Implementing PageIndex for Tabular Code Datasets in a GitHub RAG Chatbot

I am building a chatbot that looks at a GitHub repo and answers questions about it. To build it, I'm using the code_search_net dataset.

My question is: how can I implement PageIndex on this tabular dataset, which contains repository_name, func_path_in_repository, func_name, whole_func_string, language, func_code_string, func_documentation_string, and func_code_url?

All the examples I've found are about PDFs, but RAG isn't only about PDFs.