Fix multi tries duration result extraction from traces (#34)

When benchmarking with traces, each event generated does not contain a run_id field, and it is therefore not possible to aggregate the all events according to their run ids. This commit removes all logic related to run_ids. Co-authored-by: Boon Jun Soh <boonjun@google.com>

Author: boonjun-google

src/benchmark_utils.py

@@ -74,16 +74,11 @@ def get_metrics_from_trace(trace: dict[str, Any], task: str) -> float:
  if "name" in e and event_matcher.match(e["name"]):
  events.append(e)
 
- events_by_run_id = defaultdict(list)
- for e in events:
- run_id = e["args"]["run_id"] if "args" in e and "run_id" in e["args"] else "0"
- events_by_run_id[run_id].append(e)
  durations_ms = []
  try:
  # Duration is in us.
- durations_ms = [
- max([e["dur"] for e in es]) / 1e3 for run_id, es in events_by_run_id.items()
- ]
+ for e in events:
+ durations_ms.append(e["dur"] / 1e3)
  except KeyError:
  print("KeyError: Key 'dur' not found in the event object")
  raise