Shiny Widgets Integration

0.8.0 · active · verified Fri Apr 17

shinywidgets facilitates the rendering and interactive use of `ipywidgets` within `Shiny for Python` applications. It bridges the gap between the rich interactive capabilities of `ipywidgets` and the reactive framework of Shiny, allowing users to embed existing widget ecosystems like `ipyleaflet`, `plotly.graph_objects.FigureWidget`, and `ipyvolume` directly into their Shiny apps. The current version is 0.8.0, with a fairly active release cadence, often addressing compatibility with new versions of `ipywidgets`, `plotly`, and `websockets`.

Common errors

```
AttributeError: object has no attribute '_repr_mimebundle_'
```
cause This error occurred in `shinywidgets` versions prior to 0.5.1 due to an internal handling issue with widget representation, particularly with certain `ipywidgets`.

fix Upgrade `shinywidgets` to version `0.5.1` or higher: `pip install --upgrade shinywidgets`.
```
Plotly graph not rendering or failing to update in Shiny app.
```
cause Common causes include missing `anywidget` dependency, an outdated `plotly` version (pre-6.0.0), or incompatibility with `websockets>=16.0` in `shinywidgets < 0.7.1`.

fix Ensure `pip install anywidget plotly>=6.0.0` and `pip install --upgrade shinywidgets` (to at least v0.7.1) are executed.
```
ipyleaflet erroring out when attempting to read the .model_id property of a closed widget object.
```
cause This issue was specific to `ipyleaflet` integration in `shinywidgets` versions prior to 0.5.0, related to how closed widget objects were handled.

fix Upgrade `shinywidgets` to version `0.5.0` or higher: `pip install --upgrade shinywidgets`.

Warnings

gotcha Widgets initialized inside a `reactive.effect()` are no longer automatically removed when the effect invalidates. This change requires manual management of widget lifecycle if automatic cleanup was previously relied upon.
Fix: If automatic cleanup is required, consider re-evaluating widget initialization placement or explicitly managing the widget's lifecycle with `widget.close()` within the effect.
breaking For `altair` and `plotly` integrations, `anywidget` is now a direct dependency and must be installed separately. Additionally, `plotly` versions prior to 6.0.0 might be incompatible.
Fix: Ensure you `pip install anywidget` and, if using Plotly, `pip install plotly>=6.0.0`.
gotcha `anywidget`-based widgets (e.g., Plotly, Altair) may fail to initialize or update correctly with `websockets` versions 16.0 or higher due to a fix introduced in `shinywidgets` v0.7.1.
Fix: Upgrade `shinywidgets` to version `0.7.1` or newer (`pip install --upgrade shinywidgets`).
gotcha In `shinywidgets` versions before 0.7.2, widget cleanup might not occur correctly if the widget was created inside a render output's context that differs from the current context, potentially leading to resource leaks or unexpected behavior.
Fix: Upgrade `shinywidgets` to version `0.7.2` or newer (`pip install --upgrade shinywidgets`) to ensure proper widget cleanup.

Install

pip install shinywidgets Install core library
pip install shinywidgets ipywidgets ipyleaflet Install with common dependencies (e.g., ipyleaflet)

Imports

output_widget
```
from shinywidgets import output_widget
```
render_widget
```
from shinywidgets import render_widget
```

Quickstart

This example demonstrates how to integrate an `ipywidgets.IntSlider` into a Shiny application. The `output_widget` function creates a placeholder in the UI, and the `@render_widget` decorator on the server side defines the `ipywidget` to be rendered. The widget's value can be accessed via `input.widget_id()` just like standard Shiny inputs.

from shiny import App, ui, render
from shinywidgets import output_widget, render_widget
import ipywidgets as widgets

app_ui = ui.page_fluid(
    ui.h2('Shiny + ipywidgets'),
    output_widget('my_slider'),
    ui.output_text('slider_value'),
)

def server(input, output, session):
    @render_widget
    def my_slider():
        return widgets.IntSlider(min=0, max=100, value=50, description='Slider:')

    @output
    @render.text
    def slider_value():
        return f'Slider value: {input.my_slider()}'

app = App(app_ui, server)

view raw JSON →