llm_wiki/raw/教育AI研究/tools/add_cross_references.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import sys
import io
import re
import datetime
from pathlib import Path
from typing import Dict, List, Set, Tuple
from collections import defaultdict

if sys.platform == "win32":
    sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding="utf-8")
    sys.stderr = io.TextIOWrapper(sys.stderr.buffer, encoding="utf-8")


class WikilinkManager:
    def __init__(self, project_root: Path):
        self.project_root = project_root
        self.knowledge_cards_dir = project_root / "知识卡片"
        self.institution_dir = project_root / "机构档案"
        self.weekly_reports_dir = project_root / "每周报告"
        self.deep_reports_dir = project_root / "深度研究报告"
        self.wikilinks: Dict[str, Set[str]] = defaultdict(set)

    def extract_title_from_file(self, file_path: Path) -> str:
        try:
            content = file_path.read_text(encoding="utf-8")
            fm_match = re.match(r"^---\n(.*?)\n---", content, re.DOTALL)
            if fm_match:
                for line in fm_match.group(1).split("\n"):
                    if line.startswith("title:"):
                        return line.replace("title:", "").strip().strip('"')
            first_heading = re.search(r"^#\s+(.+)$", content, re.MULTILINE)
            if first_heading:
                return first_heading.group(1).strip()
        except:
            pass
        return file_path.stem

    def build_relationship_map(self) -> Dict[str, List[str]]:
        relationships = {
            "LLM教育应用": ["智能辅导系统", "自适应学习系统", "个性化学习系统"],
            "智能辅导系统": ["自适应学习系统", "CMU LearnLab", "Khanmigo"],
            "自适应学习系统": ["智能辅导系统", "个性化学习系统", "自适应学习环境"],
            "个性化学习系统": ["自适应学习系统", "自适应学习环境"],
            "智能评测技术": ["教育大数据分析", "教育机器人应用"],
            "教育大数据分析": ["智能评测技术", "教育机器人应用"],
            "教育机器人应用": ["智能评测技术", "教育大数据分析"],
            "自适应学习环境": ["自适应学习系统", "个性化学习系统"],
            "RCT研究与Cohen's d指标": ["哈佛CS50课程模式", "哈佛教务长框架"],
            "哈佛CS50课程模式": ["MIT RAISE框架", "斯坦福AI Accelerator", "牛津AIEOU"],
            "哈佛教务长框架": ["斯坦福AI Accelerator", "牛津AIEOU", "CMU LearnLab"],
            "MIT RAISE框架": ["斯坦福AI Accelerator", "Khanmigo"],
            "斯坦福AI Accelerator": ["牛津AIEOU", "CMU LearnLab", "Khanmigo"],
            "牛津AIEOU": ["CMU LearnLab", "以人为本AI教育观"],
            "CMU LearnLab": ["Khanmigo", "智能辅导系统"],
            "Khanmigo": ["斯坦福AI Accelerator", "智能辅导系统"],
            "以人为本AI教育观": ["高等教、AI全球图景"],
            "高等教育AI全球图景": ["以人为本AI教育观", "哈佛CS50课程模式"],
        }
        return relationships

    def add_wikilinks_to_file(
        self, file_path: Path, related_concepts: List[str]
    ) -> bool:
        try:
            content = file_path.read_text(encoding="utf-8")

            if "## 相关概念" in content or "## 相关链接" in content:
                print(f"  ⏭️ Skipping (already has related section): {file_path.name}")
                return False

            section = f"""

---

## 相关概念

- [[{"|".join(related_concepts)}]]

"""

            if content.rstrip().endswith("---"):
                content = content.rstrip() + section
            else:
                content = content.rstrip() + section

            file_path.write_text(content, encoding="utf-8")
            print(f"  ✅ Added related concepts to: {file_path.name}")
            return True

        except Exception as e:
            print(f"  ❌ Error adding wikilinks to {file_path.name}: {e}")
            return False

    def process_knowledge_cards(self):
        relationships = self.build_relationship_map()

        if not self.knowledge_cards_dir.exists():
            print(f"⚠️ Knowledge cards directory not found")
            return

        card_files = list(self.knowledge_cards_dir.glob("*.md"))
        print(f"\nProcessing {len(card_files)} knowledge cards...")

        updated_count = 0
        for card_file in card_files:
            card_name = card_file.stem.replace("-知识卡片", "")

            related = relationships.get(card_name, [])
            if related:
                result = self.add_wikilinks_to_file(card_file, related)
                if result:
                    updated_count += 1

        return updated_count

    def generate_cross_reference_report(self) -> str:
        relationships = self.build_relationship_map()

        report = f"""# 跨文档引用增强报告

## 基本信息

- **生成时间**: {datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")}
- **涉及知识卡片**: {len(relationships)} 个

## 关系映射

| 概念 | 相关概念 |
|------|----------|
"""

        for concept, related in sorted(relationships.items()):
            report += f"| {concept} | {', '.join(related)} |\n"

        report += f"""

## 更新统计

- **已更新文件**: 在知识卡片末尾添加了"相关概念"章节
- **链接格式**: 使用Obsidian wikilink格式 `[[概念名]]`
- **关系类型**: 学科关联、技术关联、应用关联

## 下一步建议

1. 在机构档案中添加相关知识卡片链接
2. 在周报和深度报告中添加更多内部引用
3. 定期检查和更新跨文档引用
4. 考虑建立双向链接（反向引用）

---

*报告生成时间: {datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")}*
"""

        return report


def main():
    print("=" * 60)
    print("Cross-Document References Enhancement Tool")
    print("=" * 60)
    print(f"Execution time: {datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
    print()

    project_dir = Path(__file__).parent.parent
    manager = WikilinkManager(project_dir)

    updated = manager.process_knowledge_cards()

    report = manager.generate_cross_reference_report()
    report_path = (
        project_dir
        / "outputs"
        / f"cross_reference_enhancement_{datetime.date.today()}.md"
    )
    report_path.parent.mkdir(parents=True, exist_ok=True)
    with open(report_path, "w", encoding="utf-8") as f:
        f.write(report)

    print()
    print("=" * 60)
    print("Enhancement Complete")
    print("=" * 60)
    print(f"Files updated: {updated}")
    print(f"Report saved to: {report_path.name}")
    print("=" * 60)


if __name__ == "__main__":
    main()