wip: phase_z2 evidence 파이프라인 + matching 실험(phase2~26) + 프론트 trace 패널 진행분 스냅샷

- src: phase_z2 composition/mapper/pipeline/placement_planner/retry, ai_fallback(prompts/schema/validate), mdx_text_atoms 신규
- Front: PipelineTracePanel 신규, FramePanel/SlideCanvas/Home/designAgentApi 등 갱신 + 테스트 4종 추가
- templates/phase_z2: catalog(component_expansion_registry, node_slot_mapping 신규), frames, families, slide_base 갱신
- tests/matching: phase2~26 매칭 실험 스크립트·리포트·온톨로지 전체 (미커밋 진행분)
- tests: b4_v4 evidence, task5~28.5 시리즈, regression(imp95 baseline) 등 신규 테스트 대량 추가
- docs/reference: MDX 구조 인벤토리, MDX→Frame 구조 계약 문서
- scripts: mdx 계약/parity/coverage/viewport 체크, gitea comment, run sync 유틸
- .gitignore: tmp*.json, chromedriver, .orchestrator, *.pkl, Front_test* 등 임시/스냅샷 제외

미완성 작업의 보존용 스냅샷 커밋 (2026-07-02)

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
2026-07-02 17:03:42 +09:00
co-authored by Claude Opus 4.8
parent 97b7833a1b
commit b836e79ee1
527 changed files with 673036 additions and 717 deletions
@@ -0,0 +1,382 @@
meta:
pipeline_step: 8.v3.r5
description: r4 MDX 하네스 + Frame 13 SSOT 확정 ontology
top_k: 5
answer_map:
01-2: 18
02-2.2: 14
03-1: 13
03-2: 29
holdout_sections:
- 01-1
- 02-1
- 02-2.1
lock_snapshot:
timestamp: '2026-04-24T11:42:05'
top_k: 5
files:
pipeline_08_v3_r5_structure_rerank.py: 217f67497c12c0e92cfc10d0d56ac962a983532fa3e28efbdd7c3bd8d682be39
pipeline_08_v3_r4_structure_rerank.py: 6836e889d15d4591dcb170556cc38ab39f7d17294728c919966099d855c80f30
structure_ontology_v2_final_r2.yaml: b009417abb7905e40c9bb57cf87ca99c6e58e5ef8008861e2d2d0c423b599979
v2_semantic_rerank_result.yaml: 6e4e2917bd0f1849e236f841b0ba85ad7333faae62abe1bbe08bb091c76fc325
r5_changes:
- 'ontology 입력: final.yaml → final_r2.yaml (Frame 13 review_queue 해제 반영)'
- 'MDX 하네스: r4 (B-1 + B-2) 그대로'
mdx_sections:
01-2:
section_type: null
answer_frame_number: 18
mdx_title: 2. 용어간 상호관계
mdx_layout: compare-rows
v3_rerank:
- frame_id: '1171281195'
frame_number: 18
v1_rank: 1
v2_rank: 1
v3_rank: 1
v1_score: 0.937
semantic_score: 0.7295
structure_compat: 0.55
fig_layout: compare-rows
- frame_id: '1171281208'
frame_number: 27
v1_rank: 2
v2_rank: 2
v3_rank: 2
v1_score: 0.392
semantic_score: 0.6442
structure_compat: 0.49
fig_layout: central-split
- frame_id: '1171281180'
frame_number: 9
v1_rank: 5
v2_rank: 5
v3_rank: 3
v1_score: 0.358
semantic_score: 0.5556
structure_compat: 0.35
fig_layout: list-stacked
- frame_id: '1171281172'
frame_number: 1
v1_rank: 4
v2_rank: 4
v3_rank: 4
v1_score: 0.362
semantic_score: 0.5741
structure_compat: 0.25
fig_layout: circular-nodes
- frame_id: '1171281182'
frame_number: 11
v1_rank: 3
v2_rank: 3
v3_rank: 5
v1_score: 0.379
semantic_score: 0.5853
structure_compat: 0.21
fig_layout: 3-category
02-2.2:
section_type: null
answer_frame_number: 14
mdx_title: 2.2 DX 시행 주체별 기대효과
mdx_layout: persona-3col
v3_rerank:
- frame_id: '1171281191'
frame_number: 14
v1_rank: 1
v2_rank: 1
v3_rank: 1
v1_score: 0.84
semantic_score: 0.6416
structure_compat: 0.8
fig_layout: persona-3col
- frame_id: '1171281176'
frame_number: 5
v1_rank: 5
v2_rank: 5
v3_rank: 2
v1_score: 0.404
semantic_score: 0.3762
structure_compat: 0.5
fig_layout: side-card
- frame_id: '1171281174'
frame_number: 3
v1_rank: 4
v2_rank: 3
v3_rank: 3
v1_score: 0.411
semantic_score: 0.4471
structure_compat: 0.23
fig_layout: list-numbered
- frame_id: '1171281180'
frame_number: 9
v1_rank: 2
v2_rank: 4
v3_rank: 4
v1_score: 0.495
semantic_score: 0.3861
structure_compat: 0.23
fig_layout: list-stacked
- frame_id: '1171281189'
frame_number: 12
v1_rank: 3
v2_rank: 2
v3_rank: 5
v1_score: 0.459
semantic_score: 0.4858
structure_compat: 0.215
fig_layout: cycle-3way
03-1:
section_type: null
answer_frame_number: 13
mdx_title: 1. DX 시행을 위한 필수 요건
mdx_layout: 3col-parallel
v3_rerank:
- frame_id: '1171281190'
frame_number: 13
v1_rank: 1
v2_rank: 1
v3_rank: 1
v1_score: 0.598
semantic_score: 0.634
structure_compat: 0.8
fig_layout: 3col-parallel
- frame_id: '1171281198'
frame_number: 20
v1_rank: 3
v2_rank: 3
v3_rank: 2
v1_score: 0.388
semantic_score: 0.5188
structure_compat: 0.7648
fig_layout: 3col-cards
- frame_id: '1171281197'
frame_number: 19
v1_rank: 5
v2_rank: 5
v3_rank: 3
v1_score: 0.269
semantic_score: 0.4209
structure_compat: 0.435
fig_layout: 3col-compare
- frame_id: '1171281192'
frame_number: 15
v1_rank: 2
v2_rank: 2
v3_rank: 4
v1_score: 0.542
semantic_score: 0.5959
structure_compat: 0.23
fig_layout: policy-4card-plus-list
- frame_id: '1171281208'
frame_number: 27
v1_rank: 4
v2_rank: 4
v3_rank: 5
v1_score: 0.363
semantic_score: 0.4577
structure_compat: 0.0
fig_layout: central-split
03-2:
section_type: null
answer_frame_number: 29
mdx_title: 2. Process의 혁신과 Product의 변화
mdx_layout: compare-2banner
v3_rerank:
- frame_id: '1171281210'
frame_number: 29
v1_rank: 1
v2_rank: 1
v3_rank: 1
v1_score: 0.705
semantic_score: 0.5991
structure_compat: 0.74
fig_layout: compare-2banner-top-2col-bottom
- frame_id: '1171281182'
frame_number: 11
v1_rank: 4
v2_rank: 3
v3_rank: 2
v1_score: 0.42
semantic_score: 0.465
structure_compat: 0.21
fig_layout: 3-category
- frame_id: '1171281189'
frame_number: 12
v1_rank: 2
v2_rank: 4
v3_rank: 3
v1_score: 0.521
semantic_score: 0.4395
structure_compat: 0.1
fig_layout: cycle-3way
- frame_id: '1171281179'
frame_number: 8
v1_rank: 5
v2_rank: 2
v3_rank: 4
v1_score: 0.383
semantic_score: 0.4832
structure_compat: 0.0
fig_layout: 3-section
- frame_id: '1171281180'
frame_number: 9
v1_rank: 3
v2_rank: 5
v3_rank: 5
v1_score: 0.426
semantic_score: 0.3975
structure_compat: 0.0
fig_layout: list-stacked
01-1:
section_type: null
answer_frame_number: null
mdx_title: 1. 용어 정의
mdx_layout: 3col-parallel
v3_rerank:
- frame_id: '1171281198'
frame_number: 20
v1_rank: 5
v2_rank: 5
v3_rank: 1
v1_score: 0.238
semantic_score: 0.5005
structure_compat: 0.7248
fig_layout: 3col-cards
- frame_id: '1171281182'
frame_number: 11
v1_rank: 3
v2_rank: 2
v3_rank: 2
v1_score: 0.249
semantic_score: 0.5507
structure_compat: 0.7152
fig_layout: 3-category
- frame_id: '1171281189'
frame_number: 12
v1_rank: 4
v2_rank: 3
v3_rank: 3
v1_score: 0.248
semantic_score: 0.547
structure_compat: 0.612
fig_layout: cycle-3way
- frame_id: '1171281172'
frame_number: 1
v1_rank: 1
v2_rank: 4
v3_rank: 4
v1_score: 0.551
semantic_score: 0.5368
structure_compat: 0.29
fig_layout: circular-nodes
- frame_id: '1171281208'
frame_number: 27
v1_rank: 2
v2_rank: 1
v3_rank: 5
v1_score: 0.298
semantic_score: 0.5881
structure_compat: 0.0
fig_layout: central-split
02-1:
section_type: null
answer_frame_number: null
mdx_title: 1. DX의 궁극적 목표
mdx_layout: 3col-parallel
v3_rerank:
- frame_id: '1171281191'
frame_number: 14
v1_rank: 2
v2_rank: 2
v3_rank: 1
v1_score: 0.382
semantic_score: 0.5779
structure_compat: 0.66
fig_layout: persona-3col
- frame_id: '1171281198'
frame_number: 20
v1_rank: 5
v2_rank: 1
v3_rank: 2
v1_score: 0.237
semantic_score: 0.5895
structure_compat: 0.6198
fig_layout: 3col-cards
- frame_id: '1171281189'
frame_number: 12
v1_rank: 1
v2_rank: 3
v3_rank: 3
v1_score: 0.766
semantic_score: 0.57
structure_compat: 0.612
fig_layout: cycle-3way
- frame_id: '1171281174'
frame_number: 3
v1_rank: 4
v2_rank: 5
v3_rank: 4
v1_score: 0.302
semantic_score: 0.4707
structure_compat: 0.33
fig_layout: list-numbered
- frame_id: '1171281208'
frame_number: 27
v1_rank: 3
v2_rank: 4
v3_rank: 5
v1_score: 0.356
semantic_score: 0.5467
structure_compat: 0.0
fig_layout: central-split
02-2.1:
section_type: null
answer_frame_number: null
mdx_title: 2.1 업무 수행 과정(Process)의 변화
mdx_layout: single-column
v3_rerank:
- frame_id: '1171281182'
frame_number: 11
v1_rank: 1
v2_rank: 4
v3_rank: 1
v1_score: 0.246
semantic_score: 0.4311
structure_compat: 0.43
fig_layout: 3-category
- frame_id: '1171281174'
frame_number: 3
v1_rank: 3
v2_rank: 3
v3_rank: 2
v1_score: 0.184
semantic_score: 0.4648
structure_compat: 0.18
fig_layout: list-numbered
- frame_id: '1171281180'
frame_number: 9
v1_rank: 5
v2_rank: 5
v3_rank: 3
v1_score: 0.171
semantic_score: 0.4191
structure_compat: 0.18
fig_layout: list-stacked
- frame_id: '1171281179'
frame_number: 8
v1_rank: 2
v2_rank: 2
v3_rank: 4
v1_score: 0.193
semantic_score: 0.5021
structure_compat: 0.08
fig_layout: 3-section
- frame_id: '1171281210'
frame_number: 29
v1_rank: 4
v2_rank: 1
v3_rank: 5
v1_score: 0.176
semantic_score: 0.5511
structure_compat: 0.0
fig_layout: compare-2banner-top-2col-bottom