BeeSwarm

Model

Run instructions

1. cd model.
2. If you want to run natively using your GPU natively go to step 3, otherwise skip to step 4 to run in Docker.
3. Run ./run_model_mac.sh or ./run_model_* depending on your operating system.
4. To run it in Docker you can run docker compose up --build

Modules

The models folder contain a modules folder which currently only supports access to the internet. This is not implemented yet but only implemented as a theoretical utility. It should be fully implemented.

It can be started using npm i followed by npm start. Currently it only supports a single query hard coded in the internet.ts. This should be modified to allow for any query via the api, and the chatbot should be able to access this endpoint somehow.

Load balancer

Run instructions

1. cd lb.
2. If you want to run locally for testing, skip to step 4, otherwise continue with step 3.
3. docker compose up --build.
4. If you want to test the frontend run cd ui and npm run dev or for the api cd api and npm start.

The API is used directly by the script and should not be started manually.