Wrongbook Suite

A local-first toolkit for learning from mistakes — from quant interview prep to AI-powered knowledge graphs.

---

🌐 Live Demo

Try the apps directly in your browser — no installation needed:

• QR Wrongbook (recommended)
  A feature-rich, local-first wrong question notebook tailored for quant & QR interview prep.

• AI Mistake Notebook (experimental)
  A lightweight mistake analyzer for code/math/ML with automatic concept extraction and knowledge graph insights.

---

📖 Overview

You can use either of the two HTML files included in this repository:

• qr_wrongbook_v1.1.html → the recommended, feature-rich version for quant & QR interview prep.
• ai-mistake-notebook-v1.1.html → the experimental version focusing on code/math/ML with knowledge graph insights.

Simply download the file you want, then double‑click it to open in your browser — no installation needed. This project combines two complementary tools:

• QR Wrongbook — a rich, feature-complete wrong question notebook designed for quant & QR interview preparation.
• AI Mistake Notebook — a lightweight, extensible notebook for code, math, and ML mistakes, featuring automatic concept extraction and a knowledge graph.

Both are single-file, offline-first apps: no backend, no tracking, just open in your browser.

---

🔹 Module A: QR Wrongbook

A local-first “wrong question notebook” for quant & QR interview prep.

Features

• 🗃️ Local-first — data stored entirely in your browser (IndexedDB), works offline
• 🧠 Rich metadata — category (Code / Math / Behavior / Machine Learning), difficulty (Easy → Deadly), mastery (Not at all → Mastered)
• 🖼️ Screenshot support — drag & drop, paste, or file picker
• 🧪 Focus practice mode — full-screen practice view, collapsible Hint/Solution, timer, and logs
• 💬 Comments — add & delete with confirmation
• 📊 Stats dashboard — study time, sessions, 30-day charts
• 🔎 Search & filter — full-text, multi-select filters, sorting
• 🎨 Themes — Light / Dark / Green gradient header
• 🔐 Backup/restore — Export/Import JSON

Quick Start

# Clone
git clone https://github.com/yw562/qr-wrongbook.git
cd qr-wrongbook

# Local use
# Just double-click index.html in Chrome/Firefox/Edge/Safari

Deploy on GitHub Pages

1. Push index.html + README.md to the main branch.
2. Settings → Pages → Deploy from branch → main / root (/).
3. Live at https://yw562.github.io/qr-wrongbook/.

---

🔹 Module B: AI Mistake Notebook

A lightweight notebook for code, math, and ML mistakes with knowledge graph insights.

Features

• Input problem + optional answer/attempt.
• Auto/manual type detection (code, math, ml).
• Auto-generate concepts, insight, explanation, solution sketch.
• Unlimited storage in localStorage.
• Knowledge graph (node frequency & color by type).
• Import/Export JSON.
• Optional: integrate local Ollama LLM for richer analysis.

Example Concepts

• Math: size-biased sampling, Horvitz–Thompson, unbiased estimator
• Code: enumerate, indexing, off-by-one
• ML: overfitting, ROC–AUC, cross entropy

Optional LLM (Ollama)

# Install Ollama
brew install ollama
ollama pull llama3.1:8b-instruct
ollama serve

Check Use Local LLM (Ollama) in UI.

---

📥 Bulk Import / Export

The QR Wrongbook app supports exporting and importing problems as JSON files, so you can back up your data or add many questions at once.

• Export → Use the Export button in the app to download all your problems as a JSON file.
• Import → Use the Import button and select a JSON file in the correct format.

Ready-to-use helpers

• BULK_IMPORT.md — full guide on how to structure import files (schema, examples, and an LLM prompt template for auto-generation).
• starter_import.json — a sample JSON file with one demo problem. You can copy, replace, or expand it to create your own batch of problems.

Typical workflow

1. Write problems in plain text (Math, Code, ML, Behavioral).
2. Convert them into the JSON format (see BULK_IMPORT.md or generate with an LLM).
3. Save the JSON as a file.
4. Click Import in the app and select it.
5. Done 🎉 — all problems are added locally in your browser.

---

🔮 Roadmap

• Richer LLM support (more models, better prompts)
• Editable Hint / Solution / Note fields
• Graph interactivity: click node → show related problems
• Spaced repetition scheduler based on difficulty
• Progress charts (error frequency over time)

---

🤝 Contributing

Contributions, issues, and feature requests are welcome!

• Fork and make a PR.
• Open issues for ideas/bugs.
• Suggestions for heuristics, prompts, or visualization especially appreciated.

---

📚 Citation

If you use this project, please cite:

APA:
Wang, Y. (2025). Wrongbook Suite: Local-first mistake notebooks for quant prep and AI error analysis. GitHub. https://github.com/yw562/qr-wrongbook

BibTeX:

@misc{wang2025wrongbook,
  author       = {Yueyi Wang},
  title        = {Wrongbook Suite: Local-first mistake notebooks for quant prep and AI error analysis},
  year         = {2025},
  howpublished = {GitHub},
  url          = {https://github.com/yw562/qr-wrongbook}
}

---

📜 License

Released under CC BY-NC 4.0:

• Share — copy & redistribute the material
• Adapt — remix, transform, build upon it
• Attribution — give credit
• NonCommercial — no commercial use without permission

For commercial use, please contact the author.