Files
2nd/01_Archive/2026-04-20/p_reinforce_worker.py
T

126 lines
4.8 KiB
Python

import os
import re
import uuid
import sys
import shutil
from datetime import datetime
# UTF-8 Output support
if sys.stdout.encoding != 'utf-8':
import io
sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding='utf-8')
base_dir = r"e:\Wiki\2nd"
raw_dir = os.path.join(base_dir, "00_Raw", "2026-04-20")
wiki_base = os.path.join(base_dir, "10_Wiki", "💡 Topics")
# Simple keyword-based categorizer
CATEGORY_KEYWORDS = {
"AI": ["AI", "Artificial Intelligence", "LLM", "GPT", "Neural", "Deep Learning", "Machine Learning", "Adversarial"],
"Graphics & Performance": ["Graphics", "Rendering", "Three.js", "WebGL", "WebGPU", "Shader", "3D", "Gaussian Splatting", "Splat", "Frame"],
"Design & Experience": ["UX", "UI", "Design", "Accessibility", "A11y", "Interface", "HCI", "Cognitive", "Gamification"],
"Programming & Language": ["TypeScript", "JS", "C++", "Java", "Python", "Syntax", "AST", "Type", "Core", "Pattern", "Compiler"],
"Software Architecture": ["Architecture", "Microservices", "DDD", "API", "Contract", "System Design", "Cloud", "Distributed"],
"Psychology & Behavior": ["Psychology", "Behavior", "ABA", "Neuroscience", "Mind", "Cognitive", "Emotion", "Addiction"],
"Game Design": ["Game", "Level", "Narrative", "Player", "Quest", "Mechanic", "Simulation"],
"Health & Science": ["Health", "Medical", "Biomedical", "Biology", "Clinical", "Injury", "ACL", "Performance Optimization"],
"Security": ["Security", "OWASP", "Encryption", "Auth", "Hack", "Attack", "Malware", "Privacy"]
}
def get_category(filename, content):
filename_lower = filename.lower()
content_lower = content[:500].lower() # Check first 500 chars
for cat, keywords in CATEGORY_KEYWORDS.items():
for kw in keywords:
if kw.lower() in filename_lower or kw.lower() in content_lower:
return cat
return "General Knowledge"
def process_batch(limit=200):
files = [f for f in os.listdir(raw_dir) if f.endswith(".md")]
processed_count = 0
# Get existing wiki titles to skip
existing_titles = set()
for root, dirs, f_list in os.walk(wiki_base):
for f in f_list:
existing_titles.add(f.replace(".md", ""))
for filename in files:
if processed_count >= limit:
break
title_raw = filename.replace(".md", "")
safe_title = re.sub(r'[^\w\s\(\)\[\]-]', '', title_raw).strip()
if safe_title in existing_titles:
continue
raw_path = os.path.join(raw_dir, filename)
try:
with open(raw_path, "r", encoding="utf-8", errors="ignore") as f:
content = f.read()
except:
continue
category = get_category(filename, content)
category_path = f"10_Wiki/💡 Topics/{category}"
# Parse basic info
summary_match = re.search(r'##?\s*📌\s*Brief Summary\n(.*?)(?=\n##|$)', content, re.S)
summary = summary_match.group(1).strip() if summary_match else "지식 요약 정보 추출 중..."
core_match = re.search(r'##?\s*📖\s*Core Content\n(.*?)(?=\n##|$)', content, re.S)
core = core_match.group(1).strip() if core_match else "본문 구조화 작업 중..."
conn_match = re.search(r'##?\s*🔗\s*Knowledge Connections\n(.*?)(?=\n##|$)', content, re.S)
conn = conn_match.group(1).strip() if conn_match else ""
doc_id = f"P-REINFORCE-AUTO-{uuid.uuid4().hex[:6].upper()}"
today = datetime.now().strftime("%Y-%m-%d")
wiki_content = f"""---
id: {doc_id}
category: "[[{category_path}]]"
confidence_score: 0.90
tags: [auto-reinforced]
last_reinforced: {today}
github_commit: "[P-Reinforce] Continuous Worker - {safe_title}"
---
# [[{safe_title}]]
## 📌 한 줄 통찰 (The Karpathy Summary)
> {summary}
## 📖 구조화된 지식 (Synthesized Content)
{core}
## ⚠️ 모순 및 업데이트 (Contradictions & RL Update)
- **과거 데이터와의 충돌:** 자동화 엔진에 의해 매핑된 지식으로, 추후 정밀 검증 필요.
- **정책 변화:** {category} 분야의 자동 자산화 수행.
## 🔗 지식 연결 (Graph)
{conn}
- Raw Source: [[00_Raw/2026-04-20/{filename}]]
---
"""
target_dir = os.path.join(base_dir, category_path.replace("/", os.sep))
if not os.path.exists(target_dir):
os.makedirs(target_dir)
target_path = os.path.join(target_dir, f"{safe_title}.md")
with open(target_path, "w", encoding="utf-8") as f:
f.write(wiki_content)
print(f"[{processed_count+1}] Processed: {safe_title}")
processed_count += 1
return processed_count
if __name__ == "__main__":
count = process_batch(2000) # Process ALL remaining files
print(f"Total processed in this session: {count}")