Wetsy is an idea for a "Legal Search" application aiming to help the prosecusion of climate criminals in the Netherlands.
STATUS: This a proof of concept!
While many oportunities exist to prosecute climate related and environmental offences, the legal basis is often a combination of laws which is not always obvious without extensive expertise. With the use of natural language processing, a specialized search engine can help support the work of constructing the legal case.
We use the same data as the website wetten.nl. It can be downloaded using the python package wetsuite. The data is preprocessed and ingested into a Chroma DB instance, using the embedding model robbert-2022-dutch-sentence-transformers (developed by the Netherlands Forensic Institute). Via simple UI and supporting backend api, a description of the criminal activity is matched to the relevant laws by semantic search.
docker compose up
and check http://localhost:8000.
brew install uv just
uv sync
just test
just run
uv run python data/upload.py
This requires Chroma DB to be reachable on the url in the environment variable WETSY_CHROMA_HOST.
(Currently, only Het wetboek van strafrecht is ingested).