AnkiCardProducer

AnkiCardProducer is an intelligent, high-performance tool designed to convert deep learning textbooks (PDF or Markdown) into high-quality Anki flashcards. It leverages modern LLMs (like Gemini/OpenAI) to generate rigorous, bilingual (Chinese/English) Q&A pairs, preserving technical terminology and mathematical formatting.

Key Features

• Multimodal Input:
  • PDF Parsing: Robustly converts PDF textbooks to Markdown, preserving structure, tables, and extracting images.
  • Markdown Support: Directly processes existing Markdown files.
• LLM-Driven Generation:
  • Uses advanced models (Gemini 1.5/2.0, GPT-4) for context-aware flashcard generation.
  • Generates Chinese explanations while strictly referencing English technical terms.
• High Performance:
  • Parallel Processing: Multi-threaded LLM requests for rapid generation.
  • Resume Capability: Tracks progress at the section level. If interrupted (API limits, network issues), simply re-run to resume exactly where you left off.
• Smart Image Handling:
  • Extracts figures from PDFs.
  • Uploads to Cloudflare R2 (optional) or handles local references.
  • Filters out small decorative icons.

Prerequisites

• Python 3.8+
• An LLM API Key (Google Gemini or OpenAI compatible)
• (Optional) Cloudflare R2 credentials for image hosting.

Installation

1. Clone the repository:

   git clone https://github.com/gaivrt/AnkiCardProducer.git
   cd AnkiCardProducer

2. Install dependencies:

   pip install -r requirements.txt

3. Configuration: Copy .env.example to .env and configure your keys:

   # LLM Settings
   OPENAI_BASE_URL=https://generativelanguage.googleapis.com/v1beta/openai/
   OPENAI_API_KEY=your_api_key_here
   MODEL_NAME=openai:gemini-1.5-pro
   
   # R2 / S3 Settings (Optional)
   R2_ENDPOINT_URL=https://...
   R2_ACCESS_KEY_ID=...
   R2_SECRET_ACCESS_KEY=...
   R2_BUCKET_NAME=...
   R2_PUBLIC_URL=...

Usage

1. Generate Flashcards from Markdown

If you already have a Markdown version of your book:

python src/main.py --mode md --files "path/to/book.md" --prompt src/prompts/deep_learning.txt --output output/dl_cards.csv

2. Generate Flashcards from PDF

This will first convert the PDF to Markdown (extracting images) and then generate cards:

python src/main.py --mode pdf --files "path/to/book.pdf" --prompt src/prompts/deep_learning.txt --output output/dl_cards.csv

3. Output

The tool generates a CSV file in output/ ready for import into Anki (or further processing).

• Format: Front, Back, Image, Tag

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

License

MIT