Shiny

1.6.0 · active · verified Thu Apr 16

Shiny is an open-source web development framework for Python, enabling data scientists and developers to build interactive web applications and dashboards with purely Python code. Maintained by Posit, it is under active development with frequent releases that introduce features like OpenTelemetry integration, toast notifications, AI-powered test generation, and application bookmarking. It allows for the creation of rich user interfaces that react dynamically to user input.

Common errors

```
Error: An error has occurred. Check your logs or contact the app author for clarification.
```
cause This is Shiny's default sanitized error message for deployed applications, hiding the underlying Python exception for security reasons.

fix Run the application locally (`shiny run app.py --reload`) to see the full stack trace, or disable error sanitization in the `App` constructor: `app = App(app_ui, server, sanitize_errors=False)`.
```
Output not appearing (e.g., a blank space where a plot or text should be) or 'Missing output' message.
```
cause The output function (`@render.text`, `@render.plot`, etc.) is trying to access an input ID that does not exist or is misspelled. Alternatively, there might be a runtime error within the render function.

fix Verify that the `id` argument in `ui.output_*` matches the name of the decorated function in `server`. Ensure that `input.some_id()` refers to an existing `ui.input_some_id` component. Check the console for any Python exceptions or tracebacks.
```
TypeError: 'ReactiveValue' object is not callable
```
cause Attempting to access a reactive value (like `input.some_id`) without calling it as a function (`input.some_id()`) within a reactive context.

fix Always append `()` when reading the value of an `input` or `reactive.value` within `@reactive.effect`, `@render.*`, or `@reactive.calc` functions, e.g., `input.slider_val()` instead of `input.slider_val`.

Warnings

breaking Breaking changes in `shiny.playwright.controllers` for testing. `.expect_inverse()` now requires a `bool` (use `False` for previous behavior). `.expect_layout()` was renamed to `.expect_fluid()` and also requires a `bool` (use `True` for previous behavior).
Fix: Update test code to use `.expect_inverse(False)` and `.expect_fluid(True)` where applicable, or adapt to the new `bool` argument for `expect_inverse`.
deprecated The `ui.Chat().set_user_message()` method has been deprecated.
Fix: Use the new `ui.Chat().update_user_input()` method instead, which provides similar functionality and more flexibility.
gotcha Shiny applications, by default, sanitize error messages when deployed to prevent sensitive information leakage. This can make debugging deployed applications challenging.
Fix: To view full error messages for debugging purposes, set `sanitize_errors=False` in the `App` constructor: `app = App(app_ui, server, sanitize_errors=False)`.
gotcha Starting with v1.6.0, Shiny includes OpenTelemetry integration. The `SHINY_OTEL_COLLECT` environment variable controls the default collection level (e.g., 'none', 'session'). If not explicitly set, 'session' level telemetry might be collected.
Fix: Set the `SHINY_OTEL_COLLECT` environment variable to `'none'` if you wish to disable all Shiny-specific telemetry collection for sensitive operations or to reduce overhead.
gotcha If an output in a Shiny app fails to render or shows a 'Missing output' error, it's often due to an attempt to read a non-existent input ID or an error within the `@render` function.
Fix: Double-check input IDs for typos. If an error occurs within a `@render` function, the full Python stack trace will be available in the console where `shiny run` was executed. Debug the `@render` function's logic.

Install

pip install shiny Install latest version

Imports

App
```
from shiny import App
```
ui
```
from shiny import ui
```
render
```
from shiny import render
```
reactive
```
from shiny import reactive
```
input
wrong:
```
from shiny import input
```
correct:
```
from shiny.express import input
```
`input` is typically used implicitly within `@reactive` or `@render` functions, or directly imported from `shiny.express` for Express mode applications. In Core mode, it's passed as an argument to the `server` function.
output
wrong:
```
from shiny import output
```
correct:
```
from shiny.express import output
```
Similar to `input`, `output` is implicitly available in Express mode or passed to the `server` function in Core mode.
session
wrong:
```
from shiny import session
```
correct:
```
from shiny.express import session
```
Similar to `input` and `output`, `session` is implicitly available in Express mode or passed to the `server` function in Core mode.

Quickstart

This minimal Shiny application demonstrates a basic slider input (`ui.input_slider`) and a reactive text output (`ui.output_text_verbatim`). The `server` function uses the `@render.text` decorator to reactively display the current value of the slider.

from shiny import App, ui, render

app_ui = ui.page_fluid(
    ui.input_slider("n", "N", 0, 100, 20),
    ui.output_text_verbatim("txt"),
)

def server(input, output, session):
    @render.text
    def txt():
        return f"The value of N is {input.n()}"

app = App(app_ui, server)

view raw JSON →