[Python] tips for nlp-informed perspective on self-referentiality in documents [closed]

I am experimenting with using nlp as a basic reading-assistive tool for legislation. These documents are heavily self-referential, to give one example, in the sense that they "define," or pretend to define, key terms, but then also reference those definitions, or avoid referencing them. I am thinking about a 'document map' to follow the journey of a term through a document. can anyone who knows more about the current state of NLP suggest some helpful approaches/widgets/functions? I have tried some of these https://spacy.io/universe but I like to keep things simple, explicit, explainable, pure-python-y when possible. I don't just want to load a huge LLM and throw it at a doc or corpus and call that insight...I more want to think through how the nlp techniques can help or hinder more traditional conceptions of "reading." How can I use nlp to write something that follows a key term through a doc in a way the supports advocacy?

Continue reading...