class TuningParser:
"""Parser class is responsible for parsing tuning log files."""
def __init__(self, logs: list) -> None:
"""Initialize object."""
self._logs = logs
self.metric = Metric()
def process(self) -> Dict[str, Any]:
"""Process files."""
for log_file in self._logs:
log.debug(f"Read from {log_file}")
with open(log_file) as f:
for line in f:
for key in self.patterns:
prog = re.compile(self.patterns[key])
match = prog.search(line)
if match:
self.metric.insert_data(key, match.group(1))
parsed_data: Dict[str, Any] = self.metric.serialize() # type: ignore
return parsed_data
@property
def patterns(self) -> dict:
"""Set patterns to get metrics from lines."""
return {
"acc_fp32": r".*FP32 baseline is: \[(\d+.\d+),",
"acc_int8": r".*Best tune result is: \[(\d+.\d+),",
}
def update_data(
self,
request_id: Optional[str],
model_path: Optional[str] = None,
input_precision: Optional[str] = None,
model_output_path: Optional[str] = None,
output_precision: Optional[str] = None,
mode: Optional[str] = None,
metric: Optional[Union[Dict[str, Any], Metric]] = Metric(),
status: Optional[str] = None,
execution_details: Optional[Dict[str, Any]] = None,
) -> None:
"""Update data in workloads.list_json."""
self.load()
workload_info = WorkloadInfo(
workload_path=self.workload_path,
request_id=request_id,
model_path=model_path,
input_precision=input_precision,
model_output_path=model_output_path,
output_precision=output_precision,
mode=mode,
metric=metric,
status=status,
code_template_path=self.template_path,
execution_details=execution_details,
).serialize()
self.workloads_data["workloads"][request_id] = workload_info
self.dump()
def __init__(
self,
workspace_path: Optional[str] = None,
request_id: Optional[str] = None,
model_path: Optional[str] = None,
input_precision: Optional[str] = None,
model_output_path: Optional[str] = None,
output_precision: Optional[str] = None,
mode: Optional[str] = None,
metric: Optional[Union[dict, Metric]] = None,
overwrite: bool = True,
) -> None:
"""Initialize workdir class."""
self.workdir_path = os.path.join(os.environ.get("HOME", ""), ".lpot")
self.ensure_working_path_exists()
self.workloads_json = os.path.join(self.workdir_path, "workloads_list.json")
self.request_id = request_id
self.workload_path: str
if not metric:
metric = Metric()
if os.path.isfile(self.workloads_json):
self.workloads_data = self.load()
else:
self.workloads_data = {
"active_workspace_path": workspace_path,
"workloads": {},
"version": "2",
}
workload_data = self.get_workload_data(request_id)
if workload_data:
self.workload_path = workload_data.get("workload_path", "")
elif workspace_path and request_id:
workload_name = request_id
if model_path:
workload_name = "_".join(
[
Path(model_path).stem,
request_id,
],
)
self.workload_path = os.path.join(
workspace_path,
"workloads",
workload_name,
)
if request_id and overwrite:
self.update_data(
request_id=request_id,
model_path=model_path,
input_precision=input_precision,
model_output_path=model_output_path,
output_precision=output_precision,
mode=mode,
metric=metric,
)
def __init__(
self,
workspace_path: Optional[str] = None,
request_id: Optional[str] = None,
model_path: Optional[str] = None,
model_output_path: Optional[str] = None,
metric: Optional[Union[dict, Metric]] = None,
overwrite: bool = True,
) -> None:
"""Initialize workdir class."""
self.workdir_path = os.path.join(os.environ["HOME"], ".lpot")
self.ensure_working_path_exists()
self.workloads_json = os.path.join(self.workdir_path, "workloads_list.json")
self.request_id = request_id
self.workload_path: str
if not metric:
metric = Metric()
if os.path.isfile(self.workloads_json):
self.workloads_data = self.load()
else:
self.workloads_data = {
"active_workspace_path": workspace_path,
"workloads": {},
}
if (
self.workloads_data.get("workloads", None)
and self.workloads_data.get(
"workloads",
None,
).get(request_id, None)
):
self.workload_path = self.workloads_data["workloads"][request_id][
"workload_path"
]
elif workspace_path and request_id:
workload_name = request_id
if model_path:
workload_name = "_".join(
[
Path(model_path).stem,
request_id,
],
)
self.workload_path = os.path.join(
workspace_path,
"workloads",
workload_name,
)
if request_id and overwrite:
self.update_data(request_id, model_path, model_output_path, metric)
def process(self) -> Dict[str, Any]:
"""Process files."""
metric = Metric()
partial: Dict[str, List] = {}
for log_file in self._logs:
log.debug(f"Read from {log_file}")
with open(log_file) as f:
for line in f:
for key in self.patterns:
prog = re.compile(self.patterns[key])
match = prog.search(line)
if not match:
continue
metric_name = f"perf_{key}_fp32"
metric.insert_data(metric_name, match.group(1))
converted_value = getattr(metric, metric_name)
parse_result = {
key: converted_value,
}
partial = self.update_partial(partial, parse_result)
return self.summarize_partial(partial)
def __init__(self, logs: list) -> None:
"""Initialize parser."""
self._logs = logs
self.metric = Metric()