wzyqaq0v0/autocoursenote

AutoCourseNote is a Python tool that turns a folder of course materials into a polished, textbook-style document. You drop your slides, textbooks, assignments, past exam papers, and images into a folder, point the tool at that folder, and it works through several stages automatically: classifying the files by type, extracting text, drafting notes from each source, reviewing and expanding the draft, writing a full textbook version, refining each chapter, and finally exporting a structured Markdown file and optionally a PDF. The system is built for students who have all their course materials saved locally but want a well-organized reference document rather than a pile of separate files. It is optimized for Chinese-language courses, handling mixed Chinese and English terminology, formulas, and lecture text. For science and engineering courses it emphasizes worked derivations and complete problem solutions, for humanities courses it emphasizes key points, memorization outlines, and essay frameworks. Each stage in the pipeline is handled by a separate agent that calls an AI language model through an OpenAI-compatible API. You can connect DeepSeek, OpenAI, or any compatible service by setting your API key as an environment variable. The key is never written into the configuration file. The tool does not scrape any website or handle logins: it only processes files you have already gathered yourself. Configuration lives in a single TOML file where you set the course name, the materials folder, the output prefix, and the language. A dry-run flag lets you preview the pipeline steps before starting a full run. Outputs land in a structured folder containing parsed text, drafts, the polished Markdown, a LaTeX source file, and a compiled PDF if you have XeLaTeX installed. Citation records and self-check reports are also written so you can trace where each piece of content came from. The README is written in Chinese, and the AI prompts and subject-aware tuning are oriented toward Chinese academic contexts, though the file parsing itself works with documents in any language.