The script knowledge_graph/run_ontoGPT.sh can be used to run OntoGPT over a set of text files in a directory. The directions to set up the dependencies are found in the OntoGPT repo. In short, we cloned the repository, copied the desiccation schema from knowledge_graph/schema to ontogpt/src/ontogpt/templates, set up a conda environment with python 3.9 (separate from the one described in the main README of this repo) and used pip install -e . from the root directory of ontogpt to set up the package for our usage here.