NoteConnection is a visualization tool that transforms your Markdown notes into a hierarchical knowledge graph. It helps you understand the structure of your knowledge base by identifying prerequisites, learning paths, and clusters.
- Select Data Source: Upon opening the application, use the dropdown menu in the top-left corner to select a folder from your
Knowledge_Base. - Load Graph: Click the "Load" button. For large datasets, the system will use parallel processing to build the graph efficiently.
- Explore: The graph will render in the main view. You can pan and zoom using your mouse or touch gestures.
- Force-Directed: Default mode. Good for seeing clusters and groups.
- DAG (Directed Acyclic Graph): Hierarchical mode. Arranges notes based on dependencies (Top -> Bottom). Useful for learning paths.
- Enter: Double-click any node to enter Focus Mode.
- View: The selected node moves to the center.
- Helping to understand (Left/Bottom): Prerequisites and incoming connections.
- Further exploration (Right/Top): Next steps and outgoing connections.
- Controls: Use the sliders at the bottom to adjust Vertical and Horizontal spacing.
- Open Content: Click the "Specific Content" button to read the node's full text.
- Exit: Click the "Exit Focus" button or double-click the background.
- Note: "Freeze Layout" allows you to manually arrange nodes in this mode without them snapping back.
- Hover (PC): Hover over a node to see its connections and details.
- Click (Mobile/PC): Click a node to "Freeze" it and open the Statistics Popup.
- Popup: Shows In-degree (Red) and Out-degree (Blue) lists. You can drag and zoom this popup.
- Actions: Click any neighbor in the list to navigate to it.
- Panel: Located on the right side.
- Freeze Layout: Check this to stop all movement. Useful for large graphs or manual arrangement.
- Quick Button: Use the snowflake icon (❄️) in the toolbar for quick toggling.
- Speed/Damping: Adjust the slider to control how fast the graph settles.
- Access: Scroll down or click "Analysis" to view the data table.
- Interaction: Clicking a row in the table highlights the corresponding node in the graph.
- Mobile: The panel supports full-screen mode and pinch-to-zoom for better readability.
- Access: Click the Gear icon.
- Options:
- Language: Switch between English and Chinese.
- Physics: Tune gravity and repulsion forces.
- Performance:
- Enable GPU Acceleration: Toggle GPU usage for accelerated matrix operations (e.g., Cosine Similarity). Recommended for AMD Radeon 7900XT or compatible GPUs.
- Max Workers: Adjust the number of CPU threads used for parallel processing (Default: CPU Cores - 1). Higher values speed up large builds but use more memory.
- Memory Saving Mode: (v0.9.63) Enable to optimize backend memory usage for large datasets (10k+ files). Reduces Peak Heap by streaming file content and serializing data efficiently. Default: Enabled.
- Visuals: Adjust edge opacity.
- Canvas Auto-Switch: If your graph contains more than 3000 nodes, NoteConnection automatically switches to the Canvas Renderer to ensure smooth performance. You can manually switch back to SVG in the controls if needed.
- Physics Culling: For extremely dense graphs (>20,000 edges), the physics simulation will only calculate forces for a subset of edges to prevent freezing the interface, while still rendering all connections.
NoteConnection includes a one-click script to build a native Android APK (v1.1.0+).
- Prerequisites:
- Node.js (LTS)
- Java JDK (17 or higher)
- Android SDK (configured in environment)
- How to Build:
- Open the project folder in a terminal.
- Run
build_apk.bat. - Follow the on-screen instructions. The script will check your environment, install dependencies, build the web app, and compile the APK.
- Output: The APK will be generated at
android/app/build/outputs/apk/debug/app-debug.apk.
- Access: Click the "Path Mode" button in the top-left toolbar.
- Features:
- Domain Learning: Topological ordering of a specific topic cluster (Mastery).
- Diffusion Learning: Find the shortest path to a target concept (Efficiency).
- Controls:
- Strategy: Choose between "Foundational" (Base concepts first) or "Core" (Key concepts first).
- Layout: Switch between Radial, Vertical Tree, and Horizontal Tree views.
- Exit: Click "Exit" to return to the main graph.
- Desktop Mode: In the desktop application (Tauri), Path Mode opens in a high-performance native 3D window (Godot) for rendering up to 50,000 nodes smoothly, automatically syncing via WebSocket.
- Single-window behavior:
- In
npm run tauri:dev:mini:gpu, Tauri is the default visible window. - Entering Path Mode switches visibility to the Godot frontend.
- At any moment, only one primary frontend window should remain visible.
- In
- Godot close confirmation:
- Closing the Godot window should show a confirmation dialog.
- Users can either return to the main interface or close all windows.
- NoteMD embedded workflow:
- NoteMD is integrated as an embedded workflow in both Tauri and Godot flows (not a standalone desktop app window).
- File/folder/save browse actions in Tauri are expected to trigger native pickers and return selected paths.
- PDF import rule:
- PDF files must be converted to Markdown with Mineru before importing into NoteMD.
NoteConnection 是一个可视化工具,可将您的 Markdown 笔记转换为层级知识图谱。它通过识别先决条件、学习路径和聚类,帮助您理解知识库的结构。
- 选择数据源: 打开应用程序后,使用左上角的下拉菜单从
Knowledge_Base中选择一个文件夹。 - 加载图谱: 点击“Load”按钮。对于大数据集,系统将使用并行处理来高效构建图谱。
- 探索: 图谱将在主视图中渲染。您可以使用鼠标或触摸手势进行平移和缩放。
- 力导向 (Force-Directed): 默认模式。适合查看聚类和分组。
- DAG (有向无环图): 层级模式。根据依赖关系排列笔记(上 -> 下)。适用于查看学习路径。
- 进入: 双击任意节点进入专注模式。
- 视图: 选中的节点移动到中心。
- 帮助理解 (左/下): 先决条件和入度连接。
- 进一步探索 (右/上): 后续步骤和出度连接。
- 控制: 使用底部的滑块调整垂直和水平间距。
- 打开内容: 点击“打开具体内容”按钮以阅读节点的全文。
- 退出: 点击“退出专注”按钮或双击背景。
- 注意: “冻结布局”允许您在此模式下手动排列节点,而不会自动弹回。
- 悬停 (PC): 悬停在节点上以查看其连接和详情。
- 点击 (移动端/PC): 点击节点以“冻结”它并打开统计弹窗。
- 弹窗: 显示入度 (红色) 和出度 (蓝色) 列表。您可以拖动和缩放此弹窗。
- 操作: 点击列表中的任意邻居以导航到该节点。
- 面板: 位于右侧。
- 冻结布局: 选中此项以停止所有移动。适用于大图或手动排列。
- 快速按钮: 使用工具栏中的雪花图标 (❄️) 进行快速切换。
- 速度/阻尼: 调整滑块以控制图谱稳定的速度。
- 访问: 向下滚动或点击“Analysis”以查看数据表。
- 交互: 点击表格中的行会在图谱中高亮显示相应的节点。
- 移动端: 面板支持全屏模式和捏合缩放,以获得更好的可读性。
- 访问: 点击齿轮图标。
- 选项:
- 语言: 在英语和中文之间切换。
- 物理: 调整重力和排斥力。
- 性能:
- 启用 GPU 加速: 切换用于加速矩阵运算(如余弦相似度)的 GPU 使用。推荐用于 AMD Radeon 7900XT 或兼容 GPU。
- 最大 Worker 数: 调整用于并行处理的 CPU 线程数(默认:CPU 核心数 - 1)。较高的值可加快大型构建速度,但会占用更多内存。
- 省内存模式 (Memory Saving Mode): (v0.9.63) 启用以优化大数据集(10k+ 文件)的后端内存使用。通过流式传输文件内容和高效序列化数据来降低峰值堆内存。默认:已启用。
- 视觉: 调整边透明度。
- Canvas 自动切换: 如果您的图谱包含超过 3000 个节点,NoteConnection 会自动切换到 Canvas 渲染器以确保流畅的性能。如有需要,您可以在控件中手动切回 SVG。
- 物理剔除: 对于极其密集的图谱(>20,000 条边),物理模拟将仅计算边的一个子集的力,以防止界面冻结,同时仍渲染所有连接。
NoteConnection 包含一个一键式脚本,用于构建原生 Android APK (v1.1.0+)。
- 先决条件:
- Node.js (LTS)
- Java JDK (17 或更高版本)
- Android SDK (已在环境中配置)
- 如何构建:
- 在终端中打开项目文件夹。
- 运行
build_apk.bat。 - 按照屏幕上的指示操作。脚本将检查您的环境,安装依赖项,构建 Web 应用程序,并编译 APK。
- 输出: APK 将生成于
android/app/build/outputs/apk/debug/app-debug.apk。
- 访问: 点击左上角工具栏中的“Path Mode”按钮。
- 功能:
- 领域学习: 特定主题集群的拓扑排序(掌握模式)。
- 扩散学习: 寻找通往目标概念的最短路径(效率模式)。
- 控制:
- 策略: 选择“基础优先”(Foundational) 或“核心优先”(Core)。
- 布局: 在径向 (Radial)、垂直树 (Vertical Tree) 和水平树 (Horizontal Tree) 视图之间切换。
- 退出: 点击“Exit”返回主图谱。
- 桌面模式: 在桌面端应用程序 (Tauri) 中,路径模式会在高性能原生 3D 窗口 (Godot) 中打开,可流畅渲染多达 50,000 个节点,并通过 WebSocket 自动同步。
- 单窗口行为:
- 在
npm run tauri:dev:mini:gpu下,默认显示 Tauri 主窗口。 - 进入 Path Mode 时切换为 Godot 前端显示。
- 任意时刻应只显示一个主前端窗口。
- 在
- Godot 关闭确认:
- 关闭 Godot 窗口时应先弹出确认框。
- 用户可选择返回主界面,或关闭全部窗口。
- NoteMD 嵌入式流程:
- NoteMD 作为嵌入能力存在于 Tauri 和 Godot 两条前端链路中,不应作为独立桌面窗口运行。
- 在 Tauri 中点击文件/文件夹/保存 Browse 应触发原生选择器并回填路径。
- PDF 导入规则:
- PDF 需要先用 Mineru 转换成 Markdown,再导入 NoteMD。