All projects

Python 3.11+ · source application

Gaussian PREP

Build exam practice from a question dataset, check the answers, and export a question sheet with a separate key. The reviewer still needs reviewing.

What it does

  • Plans, generates or adapts, verifies, reviews, and audits practice material.
  • Checks supported mathematical answers with SymPy.
  • Exports PDF or DOCX and an offline HTML summary, with saved runs and recovery.

Compatibility

Requires Python 3.11+, a JSON, JSONL, or CSV question dataset, and a configured model provider or local Ollama server.

Windows launch scripts are included. The release is not validated on macOS or Linux.

This is a source package, not a prebuilt executable. A LaTeX engine is optional; the PDF export has a built-in fallback.

Download & installation

Available as a documented source installation. The download contains the public repository's source files.

Source code · requires the setup described below · not a prebuilt executable

  1. Download and extract the source ZIP. From the project directory, run setup.ps1 with your Python path, then START.cmd or run.ps1.
  2. For an existing environment, run python -m pip install . and then python -m gaussian_prep.
  3. Use the inspect command to check your dataset without credentials or model calls. Review generated questions and answers before using them.

Limitations

A dataset is not bundled. Automated checks do not guarantee correct answers, exam alignment, originality, or calibrated difficulty. Hosted model calls can incur provider charges.

Documentation

All projects