ltcai 0.1.0

package/p_reinforce.py

@@ -0,0 +1,148 @@
+"""
+P-Reinforce Knowledge Gardener
+Raw 데이터를 자동으로 분석해서 구조화된 마크다운 위키로 정리
+"""
+
+import json
+import os
+import re
+import time
+import shutil
+from datetime import datetime
+from pathlib import Path
+from typing import Optional
+
+BRAIN_DIR = Path(
+    os.getenv("LATTICEAI_OBSIDIAN_VAULT_DIR")
+    or os.getenv("LATTICEAI_BRAIN_DIR")
+    or Path.home() / ".ltcai-brain"
+)
+
+STRUCTURE = {
+    "10_Wiki":   "검증된 지식, 개념 설명, 레퍼런스",
+    "00_Raw":    "정제되지 않은 원시 데이터, 아이디어 메모",
+    "20_Skills": "재사용 가능한 코드 스니펫, 프롬프트, 워크플로",
+    "30_Projects": "프로젝트별 컨텍스트, 진행 상황",
+    "40_Log":    "날짜별 작업 로그",
+}
+
+
+class PReinforceGardener:
+    def __init__(self):
+        self._ensure_structure()
+
+    def _ensure_structure(self):
+        for folder in STRUCTURE:
+            (BRAIN_DIR / folder).mkdir(parents=True, exist_ok=True)
+        # 인덱스 파일
+        index_path = BRAIN_DIR / "INDEX.md"
+        if not index_path.exists():
+            index_path.write_text(self._render_index())
+
+    def _render_index(self) -> str:
+        lines = ["# 🧠 Lattice AI Brain — P-Reinforce Index\n"]
+        lines.append(f"*Generated: {datetime.now().strftime('%Y-%m-%d %H:%M')}*\n")
+        lines.append("\nThis folder is an Obsidian-compatible Markdown vault.\n")
+        for folder, desc in STRUCTURE.items():
+            lines.append(f"## [{folder}](./{folder}/)\n_{desc}_\n")
+        lines.append("## Connector Status\n")
+        lines.append(f"- OCR engine: `{'tesseract' if shutil.which('tesseract') else 'not installed'}`\n")
+        return "\n".join(lines)
+
+    # ── Classify ──────────────────────────────────────────────────────────────
+
+    def _classify(self, text: str) -> str:
+        """간단한 규칙 기반 분류 (LLM 없이도 동작)"""
+        text_lower = text.lower()
+
+        code_signals = ["def ", "class ", "import ", "```", "function ", "const ", "let ", "var "]
+        if any(s in text for s in code_signals):
+            return "20_Skills"
+
+        wiki_signals = ["개념", "원리", "이란", "what is", "how does", "definition", "explanation"]
+        if any(s in text_lower for s in wiki_signals):
+            return "10_Wiki"
+
+        project_signals = ["project", "프로젝트", "todo", "task", "작업", "기능", "feature"]
+        if any(s in text_lower for s in project_signals):
+            return "30_Projects"
+
+        return "00_Raw"
+
+    # ── File Naming ───────────────────────────────────────────────────────────
+
+    def _make_filename(self, text: str, folder: str) -> str:
+        # 첫 줄을 제목으로
+        first_line = text.strip().split("\n")[0][:60]
+        # 파일명 안전하게
+        safe = re.sub(r"[^\w\s-]", "", first_line).strip()
+        safe = re.sub(r"\s+", "_", safe)
+        timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+        return f"{timestamp}_{safe or 'note'}.md"
+
+    # ── Process ───────────────────────────────────────────────────────────────
+
+    async def process(self, raw_data: str, category: Optional[str] = None) -> dict:
+        folder = category if category in STRUCTURE else self._classify(raw_data)
+        filename = self._make_filename(raw_data, folder)
+        filepath = BRAIN_DIR / folder / filename
+
+        # 마크다운 래핑
+        content = self._wrap_markdown(raw_data, folder)
+        filepath.write_text(content, encoding="utf-8")
+
+        # 오늘 로그에도 기록
+        self._append_log(raw_data[:200], folder, filename)
+
+        return {
+            "status": "saved",
+            "folder": folder,
+            "filename": filename,
+            "path": str(filepath),
+            "classified_as": folder,
+            "description": STRUCTURE[folder],
+        }
+
+    def _wrap_markdown(self, raw: str, folder: str) -> str:
+        now = datetime.now().strftime("%Y-%m-%d %H:%M")
+        first_line = raw.strip().split("\n")[0][:80]
+        lines = [
+            f"# {first_line}",
+            f"\n> 📁 `{folder}` | 🕐 {now} | Lattice AI MLX\n",
+            "---\n",
+            raw,
+            "\n\n---",
+            f"*Auto-organized by P-Reinforce Gardener*",
+        ]
+        return "\n".join(lines)
+
+    def _append_log(self, preview: str, folder: str, filename: str):
+        today = datetime.now().strftime("%Y-%m-%d")
+        log_path = BRAIN_DIR / "40_Log" / f"{today}.md"
+        entry = f"\n- [{datetime.now().strftime('%H:%M')}] → `{folder}/{filename}`\n  > {preview[:100]}\n"
+        with open(log_path, "a", encoding="utf-8") as f:
+            if log_path.stat().st_size == 0 if log_path.exists() else True:
+                f.write(f"# 📅 Log — {today}\n")
+            f.write(entry)
+
+    # ── Tree ──────────────────────────────────────────────────────────────────
+
+    def get_relevant_context(self, query: str, limit: int = 3) -> str:
+        """질문과 관련된 지식을 검색하여 컨텍스트 문자열을 반환합니다."""
+        results = []
+        # 모든 마크다운 파일 탐색 (INDEX.md 및 Log 제외)
+        for file_path in BRAIN_DIR.rglob("*.md"):
+            if file_path.name == "INDEX.md" or "40_Log" in str(file_path):
+                continue
+            
+            try:
+                content = file_path.read_text(encoding="utf-8")
+                keywords = [k for k in re.split(r'\s+', query) if len(k) > 1]
+                if any(k.lower() in content.lower() for k in keywords):
+                    results.append(f"--- Document: {file_path.name} ---\n{content[:800]}")
+                    if len(results) >= limit:
+                        break
+            except Exception:
+                continue
+        
+        return "\n\n".join(results)

package/package.json

@@ -0,0 +1,44 @@
+{
+  "name": "ltcai",
+  "version": "0.1.0",
+  "description": "Lattice AI local MLX/cloud LLM workspace server",
+  "bin": {
+    "ltcai": "bin/ltcai.js",
+    "LTCAI": "bin/ltcai.js"
+  },
+  "scripts": {
+    "start": "LTCAI",
+    "dev": "python3 ltcai_cli.py --reload",
+    "build:python": "python3 -m build",
+    "check:python": "python3 -m py_compile ltcai_cli.py server.py llm_router.py p_reinforce.py telegram_bot.py tools.py codex_telegram_bot.py",
+    "publish:npm": "npm publish --access public",
+    "publish:pypi": "python3 -m twine upload dist/*"
+  },
+  "keywords": [
+    "ltcai",
+    "llm",
+    "mlx",
+    "LTCAI"
+  ],
+  "license": "MIT",
+  "private": false,
+  "files": [
+    "bin/ltcai.js",
+    "LICENSE",
+    "ltcai_cli.py",
+    "server.py",
+    "llm_router.py",
+    "p_reinforce.py",
+    "telegram_bot.py",
+    "tools.py",
+    "codex_telegram_bot.py",
+    "static/index.html",
+    "static/indexd.html",
+    "static/admin.html",
+    "requirements.txt",
+    "README.md"
+  ],
+  "publishConfig": {
+    "access": "public"
+  }
+}

package/requirements.txt

@@ -0,0 +1,11 @@
+fastapi
+uvicorn
+pydantic
+httpx
+pillow
+openai
+python-docx
+openpyxl
+python-pptx
+python-multipart
+keyring