandler.file_stat(file_path),
step=0,
tag=filename,
plugin_name=PluginNameEnum.GRAPH.value,
value=graph)
self._events_data.add_tensor_event(tensor_event)
def test_add_tensor_event_success(self):
"""Test add_tensor_event success."""
ev_data = self.get_ev_data()
t_event = TensorEvent(wall_time=1, step=4, tag='new_tag', plugin_name='plugin_name1',
value='value1')
ev_data.add_tensor_event(t_event)
assert 'tag0' not in ev_data._tags
assert ev_data._tags[-1] == 'new_tag'
assert 'tag0' not in ev_data._tags_by_plugin['plugin_name1']
assert 'tag0' not in ev_data._reservoir_by_tag
assert 'new_tag' in ev_data._tags_by_plugin['plugin_name1']
assert ev_data._reservoir_by_tag['new_tag'].samples()[-1] == _Tensor(t_event.wall_time,
t_event.step,
t_event.value)
def _parse_pb_file(summary_dir, filename):
"""
Parse pb file and write content to `EventsData`.
Args:
filename (str): The file path of pb file.
Returns:
TensorEvent, if load pb file and build graph success, will return tensor event, else return None.
"""
file_path = FileHandler.join(summary_dir, filename)
logger.info("Start to load graph from pb file, file path: %s.",
file_path)
filehandler = FileHandler(file_path)
model_proto = anf_ir_pb2.ModelProto()
try:
model_proto.ParseFromString(filehandler.read())
except ParseError:
logger.warning(
"The given file is not a valid pb file, file path: %s.",
file_path)
return None
graph = MSGraph()
try:
graph.build_graph(model_proto.graph)
except Exception as ex:
# Normally, there are no exceptions, and it is only possible for users on the MindSpore side
# to dump other non-default graphs.
logger.error("Build graph failed, file path: %s.", file_path)
logger.exception(ex)
raise UnknownError(str(ex))
tensor_event = TensorEvent(
wall_time=FileHandler.file_stat(file_path).mtime,
step=0,
tag=filename,
plugin_name=PluginNameEnum.GRAPH.value,
value=graph,
filename=filename)
logger.info("Build graph success, file path: %s.", file_path)
return tensor_event
def _event_parse(self, event):
"""
Transform `Event` data to tensor_event and update it to EventsData.
Args:
event (Event): Message event in summary proto, data read from file handler.
"""
if event.HasField('summary'):
for value in event.summary.value:
if value.HasField('scalar_value'):
tag = '{}/{}'.format(value.tag, PluginNameEnum.SCALAR.value)
tensor_event = TensorEvent(wall_time=event.wall_time,
step=event.step,
tag=tag,
plugin_name=PluginNameEnum.SCALAR.value,
value=value.scalar_value,
filename=self._latest_filename)
self._events_data.add_tensor_event(tensor_event)
if value.HasField('image'):
tag = '{}/{}'.format(value.tag, PluginNameEnum.IMAGE.value)
tensor_event = TensorEvent(wall_time=event.wall_time,
step=event.step,
tag=tag,
plugin_name=PluginNameEnum.IMAGE.value,
value=value.image,
filename=self._latest_filename)
self._events_data.add_tensor_event(tensor_event)
if value.HasField('histogram'):
histogram_msg = HistogramContainer(value.histogram)
# Drop steps if original_buckets_count exceeds HistogramContainer.MAX_ORIGINAL_BUCKETS_COUNT
# to avoid time-consuming re-sample process.
if histogram_msg.original_buckets_count > HistogramContainer.MAX_ORIGINAL_BUCKETS_COUNT:
logger.warning('original_buckets_count exceeds HistogramContainer.MAX_ORIGINAL_BUCKETS_COUNT')
else:
tag = '{}/{}'.format(value.tag, PluginNameEnum.HISTOGRAM.value)
tensor_event = TensorEvent(wall_time=event.wall_time,
step=event.step,
tag=tag,
plugin_name=PluginNameEnum.HISTOGRAM.value,
value=histogram_msg,
filename=self._latest_filename)
self._events_data.add_tensor_event(tensor_event)
if event.HasField('graph_def'):
graph_proto = event.graph_def
graph = MSGraph()
graph.build_graph(graph_proto)
tensor_event = TensorEvent(wall_time=event.wall_time,
step=event.step,
tag=self._latest_filename,
plugin_name=PluginNameEnum.GRAPH.value,
value=graph,
filename=self._latest_filename)
try:
graph_tags = self._events_data.list_tags_by_plugin(PluginNameEnum.GRAPH.value)
except KeyError:
graph_tags = []
summary_tags = self.filter_files(graph_tags)
for tag in summary_tags:
self._events_data.delete_tensor_event(tag)
self._events_data.add_tensor_event(tensor_event)
def _event_parse(self, event):
"""
Transform `Event` data to tensor_event and update it to EventsData.
Args:
event (Event): Message event in summary proto, data read from file handler.
"""
plugins = {
'scalar_value': PluginNameEnum.SCALAR,
'image': PluginNameEnum.IMAGE,
'histogram': PluginNameEnum.HISTOGRAM,
}
if event.HasField('summary'):
for value in event.summary.value:
for plugin in plugins:
if not value.HasField(plugin):
continue
plugin_name_enum = plugins[plugin]
tensor_event_value = getattr(value, plugin)
if plugin == 'histogram':
tensor_event_value = HistogramContainer(
tensor_event_value)
# Drop steps if original_buckets_count exceeds HistogramContainer.MAX_ORIGINAL_BUCKETS_COUNT
# to avoid time-consuming re-sample process.
if tensor_event_value.original_buckets_count > HistogramContainer.MAX_ORIGINAL_BUCKETS_COUNT:
logger.warning(
'original_buckets_count exceeds '
'HistogramContainer.MAX_ORIGINAL_BUCKETS_COUNT'
)
continue
tensor_event = TensorEvent(
wall_time=event.wall_time,
step=event.step,
tag='{}/{}'.format(value.tag, plugin_name_enum.value),
plugin_name=plugin_name_enum.value,
value=tensor_event_value,
filename=self._latest_filename)
self._events_data.add_tensor_event(tensor_event)
elif event.HasField('graph_def'):
graph = MSGraph()
graph.build_graph(event.graph_def)
tensor_event = TensorEvent(wall_time=event.wall_time,
step=event.step,
tag=self._latest_filename,
plugin_name=PluginNameEnum.GRAPH.value,
value=graph,
filename=self._latest_filename)
try:
graph_tags = self._events_data.list_tags_by_plugin(
PluginNameEnum.GRAPH.value)
except KeyError:
graph_tags = []
summary_tags = self.filter_files(graph_tags)
for tag in summary_tags:
self._events_data.delete_tensor_event(tag)
self._events_data.add_tensor_event(tensor_event)
def test_add_tensor_event_out_of_order(self):
"""Test add_tensor_event success for out_of_order summaries."""
wall_time = 1
value = '1'
tag = 'tag'
plugin_name = 'scalar'
file1 = 'file1'
ev_data = EventsData()
steps = [i for i in range(2, 10)]
for step in steps:
t_event = TensorEvent(wall_time=1,
step=step,
tag=tag,
plugin_name=plugin_name,
value=value,
filename=file1)
ev_data.add_tensor_event(t_event)
t_event = TensorEvent(wall_time=1,
step=1,
tag=tag,
plugin_name=plugin_name,
value=value,
filename=file1)
ev_data.add_tensor_event(t_event)
# Current steps should be: [1, 2, 3, 4, 5, 6, 7, 8, 9]
assert len(ev_data._reservoir_by_tag[tag].samples()) == len(steps) + 1
file2 = 'file2'
new_steps_1 = [5, 10]
for step in new_steps_1:
t_event = TensorEvent(wall_time=1,
step=step,
tag=tag,
plugin_name=plugin_name,
value=value,
filename=file2)
ev_data.add_tensor_event(t_event)
assert ev_data._reservoir_by_tag[tag].samples()[-1] == _Tensor(
wall_time, step, value, file2)
# Current steps should be: [1, 2, 3, 4, 5, 10]
steps = [1, 2, 3, 4, 5, 10]
samples = ev_data._reservoir_by_tag[tag].samples()
for step, sample in zip(steps, samples):
filename = file1 if sample.step < 5 else file2
assert sample == _Tensor(wall_time, step, value, filename)
new_steps_2 = [7, 11, 3]
for step in new_steps_2:
t_event = TensorEvent(wall_time=1,
step=step,
tag=tag,
plugin_name=plugin_name,
value=value,
filename=file2)
ev_data.add_tensor_event(t_event)
# Current steps should be: [1, 2, 3, 5, 7, 10, 11], file2: [3, 5, 7, 10, 11]
steps = [1, 2, 3, 5, 7, 10, 11]
new_steps_2.extend(new_steps_1)
samples = ev_data._reservoir_by_tag[tag].samples()
for step, sample in zip(steps, samples):
filename = file2 if sample.step in new_steps_2 else file1
assert sample == _Tensor(wall_time, step, value, filename)