I read The AI Engineer's Guide to Surviving the EU AI Act by Larysa Visengeriyeva, and it was great. It sparked my interest to learn more about the Act and also I realised it's a good candidate for a graph data model.
- the CSVs (articles.csv, annexes.csv, recitals.csv) contain the scraped documents; also given as txt files
- scraping is in the
scrape_data.ipynb
- some of the Cypher code is in
interact_with_neo4j.ipynb
, there are not many nodes (~300 for the 3 main documents) so this is fairly light - to set up Neo4j locally, run
docker-compose up -d
(username & pwd are in thedocker-compose.yml
)