transient-display-data for Jupyter Lab

This is a JupyterLab extension that allows JupyterLab to receive messages in a new transient_display_data type and display them in the console window of the associted notebook.

As summarized here, the transient display data messages are designed to send messages that are transient in nature and will not be displayed and saved with the notebooks. Such messages include but not limited to status or progress information for long calculations, and debug information. This message type is identical to display_data in content so you only need to use message type transient_display_data instead of display_data to mark the message as transient.

This new message type is currently under review. However, even before it is officially accepted, kernels can send messages of this type safely because all Jupyter clients ignore messages of unknown types, and JupyterLab with this extension will be able to display them. An an example, the SoS Kernel uses this message type to send progress information during the execution of the SoS workflows.

How to install

• If you are using conda version of JupyterLab, you can install this extension with command

  conda install jupyterlab-transient-display-data -c conda-forge

• Otherwise you can install the transient-display-data extension using command

  jupyter labextension install transient-display-data

  or go to the extension manager, search for transient-display-data, and install.

How to use transient_display_data

After you installed this extention, you can test it by

1. Create a notebook with Python 3 kernel
2. Right click and select New Console for Notebook to create a console window
3. Right click on the console window and you select Show Transient Message.
4. In the Python notebook, enter

kernel = get_ipython().kernel
kernel.send_response(kernel.iopub_socket,
                     'transient_display_data',
                     {
                         'data': {
                             'text/plain': 'I am transient'
                         }
                     }
                    );

and a message I am transient should be displayed in the console window.

5. If you are interested in trying SoS Notebook, you can click this link to start a JupyterLab session on our live server. You can create a new notebook with SoS kernel, open a console window, and execute for example a trivial workflow

%run
[1]
[2]
[3]

You can see progress messages in the console window.