nlp-tlp/MaintKG ? reverse-engineered prompt

Build me a tool that takes maintenance work order data from a CSV and turns it into a knowledge graph I can browse in Neo4j.

I want to drop my CMMS file into an input folder, set a few column names and database settings in a simple config file, then run one command to process everything. The app should clean up the text, extract useful entities and relationships from messy maintenance notes, normalize obvious terms, and save the results into Neo4j so I can query the graph later. It should also write any run outputs to a timestamped output folder so I can inspect what happened.

If there’s a pretrained model or checkpoint needed for the text extraction part, include a way to download or load it automatically. Use the README instructions as a guide, and look up current docs online if you need to.