Skip to content

Introduction🔗︎

Welcome to the readnext documentation! ☀️

The readnext package provides a hybrid recommender system for computer science papers. Its main objective is to suggest relevant research papers based on a given query document you might be currently exploring, streamlining your journey to discover more intriguing academic literature.

Check out the source code of the readnext package on GitHub. The motivation, theoretical background and results of this project are presented in the thesis. The accompanying repository contains the LaTeX source code as well as the code to generate all figures and diagrams in the thesis.

Overview🔗︎

  • Installation: Get started with the package in no time with a simple installation guide.

  • Overview: Obtain a high-level perspective of the project with a concise schematic overview.

  • Setup: Build the foundation of inference by downloading the required data and model files. Set environment variables, run user scripts to precompute data required for inference, and get tips for an efficient development workflow.

  • Inference: Get instructions on how to employ the package for inference, enabling you to extract recommendations for any query document of your choice.

  • Reproducibility: Learn how to reproduce the evaluation results presented in the thesis.

  • Customization: Customize the package to your needs by adding or replacing tokenizers, language models or evaluation metrics.