[
https://issues.apache.org/jira/browse/BEAM-14255?focusedWorklogId=770836&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-770836
]
ASF GitHub Bot logged work on BEAM-14255:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 16/May/22 13:38
Start Date: 16/May/22 13:38
Worklog Time Spent: 10m
Work Description: yeandy commented on code in PR #17671:
URL: https://github.com/apache/beam/pull/17671#discussion_r873727966
##########
sdks/python/apache_beam/ml/inference/base.py:
##########
@@ -193,13 +197,12 @@ def process(self, batch):
examples = batch
keys = None
- start_time = self._clock.get_current_time_in_microseconds()
+ start_time = _to_microseconds(self._clock.time())
result_generator = self._inference_runner.run_inference(
examples, self._model)
predictions = list(result_generator)
- inference_latency = self._clock.get_current_time_in_microseconds(
- ) - start_time
+ inference_latency = _to_microseconds(self._clock.time()) - start_time
Review Comment:
nit: to match the pattern used to compute `load_model_latency_ms`
```suggestion
end_time = _to_microseconds(self._clock.time())
inference_latency = end_time - start_time
```
##########
sdks/python/apache_beam/ml/inference/base.py:
##########
@@ -158,19 +163,18 @@ def __init__(self, model_loader: ModelLoader, clock=None):
self._inference_runner.get_metrics_namespace())
self._clock = clock
if not clock:
- self._clock = _ClockFactory.make_clock()
+ self._clock = time
Review Comment:
```suggestion
self._clock = time if not clock else clock
```
Issue Time Tracking
-------------------
Worklog Id: (was: 770836)
Time Spent: 1h 20m (was: 1h 10m)
> Drop the clock abastraction and just use time.time for time measurements
> ------------------------------------------------------------------------
>
> Key: BEAM-14255
> URL: https://issues.apache.org/jira/browse/BEAM-14255
> Project: Beam
> Issue Type: Sub-task
> Components: sdk-py-core
> Reporter: Ryan Thompson
> Assignee: Ryan Thompson
> Priority: P2
> Time Spent: 1h 20m
> Remaining Estimate: 0h
>
> Right now the TFX-BSL Runinference library uses an abstract clock class to
> get microsecond precision, but time.time should give an adequate precision.
>
> Investigate removing the clock abstraction and just using time.time.
>
> Alternatively, comment why the abstraction is useful.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
