SciLifeLab Workshop: AI Agents in Life Sciences

Presentation:

• Section 1
• Section 2

Setup Before the Workshop

Make sure your environment is ready before you begin:

1. Open a terminal and clone Github repo:

git clone https://github.com/DinhLongHuynh/SciLifeLab_agent_workshop

2. Navigate to the project directory:

cd SciLifeLab_agent_workshop

Option 1: Run workshop in Docker

Requires Makefile support and Docker engine on your machine.
If you don't have Makefile support on your machine, you can type the commands in the Makefile by hand.

1. Build the image

make build

2. Start Jupyter

make run-jupyter-lab

and then access the localhost:port address printed in the terminal in your browser.

or to start a separate shell (needed for some experiments in the workshop, keep make run-jupyter-lab running in the background):

make run-shell

and then cd /repo to find the source code.

Option 2: Run workshop natively

1. Create and activate a virtual environment:

python3 -m venv .venv
source .venv/bin/activate

2. Install the required dependencies:

pip install -r requirements.txt

3. Register your virtual environment as a Jupyter kernel:

python -m ipykernel install --user --name=SciLifeLab_venv --display-name="SciLifeLab_workshop_kernel"

4. Start Jupyter notebook:

jupyter notebook

5. From the top-right corner of the Jupyter notebook, select and change the kernel:

Example: SciLifeLab_agent_workshop/Section_1_LangGraph/langgraph_lab.ipynb

click Python 3 (ipykernel) → select Select kernel for: "langgraph_lab.ipynb" → from the drop-down menu, choose SciLifeLab_workshop_kernel → check Always start the preferred kernel → click Select

You can find the notebooks in the following directories:

SciLifeLab_agent_workshop/Section_1_LangGraph/

SciLifeLab_agent_workshop/Section_2_MCP/MCP_scratch/

---

You can now start exploring and using the notebooks

Happy learning! 😊