⚡ Bolt: Optimize dict.setdefault in Python backend loops

.jules/bolt.md

@@ -10,3 +10,8 @@
 ## 2024-05-24 - Memoizing inline array maps
 **Learning:** Inline mapping of arrays inside JSX in large React components causes O(N) recalculation on every render.
 **Action:** Wrap inline JSX elements that map over arrays (e.g., lists of tasks) in a `useMemo` hook with specific dependencies.
+
+## 2025-02-12 - Avoided unused setdefault list allocations in grouping loops
+
+**Learning:** `dict.setdefault(key, []).append(value)` evaluates the empty-list default on every iteration, including when the key already exists. In grouping loops, `defaultdict(list)` avoids those transient unused list allocations while preserving insertion order.
+**Action:** Use `defaultdict(list)` when missing keys are intentionally initialized with lists. Keep `setdefault` when its eager-default behavior or an ordinary `dict` is part of the required contract, and benchmark before claiming a material end-to-end improvement.

CHANGELOG.md

@@ -66,6 +66,7 @@
 
 ### 코드 건강성 개선 (Code Health)
 
+- 백엔드 그룹화 루프에 `defaultdict(list)`를 적용해 기존 순서와 응답을 보존하면서 `setdefault`가 매 반복마다 만들던 미사용 빈 리스트 할당을 피했습니다.
 - `WorkspaceHome`의 작업 완료 토글과 Reply SLA 팔로업 생성 로직을 `useTasks` hook으로 분리하고, 화면 쪽 formatter를 주입해 작업 제목 정규화 로직 중복을 방지했습니다.
 - `backend/api/security.py`에서 사용하지 않는 `from __future__ import annotations` 구문을 제거하고 조건 표현식을 정리했습니다.
 - `backend/alembic/env.py`에서 사용하지 않는 `from __future__ import annotations` 구문을 제거해 Alembic 환경 설정 코드를 간결하게 정리했습니다.

backend/api/data.py

@@ -1,5 +1,6 @@
 import base64
 import binascii
+from collections import defaultdict
 from datetime import datetime, timezone
 import hashlib
 import json
@@ -1695,14 +1696,14 @@ def _diligence_risk_matrix(
     groups: dict[
         tuple[RemediationPriority, str, str],
         list[DataDiligenceExceptionRegisterEntry],
-    ] = {}
+    ] = defaultdict(list)
     for exception in snapshot.diligence_exception_register:
         key = (
             exception.severity_code,
             exception.owner_area,
             exception.related_artifact,
         )
-        groups.setdefault(key, []).append(exception)
+        groups[key].append(exception)
 
     entries: list[DataDiligenceRiskMatrixEntry] = []
     for (severity, owner_area, related_artifact), exceptions in sorted(
@@ -1844,9 +1845,9 @@ def _diligence_close_decision_summary(
 def _diligence_close_artifact_review_queue(
     snapshot: DataEvidenceSnapshotResponse,
 ) -> list[DataDiligenceCloseArtifactReviewQueueEntry]:
-    groups: dict[str, list[DataDiligenceCloseProofPlanEntry]] = {}
+    groups: dict[str, list[DataDiligenceCloseProofPlanEntry]] = defaultdict(list)
     for proof in snapshot.diligence_close_proof_plan:
-        groups.setdefault(proof.required_proof_artifact, []).append(proof)
+        groups[proof.required_proof_artifact].append(proof)
 
     entries: list[DataDiligenceCloseArtifactReviewQueueEntry] = []
     for artifact, proofs in sorted(groups.items()):
@@ -1889,9 +1890,9 @@ def _diligence_close_artifact_review_queue(
 def _diligence_close_owner_handoff_queue(
     snapshot: DataEvidenceSnapshotResponse,
 ) -> list[DataDiligenceCloseOwnerHandoffQueueEntry]:
-    groups: dict[str, list[DataDiligenceCloseProofPlanEntry]] = {}
+    groups: dict[str, list[DataDiligenceCloseProofPlanEntry]] = defaultdict(list)
     for proof in snapshot.diligence_close_proof_plan:
-        groups.setdefault(proof.owner_area, []).append(proof)
+        groups[proof.owner_area].append(proof)
 
     entries: list[DataDiligenceCloseOwnerHandoffQueueEntry] = []
     for owner_area, proofs in sorted(groups.items()):

backend/services/email_import_service.py

@@ -1,5 +1,6 @@
 import asyncio
 import datetime
+from collections import defaultdict
 from email import policy as email_policy
 import hashlib
 import logging
@@ -575,7 +576,7 @@ def add_edge(
     segments_by_source: dict[
         tuple[str, str],
         list[ContentSegmentRecord],
-    ] = {}
+    ] = defaultdict(list)
     for segment in sorted(
         email_obj.content_segments,
         key=lambda item: (
@@ -585,10 +586,9 @@ def add_edge(
             item.segment_path,
         ),
     ):
-        segments_by_source.setdefault(
-            (segment.source_kind, segment.source_record_uid),
-            [],
-        ).append(segment)
+        segments_by_source[
+            (segment.source_kind, segment.source_record_uid)
+        ].append(segment)
         add_edge(
             edge_kind="node_has_segment",
             edge_path=f"{segment.content_node.node_path}/has/{segment.segment_path}",

backend/services/project_graph/project_registration.py

@@ -2,7 +2,7 @@
 
 import datetime
 import hashlib
-from collections import Counter
+from collections import Counter, defaultdict
 from dataclasses import dataclass
 from typing import Any, Iterable, Mapping
 
@@ -612,9 +612,9 @@ def _candidate_groups(
     *,
     scope: ProjectGraphQueryScope,
 ) -> tuple[_CandidateGroup, ...]:
-    records_by_email: dict[int, list[ProjectGraphObjectRecord]] = {}
+    records_by_email: dict[int, list[ProjectGraphObjectRecord]] = defaultdict(list)
     for record in records:
-        records_by_email.setdefault(record.email_id, []).append(record)
+        records_by_email[record.email_id].append(record)
 
     groups: list[_CandidateGroup] = []
     for group_records in records_by_email.values():
@@ -632,7 +632,9 @@ def _candidate_groups(
             if explicit_candidate is not None
             else _synthetic_project_uid(group_records, scope=scope)
         )
-        groups.append(_CandidateGroup(project_uid=project_uid, records=tuple(group_records)))
+        groups.append(
+            _CandidateGroup(project_uid=project_uid, records=tuple(group_records))
+        )
     return tuple(groups)
 
 
@@ -842,9 +844,9 @@ def _relation_summary(
     ``relation_count`` descending with a ``relation_type``-ascending tie-break so
     the result is deterministic regardless of relation iteration order.
     """
-    grouped: dict[str, list[ProjectTraceRelation]] = {}
+    grouped: dict[str, list[ProjectTraceRelation]] = defaultdict(list)
     for relation in relations:
-        grouped.setdefault(relation.relation_type, []).append(relation)
+        grouped[relation.relation_type].append(relation)
     type_summaries = [
         ProjectRelationTypeSummary(
             relation_type=relation_type,