EHALLM is a collection of R functions and scripts for interacting with OpenAI and other Large Language Models (LLM's). Right now the focus is on OpenAI's ChatGPT. The pipeline will eventually include:
- Automated academic literature search
- OCR and text extraction to serialize PDF text, tables, and figures
- Function calling to enforce a structured response from an LLM
- Prompts designed to extract information from literature corpora
- A pipeline to submit requests and format, clean, and structure
ehallm depends on a fork of the openai r package to enable function calling. As of January 2024 function calling is still not available in the main openai package. The appropriate fork can be installed using the following command:
if (!require(remotes)) install.packages("remotes")
remotes::install_github("nhardtskim/openai")