LLM Agents for Smart City Management: Enhancing Decision Support through Multi-Agent AI Systems

The repository contains code and data for the LLM-based city management research.

Build an image for the Pipeline with RAG and start the container:

Create the $NSS_NPA_TOKEN environment variable on the server

Pull changes:

cd /var/essdata/llm/project/BIAM-Urb
git checkout <required_branch>
git pull

Create config.env in the root of the project:

LLAMA_URL=<url>
LLAMA_FC_URL=<url>
ENDPOINT_LISTINGS_URL=<url>
ENDPOINT_CITY_URL=<url>
ENDPOINT_METRICS_URL=<url>
ENDPOINT_PROVISION_URL=<url>
ENDPOINT_TABLES_URL=<url>

Rebuild the image and run the container

docker container stop llm_city_app-container
docker container rm llm_city_app-container

# use --no-cache key if the entire image needs to be rebuilt (e.g. dependencies changed)
docker build -t llm_city_app --build-arg NSS_NPA_TOKEN=$NSS_NPA_TOKEN -f docker/app/Dockerfile --no-cache .
 
docker run -d --restart always -p <port>:80 --name llm_city_app-container llm_city_app

Test the Pipeline on the server:

In the terminal, send a request:

curl -v POST http://<ip>:<port>/question -H 'Content-Type: application/json' -d '{"question_body": "What are the problems of demographic development of St. Petersburg?"}'

In the browser:

http://<ip>:<port>/docs/

The application logs can be checked on the server:

docker logs --follow --timestamps llm_city_app-container

Test api locally:

Start the server (from the project root):

fastapi run main.py --proxy-headers --port 80

In the terminal, send a query:

curl -v POST http://0.0.0.0:80/question -H ‘Content-Type: application/json’ -d ‘{’question_body‘: “What are the problems of demographic development of St. Petersburg?”}’

In the browser:

http://localhost/docs#/

Run ChromaDB and the embedding model on the server

Copy compose.yaml to the server

Execute the command:

docker compose up