pypowsybl-jupyter 0.2.0

Widgets for pypowsybl in Jupyter notebooks

pypowsybl-jupyter

Widgets for PyPowSyBl in Jupyter notebooks.

Pypowsybl-jupyter integrates the powsybl-diagram-viewer library to display diagrams and is built on anywidget, an abstraction around the Jupyter Widget framework.

The widgets should work with versions of Jupyter Lab >= 4, Notebook >= 7.

Installation

You can install the widgets binaries using pip:

pip install pypowsybl_jupyter

Examples

In the examples directory there are some notebooks demonstrating the widgets.

Development installation

Create a virtual environment and install pypowsybl_jupyter in editable mode with the optional development dependencies:

python -m venv .venv
source .venv/bin/activate
pip install -e ".[dev]"

For example, in editable mode you can watch the source directory for changes, to automatically rebuild the widget, and run JupyterLab in different terminals. Changes made in js/ will be reflected in an open notebook where the widget is used.

Please note that pip only supports editable installs (enabled with the option -e) from a pyproject.toml files since v21.3. Make sure you have an up-to-date version of pip

pip install --upgrade pip

# Watch the source directory in one terminal, automatically rebuilding when needed
npm run dev
# Run JupyterLab in another terminal.
# Please ensure that you also activate the virtual environment in the new terminal.
source .venv/bin/activate
jupyter lab

To enable the automatic reloading in Jupyter, please define this environment variable:

%env ANYWIDGET_HMR=1

or, before launching a Jupyter session:

ANYWIDGET_HMR=1 jupyter lab

The changes should take effect after saving a source file. In case a change is not recognized, restarting the notebook kernel may help.

Packaging for distribution

To package pypowsybl_jupyter in a .whl file, for distribution:

pip install build
python -m build --wheel

The 'created .whl file' will be available in the 'dist' directory. To install the .whl file:

pip install <'created .whl file'>

Notes

Jupyter Widgets require Javascript extension code that Jupyter Lab needs to locate upon kernel startup.

In a standalone Jupyter Lab setup, the widget package is usually deployed in the same environment, allowing Jupyter Lab to find and load the necessary JS code seamlessly (e.g., if you pip install pypowsybl_jupyter in your standalone environment, then start Jupyter Lab, the extension is activated and anywidget and Jupyter Widget are listed among the installed dependencies).

However, installing Jupyter Widgets in kernels that are spawned by more controlled environment like Mybinder, doesn't apparently let Jupyter Lab find the JS extension code, causing widgets to fail to display properly (e.g., in these cases, if you pip install pypowsybl_jupyter from inside a notebook a javascript error could appear in the widget's cell, with a message similar to "Failed to load view class 'AnyView' from module 'anywidget'", despite the fact that no errors are shown during the installation nor that anywidget and Jupyter widget are still listed among the installed dependencies).

To resolve this known limitation, widget packages need to be installed in such a way that they are available to the hosting environment. In the Mybinder case it seems to be sufficient to add an anywidget entry to the project's requirements.txt file, which is used by Mybinder's infrastructure to instantiate Jupyter Lab. Once anywidget is installed succesfully, any custom widget developed using anywidget, like pypowsybl_jupyter's widgets, can be installed and managed correctly from a notebook (i.e., running pip install pypowsybl_jupyter in a notebook should work correctly).