feat: refactor best/worst record api in response time

[rokamu623]

Description

The method of calculating ResponseTime for the best and worst-with-external-latency cases has been changed to be the same as for all and worst case.

Related links

Jira: https://tier4.atlassian.net/browse/RT2-889

Notes for reviewers

Pre-review checklist for the PR author

• I've confirmed the contribution guidelines.
• The PR is ready for review.

In-review checklist for the PR reviewers

The PR reviewers must check the checkboxes below before approval.

• The PR has been properly tested.
• The PR has been reviewed.

Post-review checklist for the PR author

The PR author must check the checkboxes below before merging.

• (Optional) The PR has been properly tested with CARET_report verification.
• There are no open discussions or they are tracked via tickets.
• The PR is ready for merge.

After all checkboxes are checked, anyone who has write access can merge the PR.

[nabetetsu]

@rokamu623 I see you have some commit in the fork repository, do you need any additional changes on?

[bopeng-saitama]

@rokamu623
I think need to add test for the following cases:
End Time Earlier Than Start Time:

    records_raw = [
        {'start': 3, 'end': 2},
        {'start': 5, 'end': 4},
    ]

Only End Time, No Start Time:

    records_raw = [
        {'end': 3},
        {'end': 5},
    ]

[bopeng-saitama]

LGTM

[xygyo77]

I used "autoware_launch_trace_20221227-150253_universe_rosbag" to get the timeseries graph of ResponseTime before and after applying this PR. Before applying this PR, the version was bokeh3.3.0.
Visual inspection showed that there were differences in "best" and "worst-with-external-latency" cases. Please check the difference.

BEFORE




AFTER



Specification change?

[rokamu623]

@xygyo77
Is the difference about best-case?

Regarding this,
Before refactoring, the response time of the 1st path is not calculated.
This is because in the worst-with-external-latency case, the start time of the previous path (worst-case external input time) cannot be defined.

In fact, except in the worst-with-external-latency case, the response time can be calculated for the 1st path (output time - input time).

Therefore, after refactoring, the response time for the 1st path is recorded more than before refactoring.

[xygyo77]

LGTM
Note: With https://tier4.atlassian.net/browse/RT2-888, the "worst-with-external-latency" case has been replaced by the old "worst" case.