Files
C.E.L_Slide_test2/tests/matching/_archive/NORMALIZED_TOKEN_INVENTORY.html
T
KyeongminandClaude Opus 4.8 b836e79ee1 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>
2026-07-02 17:03:42 +09:00

1198 lines
20 KiB
HTML
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
<!DOCTYPE html>
<html lang="ko">
<head>
<meta charset="utf-8">
<title>Normalized Token Inventory — Step 4.2 snapshot</title>
<style>
body { font-family: -apple-system, "Segoe UI", Pretendard, sans-serif; max-width: 1100px; margin: 2em auto; padding: 0 1em; line-height: 1.55; color: #222; }
h1 { border-bottom: 2px solid #333; padding-bottom: 0.2em; }
h2 { margin-top: 2em; border-bottom: 1px solid #ccc; padding-bottom: 0.2em; }
h3 { margin-top: 1.5em; color: #555; }
table { border-collapse: collapse; margin: 0.5em 0; font-size: 0.92em; }
th, td { border: 1px solid #ddd; padding: 5px 9px; text-align: left; vertical-align: top; }
th { background: #f4f4f4; }
code { background: #f4f4f4; padding: 2px 4px; border-radius: 3px; font-size: 0.9em; }
strong { color: #0a6; }
</style>
</head>
<body>
<h1>NORMALIZED TOKEN INVENTORY — Step 4.2 확정 스냅샷</h1>
<h2>0. Executive Summary</h2>
<p>BEPS / Figma / MDX 에서 실제 텍스트를 수집해 중복 제거 후 <strong>2,124개</strong> 문장 / 항목을 정리하였다.
형태소 분석 단계에서는 <strong>13,911개</strong> 단어 출현을 확인했고,
정규화 후 최종 <strong>14,228개 단어 출현 / 1,743개 키워드 후보</strong>로 정리하였다.
이 후보 목록을 바탕으로 각 Figma 프레임별 핵심 키워드 (anchor keyword) 를 선별한다.</p>
<blockquote>
<p><strong>1,758개 = 전체 키워드 후보 풀</strong> (anchor keyword 는 이 중에서 프레임별로 선별할 핵심 키워드)</p>
</blockquote>
<h2>1. 파이프라인 처리 흐름</h2>
<table>
<thead>
<tr>
<th>단계</th>
<th>산출</th>
<th>역할</th>
</tr>
</thead>
<tbody>
<tr>
<td>Step 1</td>
<td><code>actual_text_nodes.yaml</code></td>
<td>Figma texts.md 의 <code>-</code> list item + MDX 본문에서 text_node 추출. HTML tag / markdown / HTML entity 정리. 1글자 / 순수숫자 / 단일기호 제외.</td>
</tr>
<tr>
<td>Step 2</td>
<td><code>actual_text_tokens.yaml</code></td>
<td>Kiwi 형태소 분석 (NNG / NNP / SL / SN). 1글자 / 순수숫자 제외. synonym 미적용.</td>
</tr>
<tr>
<td>Step 2.5</td>
<td><code>special_forms_inventory.yaml</code></td>
<td>S/W, H/W, 2D, 3D, BIM, DX, AS-IS, TO-BE 실제 표기형 관찰 (occurrences + nodes + source breakdown).</td>
</tr>
<tr>
<td>Step 4.2</td>
<td><code>normalized_text_tokens.yaml</code>, <code>replacement_report.yaml</code></td>
<td>PRE_COLLAPSE (4 regex) + phrase_variants 치환 → Kiwi user_dict (SL 8 + NNG 5) → 동일 필터.</td>
</tr>
</tbody>
</table>
<h2>2. 전체 카운트</h2>
<table>
<thead>
<tr>
<th>소스</th>
<th>raw tokens</th>
<th>unique tokens</th>
</tr>
</thead>
<tbody>
<tr>
<td>BEPS (1 frame)</td>
<td>10,517</td>
<td>1,708</td>
</tr>
<tr>
<td>Frames (32)</td>
<td>2,967</td>
<td>avg 63.6 / frame</td>
</tr>
<tr>
<td>MDX (3)</td>
<td>744</td>
<td>avg 117.2 / MDX</td>
</tr>
<tr>
<td><strong>Corpus</strong></td>
<td><strong>14,228</strong></td>
<td><strong>1,743</strong></td>
</tr>
</tbody>
</table>
<p><strong>총 occurrence</strong>: 14,228</p>
<h2>3. Special tokens 통계</h2>
<h3>3.1 Step 4.2 정규화 후 (token 기준)</h3>
<table>
<thead>
<tr>
<th>token</th>
<th>occurrence</th>
<th>frame_df</th>
<th>mdx_df</th>
</tr>
</thead>
<tbody>
<tr>
<td>S/W</td>
<td>286</td>
<td>22 / 32</td>
<td>4 / 3</td>
</tr>
<tr>
<td>H/W</td>
<td>28</td>
<td>6 / 32</td>
<td>1 / 3</td>
</tr>
<tr>
<td>2D</td>
<td>14</td>
<td>3 / 32</td>
<td>3 / 3</td>
</tr>
<tr>
<td>3D</td>
<td>83</td>
<td>13 / 32</td>
<td>3 / 3</td>
</tr>
<tr>
<td>DX</td>
<td>83</td>
<td>9 / 32</td>
<td>4 / 3</td>
</tr>
<tr>
<td>BIM</td>
<td>187</td>
<td>12 / 32</td>
<td>2 / 3</td>
</tr>
<tr>
<td>As-is</td>
<td>2</td>
<td>0 / 32</td>
<td>1 / 3</td>
</tr>
<tr>
<td>To-Be</td>
<td>2</td>
<td>0 / 32</td>
<td>1 / 3</td>
</tr>
</tbody>
</table>
<h3>3.2 Step 2.5 원본 표기형 관찰 (문자열 기준)</h3>
<table>
<thead>
<tr>
<th>target</th>
<th>occurrences (total)</th>
<th>source breakdown (b/f/m)</th>
<th>nodes (total)</th>
<th>observed_forms</th>
</tr>
</thead>
<tbody>
<tr>
<td>S/W</td>
<td>287</td>
<td>b197 / f84 / m6</td>
<td>230</td>
<td></td>
</tr>
<tr>
<td>H/W</td>
<td>28</td>
<td>b20 / f7 / m1</td>
<td>26</td>
<td></td>
</tr>
<tr>
<td>2D</td>
<td>14</td>
<td>b8 / f3 / m3</td>
<td>14</td>
<td></td>
</tr>
<tr>
<td>3D</td>
<td>83</td>
<td>b49 / f23 / m11</td>
<td>76</td>
<td></td>
</tr>
<tr>
<td>DX</td>
<td>50</td>
<td>b25 / f12 / m13</td>
<td>48</td>
<td></td>
</tr>
<tr>
<td>BIM</td>
<td>187</td>
<td>b148 / f28 / m11</td>
<td>169</td>
<td></td>
</tr>
<tr>
<td>AS-IS</td>
<td>2</td>
<td>b1 / f0 / m1</td>
<td>2</td>
<td><code>As-is</code>×2</td>
</tr>
<tr>
<td>TO-BE</td>
<td>2</td>
<td>b1 / f0 / m1</td>
<td>2</td>
<td><code>To-Be</code>×1, <code>To-be</code>×1</td>
</tr>
</tbody>
</table>
<h2>4. Applied replacements</h2>
<h3>4.1 PRE_COLLAPSE (regex, 4 rules)</h3>
<table>
<thead>
<tr>
<th>rule_id</th>
<th>applied</th>
</tr>
</thead>
<tbody>
<tr>
<td>digital_transition_with_dx</td>
<td>6</td>
</tr>
<tr>
<td>dx_with_digital_transition</td>
<td>4</td>
</tr>
</tbody>
</table>
<p><strong>pre_collapse_total</strong>: 10</p>
<h3>4.2 phrase_variants 치환 카운트</h3>
<table>
<thead>
<tr>
<th>canonical</th>
<th>applied</th>
</tr>
</thead>
<tbody>
<tr>
<td>DX</td>
<td>32</td>
</tr>
<tr>
<td>과정혁신</td>
<td>19</td>
</tr>
<tr>
<td>시행착오</td>
<td>1</td>
</tr>
<tr>
<td>의사소통</td>
<td>1</td>
</tr>
<tr>
<td>필수조건</td>
<td>1</td>
</tr>
<tr>
<td>결과혁신</td>
<td>1</td>
</tr>
<tr>
<td>To-Be</td>
<td>1</td>
</tr>
</tbody>
</table>
<p><strong>applied_replacements total</strong>: 56
<strong>total_text_changes</strong> (pre_collapse + phrase): 66</p>
<h3>4.3 phrase_variants 룰 정의 (<code>synonyms.yaml</code>)</h3>
<table>
<thead>
<tr>
<th>canonical</th>
<th>variants</th>
</tr>
</thead>
<tbody>
<tr>
<td>DX</td>
<td>디지털 전환, 디지털전환, 디지털 트랜스포메이션</td>
</tr>
<tr>
<td>결과혁신</td>
<td>결과 혁신, 결과의 혁신, Product의 변화, Product 혁신</td>
</tr>
<tr>
<td>과정혁신</td>
<td>과정 혁신, 과정의 혁신, Process 혁신, 프로세스 혁신</td>
</tr>
<tr>
<td>필수조건</td>
<td>필수 요건, 필수요건</td>
</tr>
<tr>
<td>의사소통</td>
<td>의사 소통</td>
</tr>
<tr>
<td>시행착오</td>
<td>시행 착오</td>
</tr>
<tr>
<td>As-is</td>
<td>As Is, As_is, as-is</td>
</tr>
<tr>
<td>To-Be</td>
<td>To-be, To Be, To_be</td>
</tr>
</tbody>
</table>
<h2>5. Frame 별 unique token count (32개)</h2>
<table>
<thead>
<tr>
<th>frame #</th>
<th>frame_id</th>
<th>raw tokens</th>
<th>unique tokens</th>
</tr>
</thead>
<tbody>
<tr>
<td>1</td>
<td>1171281172</td>
<td>19</td>
<td>18</td>
</tr>
<tr>
<td>2</td>
<td>1171281173</td>
<td>42</td>
<td>32</td>
</tr>
<tr>
<td>3</td>
<td>1171281174</td>
<td>18</td>
<td>17</td>
</tr>
<tr>
<td>4</td>
<td>1171281175</td>
<td>59</td>
<td>45</td>
</tr>
<tr>
<td>5</td>
<td>1171281176</td>
<td>28</td>
<td>21</td>
</tr>
<tr>
<td>6</td>
<td>1171281177</td>
<td>62</td>
<td>43</td>
</tr>
<tr>
<td>7</td>
<td>1171281178</td>
<td>243</td>
<td>147</td>
</tr>
<tr>
<td>8</td>
<td>1171281179</td>
<td>87</td>
<td>60</td>
</tr>
<tr>
<td>9</td>
<td>1171281180</td>
<td>42</td>
<td>28</td>
</tr>
<tr>
<td>10</td>
<td>1171281181</td>
<td>24</td>
<td>22</td>
</tr>
<tr>
<td>11</td>
<td>1171281182</td>
<td>45</td>
<td>28</td>
</tr>
<tr>
<td>12</td>
<td>1171281189</td>
<td>71</td>
<td>56</td>
</tr>
<tr>
<td>13</td>
<td>1171281190</td>
<td>84</td>
<td>57</td>
</tr>
<tr>
<td>14</td>
<td>1171281191</td>
<td>109</td>
<td>69</td>
</tr>
<tr>
<td>15</td>
<td>1171281192</td>
<td>115</td>
<td>83</td>
</tr>
<tr>
<td>16</td>
<td>1171281193</td>
<td>229</td>
<td>149</td>
</tr>
<tr>
<td>17</td>
<td>1171281194</td>
<td>150</td>
<td>97</td>
</tr>
<tr>
<td>18</td>
<td>1171281195</td>
<td>160</td>
<td>106</td>
</tr>
<tr>
<td>19</td>
<td>1171281197</td>
<td>64</td>
<td>45</td>
</tr>
<tr>
<td>20</td>
<td>1171281198</td>
<td>61</td>
<td>42</td>
</tr>
<tr>
<td>21</td>
<td>1171281201</td>
<td>44</td>
<td>34</td>
</tr>
<tr>
<td>22</td>
<td>1171281202</td>
<td>46</td>
<td>43</td>
</tr>
<tr>
<td>23</td>
<td>1171281203</td>
<td>148</td>
<td>100</td>
</tr>
<tr>
<td>24</td>
<td>1171281204</td>
<td>214</td>
<td>127</td>
</tr>
<tr>
<td>25</td>
<td>1171281205</td>
<td>48</td>
<td>43</td>
</tr>
<tr>
<td>26</td>
<td>1171281206</td>
<td>45</td>
<td>36</td>
</tr>
<tr>
<td>27</td>
<td>1171281208</td>
<td>38</td>
<td>30</td>
</tr>
<tr>
<td>28</td>
<td>1171281209</td>
<td>68</td>
<td>48</td>
</tr>
<tr>
<td>29</td>
<td>1171281210</td>
<td>164</td>
<td>115</td>
</tr>
<tr>
<td>30</td>
<td>1171281211</td>
<td>135</td>
<td>96</td>
</tr>
<tr>
<td>31</td>
<td>1171281212</td>
<td>222</td>
<td>126</td>
</tr>
<tr>
<td>32</td>
<td>1171281213</td>
<td>83</td>
<td>73</td>
</tr>
</tbody>
</table>
<h2>6. TARGET frames 상세 (13 / 14 / 18 / 29)</h2>
<h3>Frame 13 / 1171281190</h3>
<ul>
<li>raw_token_count: <strong>84</strong></li>
<li>unique_token_count: <strong>57</strong></li>
<li>unique_tokens:</li>
</ul>
<p>필수조건 / 기술 / 디지털 / 기반 / 지식 / 건설 / 산업 / 토목 / 단계 / 근본 / 이해 / 경험 / 전문 / 보유 / S/W / Digital / Technology / 다양 / 도구 / H/W / Process / 효과 / 통합 / 사람 / 역량 / 분야 / 기술자 / 엔지니어 / Know / how / 융합 / 축적 / 개발 / 개발자 / 형태 / Solution / Programmer / 자연 / 여건 / 사회 / 기업 / 제도 / 목표 / 설정 / 지침 / 탈피 / 수행 / 사례 / 검증 / 시행 / 지속 / 장기 / 투자 / 능력 / 의지 / 규모 / 과정혁신</p>
<h3>Frame 14 / 1171281191</h3>
<ul>
<li>raw_token_count: <strong>109</strong></li>
<li>unique_token_count: <strong>69</strong></li>
<li>unique_tokens:</li>
</ul>
<p>발주 / 목표 / 시공 / 설계자 / 민원 / 작업 / 예방 / 최소 / 직관 / 품질 / 향상 / 안정 / 제고 / 수행 / 공정 / 이해 / 관리 / 편의 / 증진 / 실무자 / 소통 / 오류 / 행정 / 자동 / 생산 / 건설 / 정보 / 통합 / 활용 / 강화 / 생애 / 주기 / 효율 / 디지털 / 자산 / 공사 / Risk / 시각 / 안전 / 중간 / 완성 / 측량 / 공간 / 편리 / 관계자 / 의사소통 / Model / 세도 / 관련 / 도서 / 작성 / System / 구축 / 기술 / 개발 / 원활 / 3D / 모델 / Claim / 상호 / 신뢰 / 부가 / 가치 / 창출 / 산업 / 전환 / 프로젝트 / 일관 / 유지</p>
<h3>Frame 18 / 1171281195</h3>
<ul>
<li>raw_token_count: <strong>160</strong></li>
<li>unique_token_count: <strong>106</strong></li>
<li>unique_tokens:</li>
</ul>
<p>BIM / DX / 이해 / vs. / Only / 3D / 모델 / 제작 / 상용 / S/W / Civil / Revit / Navisworks / Autocad / 기존 / 2D / 설계 / 방식 / 유지 / 중심 / 성과 / 일반 / 향상 / 분야 / 단절 / 시공 / 운영 / 단순 / 수동 / 집단 / 동질 / 오류 / 소극 / 기술 / 의존 / 제작사 / 판매 / 정책 / 평준 / 국내 / 규모 / 교육 / 집중 / 토목 / 해외 / 증가 / 프로세스 / 성과물 / 활용 / 확장 / 수행 / 개념 / CIVIL / IT / 주체 / 발주처 / 설계사 / 시공사 / ENG. / Management / 포함 / Rhino / Sketchup / Blender / EG / 전용 / 근본 / 문제 / 의식 / 개선 / 공학 / 정보 / 콘텐츠 / 연계 / 도면 / 수량 / 계획 / 일식 / 혁신 / 정립 / 생애 / 주기 / 시스템 / 구체 / 적극 / 실현 / 방안 / 접목 / 융합 / 자체 / 능력 / 지속 / 가능 / 확보 / 차별 / 경쟁력 / 진출 / ENG / 개발 / 건설 / 산업 / 과정 / 기초 / 일부분 / 인지 / 중요</p>
<h3>Frame 29 / 1171281210</h3>
<ul>
<li>raw_token_count: <strong>164</strong></li>
<li>unique_token_count: <strong>115</strong></li>
<li>unique_tokens:</li>
</ul>
<p>과정 / 혁신 / Process / 결과 / Products / 개념 / 도서 / 행정 / 절차 / 중심 / 2D / 도면 / 전문가 / 규정 / 업무 / 구분 / 책임 / 단절 / 시각 / 목적물 / 소통 / 투명 / 3D / 모델 / 참여자 / 실체 / 협업 / 창의 / 복합 / 위치 / 기반 / GIS / 건설 / 정보 / 포함 / BIM / 연계 / 과정혁신 / 지리 / 지형 / 지반 / 상태 / 형상 / 내용 / 속성 / 인프라 / 산업 / 등고선 / 지형도 / 수치 / 본태 / 측량 / 획득 / 구축 / 필수 / 설계 / 시공 / 관련 / 기술 / 정리 / 디지털 / S/W / 축적 / 작성 / 분절 / Analogue / 방식 / 성과물 / 가능 / Solution / 제공 / 과거 / 수작업 / 시행 / 발생 / 오류 / 최소 / 정확 / Data / 계획 / 개선 / 품질 / 의존 / 수량 / 계산서 / 시방서 / 시뮬레이션 / Digital / 추가 / Information / and / 숫자 / 확인 / 진행 / 지속 / 존속 / Graphic / 활용 / 기존 / 수준 / 공학 / 사이니지 / 시스템 / H/W / Engn. / 프로젝트 / 이슈 / 검토 / 논의 / 다양 / 단계 / 데이터 / 저장 / 통합 / 관리</p>
<h2>7. Top 50 by frequency</h2>
<table>
<thead>
<tr>
<th>token</th>
<th>count</th>
<th>frame_count</th>
<th>mdx_count</th>
</tr>
</thead>
<tbody>
<tr>
<td>S/W</td>
<td>286</td>
<td>22</td>
<td>4</td>
</tr>
<tr>
<td>정보</td>
<td>254</td>
<td>13</td>
<td>3</td>
</tr>
<tr>
<td>건설</td>
<td>200</td>
<td>15</td>
<td>4</td>
</tr>
<tr>
<td>BIM</td>
<td>187</td>
<td>12</td>
<td>2</td>
</tr>
<tr>
<td>관리</td>
<td>187</td>
<td>12</td>
<td>3</td>
</tr>
<tr>
<td>모델</td>
<td>185</td>
<td>12</td>
<td>3</td>
</tr>
<tr>
<td>기반</td>
<td>180</td>
<td>16</td>
<td>3</td>
</tr>
<tr>
<td>시공</td>
<td>177</td>
<td>15</td>
<td>4</td>
</tr>
<tr>
<td>설계</td>
<td>171</td>
<td>15</td>
<td>4</td>
</tr>
<tr>
<td>기술</td>
<td>160</td>
<td>17</td>
<td>3</td>
</tr>
<tr>
<td>개발</td>
<td>119</td>
<td>14</td>
<td>2</td>
</tr>
<tr>
<td>공사</td>
<td>111</td>
<td>4</td>
<td>1</td>
</tr>
<tr>
<td>계획</td>
<td>100</td>
<td>8</td>
<td>3</td>
</tr>
<tr>
<td>수행</td>
<td>93</td>
<td>10</td>
<td>3</td>
</tr>
<tr>
<td>산업</td>
<td>90</td>
<td>13</td>
<td>1</td>
</tr>
<tr>
<td>사용</td>
<td>90</td>
<td>13</td>
<td>0</td>
</tr>
<tr>
<td>현장</td>
<td>86</td>
<td>7</td>
<td>1</td>
</tr>
<tr>
<td>업무</td>
<td>86</td>
<td>13</td>
<td>3</td>
</tr>
<tr>
<td>활용</td>
<td>85</td>
<td>10</td>
<td>4</td>
</tr>
<tr>
<td>DX</td>
<td>83</td>
<td>9</td>
<td>4</td>
</tr>
<tr>
<td>3D</td>
<td>83</td>
<td>13</td>
<td>3</td>
</tr>
<tr>
<td>필요</td>
<td>79</td>
<td>12</td>
<td>3</td>
</tr>
<tr>
<td>효율</td>
<td>77</td>
<td>8</td>
<td>2</td>
</tr>
<tr>
<td>중심</td>
<td>73</td>
<td>7</td>
<td>4</td>
</tr>
<tr>
<td>가능</td>
<td>71</td>
<td>14</td>
<td>1</td>
</tr>
<tr>
<td>작업</td>
<td>70</td>
<td>6</td>
<td>1</td>
</tr>
<tr>
<td>디지털</td>
<td>68</td>
<td>7</td>
<td>4</td>
</tr>
<tr>
<td>단계</td>
<td>67</td>
<td>10</td>
<td>3</td>
</tr>
<tr>
<td>품질</td>
<td>65</td>
<td>10</td>
<td>2</td>
</tr>
<tr>
<td>시스템</td>
<td>64</td>
<td>6</td>
<td>2</td>
</tr>
<tr>
<td>작성</td>
<td>63</td>
<td>10</td>
<td>2</td>
</tr>
<tr>
<td>지원</td>
<td>61</td>
<td>5</td>
<td>0</td>
</tr>
<tr>
<td>다양</td>
<td>61</td>
<td>9</td>
<td>2</td>
</tr>
<tr>
<td>확인</td>
<td>60</td>
<td>4</td>
<td>1</td>
</tr>
<tr>
<td>공정</td>
<td>60</td>
<td>6</td>
<td>1</td>
</tr>
<tr>
<td>방식</td>
<td>59</td>
<td>9</td>
<td>3</td>
</tr>
<tr>
<td>데이터</td>
<td>58</td>
<td>4</td>
<td>3</td>
</tr>
<tr>
<td>생산</td>
<td>55</td>
<td>9</td>
<td>2</td>
</tr>
<tr>
<td>사업</td>
<td>54</td>
<td>7</td>
<td>0</td>
</tr>
<tr>
<td>안전</td>
<td>52</td>
<td>5</td>
<td>1</td>
</tr>
<tr>
<td>유지</td>
<td>52</td>
<td>9</td>
<td>3</td>
</tr>
<tr>
<td>이해</td>
<td>51</td>
<td>10</td>
<td>3</td>
</tr>
<tr>
<td>위치</td>
<td>51</td>
<td>5</td>
<td>2</td>
</tr>
<tr>
<td>현황</td>
<td>51</td>
<td>4</td>
<td>0</td>
</tr>
<tr>
<td>향상</td>
<td>51</td>
<td>6</td>
<td>3</td>
</tr>
<tr>
<td>기능</td>
<td>51</td>
<td>8</td>
<td>0</td>
</tr>
<tr>
<td>비용</td>
<td>50</td>
<td>7</td>
<td>2</td>
</tr>
<tr>
<td>구축</td>
<td>49</td>
<td>6</td>
<td>2</td>
</tr>
<tr>
<td>연계</td>
<td>47</td>
<td>8</td>
<td>2</td>
</tr>
<tr>
<td>지식</td>
<td>46</td>
<td>6</td>
<td>1</td>
</tr>
</tbody>
</table>
<h2>8. Top 50 by frame_df</h2>
<table>
<thead>
<tr>
<th>token</th>
<th>frame_count</th>
<th>count</th>
<th>mdx_count</th>
</tr>
</thead>
<tbody>
<tr>
<td>S/W</td>
<td>22</td>
<td>286</td>
<td>4</td>
</tr>
<tr>
<td>기술</td>
<td>17</td>
<td>160</td>
<td>3</td>
</tr>
<tr>
<td>기반</td>
<td>16</td>
<td>180</td>
<td>3</td>
</tr>
<tr>
<td>건설</td>
<td>15</td>
<td>200</td>
<td>4</td>
</tr>
<tr>
<td>시공</td>
<td>15</td>
<td>177</td>
<td>4</td>
</tr>
<tr>
<td>설계</td>
<td>15</td>
<td>171</td>
<td>4</td>
</tr>
<tr>
<td>개발</td>
<td>14</td>
<td>119</td>
<td>2</td>
</tr>
<tr>
<td>가능</td>
<td>14</td>
<td>71</td>
<td>1</td>
</tr>
<tr>
<td>정보</td>
<td>13</td>
<td>254</td>
<td>3</td>
</tr>
<tr>
<td>사용</td>
<td>13</td>
<td>90</td>
<td>0</td>
</tr>
<tr>
<td>산업</td>
<td>13</td>
<td>90</td>
<td>1</td>
</tr>
<tr>
<td>업무</td>
<td>13</td>
<td>86</td>
<td>3</td>
</tr>
<tr>
<td>3D</td>
<td>13</td>
<td>83</td>
<td>3</td>
</tr>
<tr>
<td>BIM</td>
<td>12</td>
<td>187</td>
<td>2</td>
</tr>
<tr>
<td>관리</td>
<td>12</td>
<td>187</td>
<td>3</td>
</tr>
<tr>
<td>모델</td>
<td>12</td>
<td>185</td>
<td>3</td>
</tr>
<tr>
<td>필요</td>
<td>12</td>
<td>79</td>
<td>3</td>
</tr>
<tr>
<td>수행</td>
<td>10</td>
<td>93</td>
<td>3</td>
</tr>
<tr>
<td>활용</td>
<td>10</td>
<td>85</td>
<td>4</td>
</tr>
<tr>
<td>단계</td>
<td>10</td>
<td>67</td>
<td>3</td>
</tr>
<tr>
<td>품질</td>
<td>10</td>
<td>65</td>
<td>2</td>
</tr>
<tr>
<td>작성</td>
<td>10</td>
<td>63</td>
<td>2</td>
</tr>
<tr>
<td>이해</td>
<td>10</td>
<td>51</td>
<td>3</td>
</tr>
<tr>
<td>지속</td>
<td>10</td>
<td>38</td>
<td>2</td>
</tr>
<tr>
<td>DX</td>
<td>9</td>
<td>83</td>
<td>4</td>
</tr>
<tr>
<td>다양</td>
<td>9</td>
<td>61</td>
<td>2</td>
</tr>
<tr>
<td>방식</td>
<td>9</td>
<td>59</td>
<td>3</td>
</tr>
<tr>
<td>생산</td>
<td>9</td>
<td>55</td>
<td>2</td>
</tr>
<tr>
<td>유지</td>
<td>9</td>
<td>52</td>
<td>3</td>
</tr>
<tr>
<td>개념</td>
<td>9</td>
<td>44</td>
<td>2</td>
</tr>
<tr>
<td>계획</td>
<td>8</td>
<td>100</td>
<td>3</td>
</tr>
<tr>
<td>효율</td>
<td>8</td>
<td>77</td>
<td>2</td>
</tr>
<tr>
<td>기능</td>
<td>8</td>
<td>51</td>
<td>0</td>
</tr>
<tr>
<td>연계</td>
<td>8</td>
<td>47</td>
<td>2</td>
</tr>
<tr>
<td>Solution</td>
<td>8</td>
<td>35</td>
<td>1</td>
</tr>
<tr>
<td>토목</td>
<td>8</td>
<td>31</td>
<td>2</td>
</tr>
<tr>
<td>현장</td>
<td>7</td>
<td>86</td>
<td>1</td>
</tr>
<tr>
<td>중심</td>
<td>7</td>
<td>73</td>
<td>4</td>
</tr>
<tr>
<td>디지털</td>
<td>7</td>
<td>68</td>
<td>4</td>
</tr>
<tr>
<td>사업</td>
<td>7</td>
<td>54</td>
<td>0</td>
</tr>
<tr>
<td>비용</td>
<td>7</td>
<td>50</td>
<td>2</td>
</tr>
<tr>
<td>범용</td>
<td>7</td>
<td>45</td>
<td>0</td>
</tr>
<tr>
<td>기존</td>
<td>7</td>
<td>44</td>
<td>3</td>
</tr>
<tr>
<td>한계</td>
<td>7</td>
<td>34</td>
<td>0</td>
</tr>
<tr>
<td>목표</td>
<td>7</td>
<td>32</td>
<td>0</td>
</tr>
<tr>
<td>규모</td>
<td>7</td>
<td>30</td>
<td>2</td>
</tr>
<tr>
<td>Digital</td>
<td>7</td>
<td>26</td>
<td>2</td>
</tr>
<tr>
<td>프로세스</td>
<td>7</td>
<td>26</td>
<td>4</td>
</tr>
<tr>
<td>Process</td>
<td>7</td>
<td>24</td>
<td>3</td>
</tr>
<tr>
<td>수준</td>
<td>7</td>
<td>21</td>
<td>0</td>
</tr>
</tbody>
</table>
<h2>9. 검수 메모</h2>
<ul>
<li><strong>DX(DX) 제거 완료</strong> — PRE_COLLAPSE 4개 룰 적용. <code>dx_duplicate</code> 잔여 0건.</li>
<li><strong>3D모델 / 2D도면 canonical 은 일부러 만들지 않음</strong> — 3D / 2D 일반 차원 토큰을 살리기 위함. 이후 anchor_set 단계에서 compound 후보로 별도 생성 가능.</li>
<li><strong>As-is / To-Be 는 frame 에 등장하지 않고 BEPS / MDX 중심</strong> — frame_df=0. BEPS/MDX 에만 각 2건.</li>
<li><strong>Source of truth 는 <code>actual_text_nodes.yaml</code> + <code>normalized_text_tokens.yaml</code></strong> — 이 보고서는 검수용 snapshot 이다.</li>
</ul>
</body>
</html>