Installing pywwt

Try without installing

You can try out the pywwt Jupyter widget without installing anything locally by using one of our example notebooks, served from the cloud! Click here to launch our examples using the great MyBinder service. These notebooks are defined in the pywwt-notebooks repository on GitHub, which has further links and information about these samples.

Install pywwt with pip

You can also install the latest release of pywwt using pip:

pip install pywwt

While pip will automatically install most dependencies as needed, to use the Qt widget you will need to install PyQt and PyQtWebEngine, or PySide separately. See the full list of dependencies below.

Set Up pywwt’s JupyterLab Integration

To use pywwt in an interactive computing environment, we recommend combining it with JupyterLab. It’s worth noting that the JupyterLab web application is a separate thing than just “Jupyter,” the lower-level system upon which it is built. Learn how to install JupyterLab here.

Once you’ve installed JupyterLab, you need to install a couple of other supporting components beyond pywwt to get the best experience: the WWT Kernel Data Relay (KDR) extension, the WWT JupyterLab Extension, and the JupyterLab widget manager extension. The following commands should ensure that everything is set up:

pip install wwt-kernel-data-relay  # or use conda
jupyter labextension install --no-build @jupyter-widgets/jupyterlab-manager
jupyter labextension install --no-build @wwtelescope/jupyterlab
jupyter lab build

If the installation is successful, then the next time you start up JupyterLab the “Launcher” display should now contain an AAS WorldWide Telescope icon:

Screenshot of JupyterLab launcher with WWT icon

Clicking on this icon should open up the WWT research app, which you can then control using pywwt, via pywwt.jupyter.connect_to_app().

Set up the pywwt Jupyter Widget

If you are using a Jupyter notebook, either within JupyterLab or in “vanilla “Jupyter, you may also wish to set up pywwt’s “Jupyter widget” support. This integration should be set up automatically if you install pywwt through one of the standard methods, but it can be somewhat finicky. As with the JupyterLab support, you should install the WWT Kernel Data Relay (KDR) extension:

pip install wwt-kernel-data-relay  # or use conda

To check whether the widget is working, run the following Python commands inside of a Jupyter notebook:

from pywwt.jupyter import WWTJupyterWidget
wwt = WWTJupyterWidget()
wwt  # just "print" this variable on its own line

If everything is working, you should see a WWT window open up inside your notebook, like so:

Screenshot of Jupyter notebook with WWT widget

If you get any other result, some troubleshooting may be necessary.

Troubleshooting the Jupyter Integration

The first thing to check is the “notebook extension”. Run:

jupyter nbextension install --py --sys-prefix pywwt
jupyter nbextension enable --py --sys-prefix pywwt
jupyter nbextension list

These commands should not report any error messages, and should indicate at the end that the pywwt extension is installed, enabled, and validated. The ipyevents extension should also be available. Unfortunately, if there is a problem at this stage, there might be a lot of possible reasons. For help, try filing an issue on our GitHub with a copy-paste of the output from the commands above.

Next are the Jupyter “server extension”s, which are controlled similarly:

jupyter serverextension enable --py --sys-prefix pywwt
jupyter serverextension list

Here too, the commands should report that the pywwt and wwt_kernel_data_relay extensions are installed and enabled, without any apparent errors.

The recommended way to use pywwt inside of JupyterLab (not vanilla Jupyter) is with the help of the separate WWT JupyterLab Extension:

jupyter labextension install --no-build @wwtelescope/jupyterlab
jupyter labextension list

Finally, if you wish to use the specific combination of the pywwt widget (not app) inside of JupyterLab (not vanilla Jupyter), you may also need to ensure that pywwt is installed as its own, different, “lab extension”, along with additional helpers:

jupyter labextension install --no-build @jupyter-widgets/jupyterlab-manager
jupyter labextension install --no-build ipyevents
jupyter labextension install --no-build pywwt
jupyter labextension list

If the above commands seem to be OK but report that a “build” is needed, that is OK:

jupyter lab build

Dependencies

If you install pywwt using pip or conda as described above, any required dependencies will get installed automatically (with the exception of PyQt/PySide if using pip). For the record, these dependencies are as follows:

In addition, if you want to use the Qt widget, you will need:

For the Jupyter widget, you will need:

Installing the developer version

If you want to use the very latest developer version version, you can clone this repository and install the package manually (note that this requires npm to be installed):

git clone https://github.com/WorldWideTelescope/pywwt.git
cd pywwt
pip install -e .

If you want to use the Jupyter widget, you will also need to run:

jupyter nbextension install --py --symlink --sys-prefix pywwt
jupyter nbextension enable --py --sys-prefix pywwt
jupyter nbextension list  # check that the output shows pywwt as enabled and OK
jupyter serverextension enable --py --sys-prefix pywwt
jupyter serverextension list  # check that the output shows pywwt as enabled and OK

And if you additionally want to use the widget in JupyterLab, run:

jupyter labextension install frontend
jupyter labextension list  # check that the output shows pywwt as enabled and OK

Besides pywwt, the WWT Kernel Data Relay server extension and the WWT JupyterLab extension are both strongly recommended, but not technically necessary. If you want to use pywwt as a widget in JupyterLab, the @jupyter-widgets/jupyterlab-manager and ipyevents lab-extensions must be installed and enabled.