haolpku/K12-KGraph ? reverse-engineered prompt

Build me a Python project that can take official K 12 textbook PDFs and turn them into a curriculum aligned knowledge graph for school science subjects like math, physics, chemistry, and biology.

I want the full pipeline to read the textbooks, pull out concepts and relationships, organize them by book, subject, and overall graph, and then generate two things from that graph, a multiple choice benchmark and a small instruction response training set for educational LLMs. Please include a simple way to run the pipeline from config files, a demo with sample data, and an evaluation script that can score a model on the benchmark.

Make it easy to swap in a different textbook set or model provider if needed, and add clear README style instructions so someone can get started without digging through the code. If you need to check current docs for any external tools, go ahead and look them up online.