RepoScope: A Multi-Granularity Query-by-Example Framework for Github Repository Search

This repository contains the replication package for RepoScope. The package includes: reposcope_pipeline.ipynb: Contains the code and step-by-step procedures used in the RepoScope study, specifically implementing Repository Embedding and Retrieval and Clustering and Visualization steps of RepoScope. all_repos.json: Includes a list of 7143 repository information with their metadata necessary for executing the code in the Jupyter Notebook. Requirements: The notebook includes a list of libraries that need to be installed for the code to run properly. Accessing certain models may require a Hugging Face API token before running the code.