feat: organize and categorize 1535 knowledge assets into wiki topics (2026-04-29)

2026-04-29 18:14:09 +09:00
parent 9e68271383
commit 3d9d0077cc
4106 changed files with 129792 additions and 0 deletions
@@ -0,0 +1,103 @@
+import os
+import shutil
+import re
+
+# 경로 설정
+SOURCE_ROOT = r"E:\Wiki\2nd\00_Raw\1535_Knowledge"
+TARGET_TOPICS = r"E:\Wiki\2nd\10_Wiki\Topics"
+CATEGORIES = {
+    "Art": r"E:\Wiki\2nd\10_Wiki\Topics_Art",
+    "Biz": r"E:\Wiki\2nd\10_Wiki\Topics_Biz",
+    "Blog": r"E:\Wiki\2nd\10_Wiki\Topics_Blog",
+    "GD": r"E:\Wiki\2nd\10_Wiki\Topics_GD"
+}
+
+# 키워드 정의
+KEYWORDS = {
+    "Art": ["art", "visual", "ui", "ux", "graphic", "asset", "design", "color", "icon", "stylized", "magitech", "frontend", "rendering", "shader"],
+    "Biz": ["business", "biz", "strategy", "market", "pm", "management", "operation", "growth", "revenue", "partnership", "governance", "compliance", "roadmap"],
+    "Blog": ["blog", "story", "narrative", "log", "devlog", "retro", "retrospective", "content", "storytelling", "personal"],
+    "GD": ["game", "mechanic", "balance", "gd", "level", "stage", "boss", "combat", "skill", "player", "enemy", "progression", "vampire", "survivor"]
+}
+
+# 폴더 기반 힌트
+FOLDER_HINTS = {
+    "01_Frontend_Mastery": ["Art"],
+    "AI & Games": ["GD"],
+    "AI & Narrative": ["Blog", "GD"],
+    "AI & Psychology": ["GD"],
+    "AI & ML MLOps": ["Biz"],
+    "04_Governance_Reliability": ["Biz"],
+    "AI & Tools": ["Biz", "GD"]
+}
+
+def classify(file_path, content):
+    matched_cats = set()
+    
+    # 1. 폴더 힌트 확인
+    for folder, cats in FOLDER_HINTS.items():
+        if folder in file_path:
+            matched_cats.update(cats)
+            
+    # 2. 파일명 및 내용 키워드 분석
+    lower_path = file_path.lower()
+    lower_content = content[:1000].lower()
+    
+    for cat, kws in KEYWORDS.items():
+        for kw in kws:
+            if kw in lower_path or kw in lower_content:
+                matched_cats.add(cat)
+                break
+                
+    return matched_cats
+
+def main():
+    # 디렉토리 생성
+    os.makedirs(TARGET_TOPICS, exist_ok=True)
+    for path in CATEGORIES.values():
+        os.makedirs(path, exist_ok=True)
+        
+    count = 0
+    
+    for root, dirs, files in os.walk(SOURCE_ROOT):
+        for file in files:
+            if not file.endswith(".md"):
+                continue
+                
+            src_path = os.path.join(root, file)
+            
+            try:
+                with open(src_path, 'r', encoding='utf-8') as f:
+                    content = f.read()
+            except Exception as e:
+                print(f"Error reading {src_path}: {e}")
+                continue
+                
+            # 분류
+            cats = classify(src_path, content)
+            
+            # 1. 기본 이동 (Topics)
+            # 파일명 중복 방지를 위해 폴더 구조 일부 반영하거나 유니크하게 처리 (여기서는 단순 이동 시도)
+            # 실제 1535개면 중복 가능성 높으므로 주의
+            dest_topics_path = os.path.join(TARGET_TOPICS, file)
+            
+            # 중복 회피 로직
+            if os.path.exists(dest_topics_path):
+                name, ext = os.path.splitext(file)
+                dest_topics_path = os.path.join(TARGET_TOPICS, f"{name}_{count}{ext}")
+            
+            shutil.move(src_path, dest_topics_path)
+            
+            # 2. 카테고리별 복사
+            for cat in cats:
+                cat_dest = os.path.join(CATEGORIES[cat], os.path.basename(dest_topics_path))
+                shutil.copy2(dest_topics_path, cat_dest)
+                
+            count += 1
+            if count % 100 == 0:
+                print(f"Processed {count} files...")
+
+    print(f"Total processed: {count}")
+
+if __name__ == "__main__":
+    main()