Explore with notebooks

The notebooks in this section provide examples for exploring the GKM Toolkit and working with bundles. The source notebooks live in the project's notebooks/ directory.

Available walkthroughs

The walkthroughs below are rendered examples, so you can read them directly in your browser without installing anything. The bundle examples are small, focused subsets of a producer's actual data. Published bundles contain the complete dataset released by a producer.

Explore GKM Toolkit features with CIViC bundle examples: learn the Toolkit by loading two CIViC bundle examples, inspecting their collections, and following bundle-local references.
Explore the public bundle repository — discover public resources, retrieve a producer's complete published bundle and its schema, load them into GKM models, and handle repository errors.
Validate a bundle: see how the Toolkit checks a bundle and what happens when something is wrong.

Run the notebooks with MyBinder

Binder provides a ready-to-use computing environment for running the notebooks directly in your browser, with no local installation required.

You can launch the notebooks on Binder here.

Once launched, select GKM Toolkit Kernel as the notebook kernel.

Run the notebooks locally

If you want to run the notebooks yourself, follow the Installation steps to clone the repository, create the virtual environment (.venv), and activate it.

Notebook dependency

If you did not install the notebooks optional dependency during setup, run this command from the root of the cloned project (gkm-starter-kit/), with the virtual environment activated:

python3 -m pip install "ga4gh.gkm[notebooks]" --pre

--pre is provided since this project uses pre-release versions of the GKM reference implementations

Open a notebook

We recommend either of the following options for opening a notebook.

JupyterLab

In the terminal, activate the project's .venv virtual environment:
```
 source .venv/bin/activate
```
In the activated virtual environment, start JupyterLab:
```
 python3 -m jupyterlab
```
JupyterLab usually opens automatically in a new browser window. If it does not, open the URL shown in the terminal.
Open a notebook from the notebooks/ directory.

VS Code

Install the Python extension and Jupyter extension.
Open the project root in VS Code and open a notebook from the notebooks/ directory.
Select Select Kernel in the upper-right corner and choose the Python interpreter from the repository's .venv virtual environment.