chore: sort trace by loop_id

[qew21]

Description

Motivation and Context

How Has This Been Tested?

• If you are adding a new feature, test on your own test scripts.

Screenshots of Test Results (if appropriate):

1. Your own tests:

Types of changes

• Fix bugs
• Add new feature
• Update documentation

---

📚 Documentation preview 📚: https://RDAgent--977.org.readthedocs.build/en/977/

[Copilot]

Pull Request Overview

This PR refactors how SOTA experiment statistics are determined by sorting trace files by loop ID, propagates the chosen loop ID into downstream summary logic, and improves error logging with full tracebacks.

• Updated get_sota_exp_stat to accept an optional sota_loop_id and iterate through sorted feedback traces to find the decision point.
• Enhanced error handlers in mle_summary.py to include full stack traces and sorted messages by loop ID in summarize_folder.
• Propagated sota_loop_id into summary and added a new sota_exp_stat_new field.

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
rdagent/log/ui/utils.py	Changed get_sota_exp_stat signature, replaced max-based selection with sorting by loop ID
rdagent/log/ui/utils.py	Updated summary builder to pass through sota_loop_id when calling get_sota_exp_stat
rdagent/log/mle_summary.py	Added import traceback; improved exception logging; sorted log messages by loop ID

Comments suppressed due to low confidence (3)

rdagent/log/ui/utils.py:276

• [nitpick] The field name sota_exp_stat_new may be confusing alongside the existing sota_exp_stat. Consider renaming or documenting it more clearly (e.g., sota_exp_stat_resolved) to improve maintainability.

            v["sota_exp_stat_new"] = get_sota_exp_stat(Path(lf) / k, v.get("sota_loop_id"))

rdagent/log/ui/utils.py:146

• New behavior branches depending on sota_loop_id and sorted feedback traces. Please add unit tests for both flows: when sota_loop_id is provided and when it must be deduced from trace files.

def get_sota_exp_stat(log_path: Path, sota_loop_id: int | None = None):

rdagent/log/mle_summary.py:44

• Using print for exception output can make log aggregation harder. Consider using a logger (e.g., logger.exception(...)) to automatically include the traceback in structured logs.

                    print(f"Error in {log_trace_path}: {e}", traceback.format_exc())