perf(pipeline): dedupe the TemplatesToMiGroups bridge and stop cloning records

[nh13]

Wave 2 perf cleanup (#8 of the feat-runall review series) — the highest-value item: it hits the primary runall workload (fused sort→group→consensus / group→consensus). Two commits, read in order.

1. refactor(pipeline): extract the bridge into one factory

add_simplex, add_duplex, and add_codec each inlined a ~70-line copy of the TemplatesToMiGroups bridge (the fused group→consensus path that splices each template's MI into the MI tag and runs records into per-MI groups). The three copies were identical except for one flag — duplex_strip_strand (true for duplex, false for simplex/codec). Hoisted into a single templates_to_mi_step(limit_bytes, duplex_strip_strand) factory called from all three sites. Pure dedup, no behavior change.

2. perf(pipeline): move tagged bytes into records instead of cloning

The bridge tags each record's MI in a reused scratch buffer, then did RawRecord::from(state.scratch.clone()) — a full per-record heap copy on top of the extend_from_slice, once per read on the main runall workload. Now moves the buffer into the record with std::mem::replace, reinstalling a fresh Vec pre-sized to the previous capacity so the next record reuses it without a realloc (the per-record clear() keeps it warm). Same allocation count as the clone, minus the record-body memcpy. Output unchanged.

Verification

• cargo ci-fmt / cargo ci-lint (clippy pedantic) clean.
• cargo nextest -E 'test(runall) or test(group_determinism)' — 149 tests pass, covering all three fused group→{simplex,duplex,codec} shapes and MI determinism.
• Net −132 lines.
• CodeRabbit CLI (--agent) and a CodeRabbit-style self-review: 0 findings.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
⚠️ Please upload report for BASE (feat-runall@f9e7817). Learn more about missing BASE report.

Additional details and impacted files

@@              Coverage Diff               @@
##             feat-runall     #431   +/-   ##
==============================================
  Coverage               ?   93.40%           
==============================================
  Files                  ?      139           
  Lines                  ?    54817           
  Branches               ?        0           
==============================================
  Hits                   ?    51200           
  Misses                 ?     3617           
  Partials               ?        0           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[nh13]

@coderabbitai review

[coderabbitai]

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 0b5df7d1-bfaf-482a-a7b2-39caf7b9d2ad

📥 Commits

Reviewing files that changed from the base of the PR and between f9e7817 and 3065da4.

📒 Files selected for processing (1)

• src/lib/pipeline/chains/builder.rs

---

📝 Walkthrough

Walkthrough

In builder.rs, a new module-level helper templates_to_mi_step(limit_bytes, duplex_strip_strand) is introduced. It encapsulates the process_with_worker_state step that converts BatchedProcessedPositionGroups to BatchedMiGroups: iterating templates, computing MI keys (stripping duplex strand when duplex_strip_strand = true, otherwise deriving the base MI string), updating raw MI tags via fgumi_raw_bam::update_string_tag, and emitting grouped MiGroups. The three inline copies of this logic in add_simplex, add_duplex, and add_codec are replaced with calls to the helper, reducing net line count by ~132 lines.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately summarizes the main change: deduplicating the TemplatesToMiGroups bridge and eliminating record cloning in the pipeline.
Description check	✅ Passed	The description is directly related to the changeset, detailing the refactor and perf improvements with clear context, verification steps, and impact metrics.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feat-runall-31-mi-bridge-clone

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}