Overview overview

Rerun can be used directly within Jupyter notebooks. Many additional environments beyond Jupyter are supported such as Google Colab or VSCode.

You need the separate support package rerun-notebook to use this feature. Typically this is installed using:

pip install "rerun-sdk[notebook]"

In order to show a Rerun Viewer inline within a notebook, you can call:

rr.init("rerun_example_notebook")

rr.log(...)

rr.notebook_show()

This will show the contents of the current global recording stream. Note that the global stream will accumulate data in-memory. You can reset the stream by calling rr.init again to establish a new global context.

As with the other stream viewing APIs (rr.show, rr.connect_grpc, rr.spawn), you can alternatively pass a specific recording instance to notebook_show

rec = rr.RecordingStream("rerun_example_notebook_local")

rec.log(...)
rec.notebook_show()

Running in Jupyter running-in-jupyter

The easiest way to get a feel for working with notebooks is to use it.

First, install the requirements (this includes Jupyter, the Rerun SDK, and the notebook support package)

pip install -r requirements.txt

Then, open the notebook

jupyter notebook cube.ipynb

Follow along in the browser that opens.