Deprecated. New repository is moved to: [https://github.com/INTEND-Project/inSwitch] LLM agents for cognitive computing continuum.
We recommend using VS Code + devcontainer. You need to install the devcontainers extension to VS Code. After that, when you open this workspace in Code, there will be a pop-up window asking if you want to re-open it in devcontainer. If you missed it, use cmd+shift+p, and select "Dev Containers: Reopen in Container". Make sure you have a docker environment running.
Once the workspace is reopened inside a container (to make sure it is inside a container, launch a terminal and check the user name and node name), you may need to manually install two extensions to the "new VS code", i.e., Python (ms-python.python) and Jupyter (ms-toolsai.jupyter).
You need a valid token for OpenAI API, stored in ./openai.credential. Other LLM models will be supported later.
- ./inswitch: generic components for inSwitch
- ./usecases: data, scripts, components, etc., for different use cases
Most of the running examples and demos are in Jupyter Notebooks. To run them, you need to select a kernal - choose a python environment called "base (Python 3.12.7)". The version may vary. It is the under path "/opt/conda/bin/python".
If you face problem running python scripts, you may also try to select this same "base" python as the Python Interpreter (this can be done also via cmd+shift+p).
All the use cases can be found in usecase. Four each use case, there is a README.md file for further information.
So far, we have one simplified example running. You can check the snapshot here