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>

tests/matching/DECK_06_FIGMA_TO_DB.html

@@ -0,0 +1,105 @@
+<!DOCTYPE html>
+<html lang="ko"><head><meta charset="utf-8"><title>06. Frame → DB 저장 샘플</title>
+<style>
+@page { size: A4 portrait; margin: 12mm; }
+* { box-sizing: border-box; }
+body {
+    font-family: -apple-system, "Segoe UI", Pretendard, "Malgun Gothic", sans-serif;
+    background: white; color: #1e293b; margin: 0; font-size: 9.5pt; line-height: 1.4;
+}
+@media screen {
+    body { max-width: 210mm; margin: 10mm auto; box-shadow: 0 4px 24px rgba(0,0,0,0.12); padding: 12mm; background: #fff; }
+}
+h1 { font-size: 13pt; border-bottom: 2pt solid #2563eb; padding-bottom: 3pt; margin: 0 0 4pt 0; color: #0f172a; }
+h2 { font-size: 10pt; color: #1e293b; margin: 6pt 0 3pt; padding: 2pt 5pt; background: #e0e7ff; border-left: 3pt solid #0a6; border-radius: 2pt; }
+.meta { color: #64748b; font-size: 8pt; font-style: italic; margin-bottom: 5pt; }
+.mdx-card { background: #f8fafc; border: 0.5pt solid #cbd5e1; border-radius: 3pt; padding: 5pt 8pt; margin: 3pt 0 5pt; }
+.mdx-card .title { font-weight: 700; color: #0f172a; font-size: 9.5pt; margin-bottom: 2pt; }
+.popup-tag { display: inline-block; padding: 1pt 5pt; background: #eef2ff; color: #3730a3; border-radius: 8pt; font-size: 7.5pt; font-weight: 700; margin-left: 4pt; vertical-align: middle; }
+.mdx-card ul { margin: 2pt 0 0 0; padding-left: 12pt; font-size: 8pt; color: #475569; }
+table { border-collapse: collapse; width: 100%; font-size: 9pt; }
+th, td { border: 0.5pt solid #cbd5e1; padding: 3pt 5pt; text-align: left; vertical-align: top; }
+th { background: #1e293b; color: white; font-weight: 700; font-size: 8.5pt; text-align: center; }
+code { background: #f1f5f9; padding: 0 3pt; border-radius: 2pt; font-family: ui-monospace, Menlo, monospace; font-size: 8.5pt; color: #0f172a; }
+.formula-box { background: #eff6ff; border: 0.7pt solid #93c5fd; border-radius: 3pt; padding: 5pt 8pt; margin: 3pt 0; font-size: 9pt; color: #1e3a8a; }
+.label-use_as_is { background: #d1fae5; color: #065f46; padding: 1pt 5pt; border-radius: 2pt; font-size: 7.5pt; font-weight: 700; }
+.label-light_edit { background: #dbeafe; color: #1e40af; padding: 1pt 5pt; border-radius: 2pt; font-size: 7.5pt; font-weight: 700; }
+.label-restructure { background: #fef3c7; color: #92400e; padding: 1pt 5pt; border-radius: 2pt; font-size: 7.5pt; font-weight: 700; }
+.label-reject { background: #fee2e2; color: #991b1b; padding: 1pt 5pt; border-radius: 2pt; font-size: 7.5pt; font-weight: 700; }
+
+@page { size: A4 portrait; margin: 12mm; }
+.db-header { display: table; width: 100%; margin-bottom: 6pt; }
+.db-header > div { display: table-cell; vertical-align: middle; }
+.db-header .img-col { width: 30%; text-align: center; padding-right: 10pt; }
+.db-header .img-col img { max-width: 100%; max-height: 35mm; border: 0.5pt solid #94a3b8; border-radius: 3pt; }
+.db-header .info-col .name { font-weight: 700; color: #2563eb; font-size: 11pt; }
+.db-header .info-col .desc { font-size: 8.5pt; color: #475569; margin-top: 3pt; line-height: 1.5; }
+.db-section { background: #f8fafc; border: 0.7pt solid #cbd5e1; border-radius: 4pt; padding: 6pt 10pt; margin: 5pt 0; }
+.db-section .sec-title { font-size: 10pt; font-weight: 700; color: #2563eb; margin-bottom: 4pt; padding-bottom: 2pt; border-bottom: 1pt solid #cbd5e1; }
+.db-section .sec-desc { font-size: 8pt; color: #64748b; margin-bottom: 4pt; }
+pre.yaml { background: #1e293b; color: #e2e8f0; border-radius: 3pt; padding: 6pt 10pt; font-family: ui-monospace, "Menlo", monospace; font-size: 7.5pt; line-height: 1.4; margin: 3pt 0; overflow-x: auto; white-space: pre-wrap; word-break: break-all; }
+pre.yaml .key { color: #7dd3fc; }
+pre.yaml .str { color: #fde68a; }
+pre.yaml .num { color: #86efac; }
+pre.yaml .com { color: #94a3b8; font-style: italic; }
+</style></head><body><h1>06. Figma 디자인이 들어오면 DB 에 어떻게 저장되나 — Frame 18 샘플</h1><div class="meta">새 Figma 프레임이 입력되면 「키워드 추출」 + 「구조 분석」 두 갈래로 DB 에 저장. 매칭 시 이 DB 가 후보 풀이 됨. Frame 18 (BIM과 DX의 이해) 실제 저장 형태.</div><div class="db-header"><div class="img-col"><img src="data:image/png;base64,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 truncated
+<span class="key">standalone</span>:
+  <span class="key">total_count</span>: <span class="num">28</span>
+  <span class="key">terms</span>:
+    - <span class="str">Autocad</span>
+    - <span class="str">CIVIL</span>
+    - <span class="str">ENG</span>
+    - <span class="str">ENG.</span>
+    - <span class="str">IT</span>
+    - <span class="str">Management</span>
+    - <span class="str">Navisworks</span>
+    - <span class="str">Only</span>
+    <span class="com"># … 외 20개</span>
+
+<span class="com"># 키워드 세트 (같은 줄에 함께 등장하는 묶음 — 구조 단위)</span>
+<span class="key">source_text_sets</span>:
+  <span class="key">total_count</span>: <span class="num">39</span>
+  <span class="key">sets</span>:
+    - <span class="key">raw</span>: <span class="str">"BIM과 DX의 이해"</span>
+      <span class="key">tokens</span>: [<span class="str">BIM</span>, <span class="str">DX</span>, <span class="str">이해</span>]
+    - <span class="key">raw</span>: <span class="str">"BIM vs. DX"</span>
+      <span class="key">tokens</span>: [<span class="str">BIM</span>, <span class="str">vs.</span>, <span class="str">DX</span>]
+    - <span class="key">raw</span>: <span class="str">"Only 3D"</span>
+      <span class="key">tokens</span>: [<span class="str">Only</span>, <span class="str">3D</span>]
+    <span class="com"># … 외 36개 세트</span>
+
+<span class="com"># 연관 키워드 (여러 프레임 공통 — 보조 신호)</span>
+<span class="key">related</span>:
+  <span class="key">total_count</span>: <span class="num">78</span>
+  <span class="key">terms</span>: [<span class="str">2D</span>, <span class="str">3D</span>, <span class="str">BIM</span>, <span class="str">Blender</span>, <span class="str">Civil</span>, <span class="str">DX</span>, <span class="str">EG</span>, <span class="str">Revit</span> <span class="com"># … 외 70개</span>]</pre></div><div class="db-section"><div class="sec-title">구조 DB — 프레임 분석 결과</div><div class="sec-desc">Figma 프레임의 레이아웃·항목수·관계를 분석 + AI / 사람이 콘텐츠 성격·시각 의도·슬롯·대안 레이아웃 라벨링 → 저장.</div><pre class="yaml"><span class="com"># 시각 패턴 (자동 감지)</span>
+<span class="key">visual_pattern</span>:
+  <span class="key">layout</span>: <span class="str">compare-rows</span>         <span class="com"># 표/카드 등 레이아웃 ID</span>
+  <span class="key">family</span>: <span class="str">table</span>              <span class="com"># table / cards / visual / emphasis</span>
+  <span class="key">relation_type</span>: <span class="str">compare</span>        <span class="com"># compare / parallel / sequence</span>
+  <span class="key">cardinality</span>:
+    <span class="key">ideal</span>: <span class="num">2</span>  <span class="key">min</span>: <span class="num">2</span>  <span class="key">max</span>: <span class="num">2</span>
+
+<span class="com"># 콘텐츠 성격 (12 카테고리)</span>
+<span class="key">content_affinity</span>:
+  <span class="key">primary</span>: <span class="str">comparative_matrix</span>    <span class="com"># 행렬형 비교</span>
+  <span class="key">secondary</span>: [<span class="str">capability_requirements</span>, <span class="str">interrelation</span>]
+
+<span class="com"># 시각 의도 (9 카테고리)</span>
+<span class="key">structure_intent</span>:
+  <span class="key">primary</span>: <span class="str">matrix_coverage</span>    <span class="com"># 매트릭스</span>
+  <span class="key">secondary</span>: [<span class="str">binary_compare</span>, <span class="str">cycle_interrelation</span>]
+
+<span class="com"># 편집 슬롯 (텍스트 교체 가능 영역)</span>
+<span class="key">slots</span>:
+  - <span class="str">title</span>
+  - <span class="str">col_a_label</span>
+  - <span class="str">col_b_label</span>
+  - <span class="str">rows</span>
+
+<span class="com"># 대안 레이아웃 (이 프레임으로 못 담을 때 대체 가능)</span>
+<span class="key">alternative_patterns</span>:
+  - <span class="key">pattern</span>: <span class="str">table-2col</span>  <span class="key">confidence</span>: <span class="num">1.00</span>
+  - <span class="key">pattern</span>: <span class="str">table-3col</span>  <span class="key">confidence</span>: <span class="num">0.93</span>
+  - <span class="key">pattern</span>: <span class="str">paired-rows</span>  <span class="key">confidence</span>: <span class="num">0.50</span>
+  - <span class="key">pattern</span>: <span class="str">compare-2col</span>  <span class="key">confidence</span>: <span class="num">0.45</span>
+  <span class="com"># … 외 1개</span></pre></div><div class="db-section" style="background:#f0fdf4;border-color:#86efac"><div class="sec-title" style="color:#065f46;border-color:#86efac">저장된 DB 가 매칭 시 어떻게 쓰이나</div><div style="font-size:8.5pt;color:#1e293b;line-height:1.6">· <b>키워드 매칭 (V1)</b> : <code>standalone</code> / <code>source_text_sets</code> / <code>related</code> 가 MDX 본문 단어와 대조 → 키워드 점수<br/>· <b>구조 매칭 (V3)</b> : <code>visual_pattern.layout</code> / <code>content_affinity</code> / <code>structure_intent</code> 가 MDX 의 자동 분석 결과와 대조 → 구조 점수<br/>· <b>종합 판정 (V4)</b> : <code>cardinality</code> / <code>relation_type</code> / <code>slots</code> 가 V4 5축의 cardinality/relation/slot 점수에 사용 → confidence<br/>· <b>적용 불가 시</b> : <code>alternative_patterns</code> 의 후보로 fallback</div></div></body></html>