SuperAPI AI Directory

learning-notes-automation

General & other

View repository

Updated 10-15 17:34

About this skill

Learning Note Automation

Core Value

Passive learning → Active memory: Watching videos ≠ learning, generating flashcards = remembering

Input Sources

Input Type Tool Output Format
YouTube video youtube-transcribe Transcript text
Bilibili video bilibili-transcribe Transcript text
Podcast podcast-transcribe Transcript text
Douyin douyin-transcribe Transcript text
WeChat Official Account article wechat-article-ingest Markdown

Workflow

Phase 1: Content Transcription

Reuse this repository's transcription skill (relative path, replace with your installation location as needed):

# YouTube(自动翻译)
python3 ../youtube-transcribe/scripts/transcribe.py "https://youtube.com/watch?v=xxx" -o ./

# B站
python3 ../bilibili-transcribe/scripts/transcribe.py "https://bilibili.com/video/xxx" ./

# 播客
python3 ../podcast-transcribe/scripts/transcribe.py "https://xiaoyuzhoufm.com/episode/xxx" ./

The output is a transcription Markdown, used as the input for Phase 2.

Phase 2 + 3: Knowledge Point Extraction + Flashcard Generation (in one step)

Hand the transcript to make_notes.py, automatically extract knowledge points and generate Anki-compatible flashcards:

export DEEPSEEK_API_KEY=***
python3 scripts/make_notes.py 转录稿.md --output ./notes --max-cards 20
# 产出:<标题>-学习笔记.md(核心要点+闪卡+关联知识) 和 <标题>-闪卡.csv(直接导入 Anki)

The script uses DeepSeek to output structured JSON, then renders it into the note and flashcard formats specified below. The following are design notes for each dimension, for reference when understanding and manually fine-tuning.

Extraction Dimensions:

  1. Core Concepts(must master)

    • Definition
    • Principle
    • Application Scenarios
  2. Key Facts(need to memorize)

    • Data
    • Timeline
    • People/Companies
  3. Methodologies(can be reused)

    • Steps
    • Framework
    • Best Practices
  4. Quotable Lines/Insights(worth quoting)

    • Incisive Summaries
    • Unique Viewpoints

Phase 3: Flashcard Generation

Flashcard Format(Anki-compatible):

## 闪卡类型

### 1. 概念卡(Cloze Deletion)
Q: {{c1::Transformer}} 架构的核心机制是 {{c2::自注意力(Self-Attention)}}
A: Transformer, 自注意力(Self-Attention)

### 2. 问答卡(Basic)
Q: 什么是 RAG?
A: Retrieval-Augmented Generation,检索增强生成。通过检索外部知识库来增强 LLM 的回答能力,解决幻觉问题。

### 3. 对比卡(Comparison)
Q: Fine-tuning vs RAG 的区别?
A: 
| 维度 | Fine-tuning | RAG |
|------|-------------|-----|
| 成本 | 高(需要训练) | 低(只检索) |
| 更新 | 需要重新训练 | 实时更新 |
| 适用 | 特定任务 | 知识问答 |

### 4. 步骤卡(Process)
Q: 如何构建一个 RAG 系统?
A: 
1. 文档分块(Chunking)
2. 向量化(Embedding)
3. 存入向量数据库
4. 检索相关片段
5. 拼接 Prompt
6. LLM 生成回答

Phase 4: Knowledge Graph Update

Entity Extraction:

  • People (Who)
  • Concepts (What)
  • Tools/Products (Tool)
  • Methodologies (How)
  • Time (When)

Relationship Mapping:

  • 发明了: Person → Concept/Tool
  • 属于: Concept → Domain
  • 替代了: New tool → Old tool
  • 依赖于: Concept → Concept

Output to Knowledge Base:

---
title: [概念名]
type: knowledge-card
platform: learning
tags: [学习, AI, ...]
source: [[视频/播客链接]]
created: YYYY-MM-DD
---

# [概念名]

## 定义
一句话定义

## 核心要点
1. ...
2. ...
3. ...

## 关联概念
- [[概念A]]:关系说明
- [[概念B]]:关系说明

## 应用场景
- 场景1:...
- 场景2:...

## 闪卡
(Anki 格式的闪卡内容)

Output Format

1. Anki Flashcard File

# 符号分隔格式,可直接导入 Anki
"问题","答案","标签"
"什么是 Transformer?","一种基于自注意力机制的深度学习架构,广泛用于 NLP 任务","AI,深度学习"

2. Learning Note Markdown

# [视频标题] - 学习笔记

**来源**:[链接]
**日期**:YYYY-MM-DD
**时长**:XX 分钟
**领域**:AI / 半导体 / ...

---

## 📝 核心要点

### 要点 1:[标题]
- 摘要:...
- 重要性:⭐⭐⭐⭐⭐

### 要点 2:[标题]
- 摘要:...
- 重要性:⭐⭐⭐⭐

---

## 🃏 闪卡(XX 张)

### 概念卡
1. Q: ... A: ...
2. Q: ... A: ...

### 问答卡
1. Q: ... A: ...

---

## 🔗 关联知识

- [[概念A]]:...
- [[概念B]]:...

---

## 💡 个人思考

(基于内容的个人见解和应用想法)

3. Knowledge Base Entry

Save to:wiki/📖 行业研究/[领域]/[概念名].md

Cron Configuration (Optional)

# 每周日整理本周学习内容
schedule: "0 20 * * 0"
prompt: "整理本周的所有学习笔记,生成周学习报告"

Usage Example

用户:帮我把这个视频做成学习笔记
https://youtube.com/watch?v=xxx

Agent:
1. 转录视频
2. 提取 15 个知识点
3. 生成 20 张闪卡
4. 更新知识图谱
5. 输出 Anki 文件 + Markdown 笔记

Anti-patterns

  • ❌ Do not only extract without digesting, there must be personal reflection
  • ❌ Do not generate too many flashcards, quality > quantity (15-30 per single video)
  • ❌ Do not ignore knowledge connections, isolated knowledge is easily forgotten
  • ✅ Flashcards must be understandable independently (not dependent on context)
  • ✅ Each flashcard has only one knowledge point (atomicity)
  • ✅ Prioritize generating "why" and "how" cards, rather than "what"