vizro-mcp

# How to extend and customize Vizro dashboards At its simplest, Vizro enables low-code configuration, but you can also customize it to go beyond its default capabilities by incorporating any Dash component, Dash callback, or Plotly chart. - **[Vizro customizations](#vizro-customizations)**. You can customize Vizro to extend the default functionality of Vizro and create Python functions as customized Plotly charts, tables, dashboard components, actions, or reactive HTML components, and then plug them directly into the existing Vizro dashboard configuration (as explained below). - **[Dash customizations](#dash-customizations)**. You can add custom Dash callbacks directly to any Vizro dashboard, enabling you to code beneath the Vizro layer and control Dash directly. - **[CSS customizations](#css-customizations)**. You can add custom CSS to any Vizro dashboard, enabling users to deviate from the default styling and create a unique look and feel for their dashboard. - **[React.js customizations](#reactjs-customizations)**. You can add custom React.js components directly to any Vizro dashboard, enabling users to go beneath both the Vizro and Dash layers and control React.js directly Vizro offers the ability to combine ease of use of low-code configuration, with the flexibility of a range of high-code extensions to expand your dashboard's functionality. ## Vizro customizations ### [Custom charts](custom-charts.md) You can create custom chart functions in Vizro by wrapping Plotly chart code inside a Vizro chart function wrapper, and then use the functions directly inside the Vizro dashboard configuration. This enables the creation of `plotly.graph_objects` charts with multiple traces, or `plotly_express` charts with post update customizations ### [Custom tables](custom-tables.md) If the available arguments for the [`dash_ag_grid`][vizro.tables.dash_ag_grid] or [`dash_data_table`][vizro.tables.dash_data_table] models are insufficient, you can create a custom Dash AG Grid or Dash DataTable. ### [Custom components](custom-components.md) You can create a custom component based on any dash-compatible component (for example, [dash-core-components](https://dash.plotly.com/dash-core-components), [dash-bootstrap-components](https://dash-bootstrap-components.opensource.faculty.ai/), [dash-html-components](https://github.com/plotly/dash/tree/dev/components/dash-html-components)). All Vizro's components are based on `Dash` and ship with a set of defaults that can be modified. If you would like to overwrite one of those defaults, or if you would like to use extra `args` or `kwargs` of those components, then this is the correct way to include those. You can use any existing attribute of any underlying [Dash component](https://dash.plotly.com/#open-source-component-libraries) with this method. ### [Custom actions](custom-actions.md) If you want to perform functions that are not available in the [built-in action functions](actions.md), you can [create your own custom actions](custom-actions.md). ### [Custom figures](custom-figures.md) Custom figures are useful when you need a component that reacts to [filter](filters.md) and [parameter](parameters.md) controls. Vizro's [`Figure`][vizro.models.Figure] model accepts the `figure` argument, where you can enter _any_ custom figure function as described in the [how-to guide for figures](figure.md). ### Use custom functions in `yaml`/`json` configuration !!! note "Exposing configuration to untrusted users" Exposing Vizro configuration to untrusted users may pose a security risk. A user with access to YAML/JSON configuration can potentially execute arbitrary Python code when the dashboard is run. The security of your `yaml`/`json` configuration should be regarded as equivalent to the security of your `app.py` file. It is possible to refer to custom functions that are used as `CapturedCallable` by their import path in a `yaml`/`json` configuration of the dashboard. In the [above guides](#vizro-customizations), you will find examples on how the Vizro schema can be extended by using custom Python code. It is possible to refer to these custom functions in the `yaml`/`json` configuration by using the `_target_` key and the correct import path. !!! example "Custom charts with YAML config example" === "app.py" ```python from pathlib import Path import vizro.plotly.express as px import yaml from vizro import Vizro from vizro.managers import data_manager from vizro.models import Dashboard data_manager["iris"] = px.data.iris() dashboard = yaml.safe_load(Path("dashboard.yaml").read_text(encoding="utf-8")) # (1)! dashboard = Dashboard(**dashboard) if __name__ == "__main__": Vizro().build(dashboard).run() ``` 1. Parse the YAML or JSON configuration that lies in a separate file. === "dashboard.yaml" ```yaml pages: - components: - figure: _target_: custom_charts.custom_bar # (1)! data_frame: iris x: sepal_length y: sepal_width type: graph title: Custom Bar Chart ``` 1. Here we refer to the import path of the custom chart function. If you define the custom chart in `app.py`, then use `__main__` as the import path. Note that the import path will be interpreted by pydantics [`ImportString` type](https://docs.pydantic.dev/dev/usage/types/string_types/#importstring). === "custom_charts.py" ```python import pandas as pd import plotly.graph_objects as go from vizro.models.types import capture @capture("graph") def custom_bar(data_frame: pd.DataFrame, x: str, y: str) -> go.Figure: # (1)! """Custom bar chart.""" return go.Figure(data=[go.Bar(x=data_frame[x], y=data_frame[y])]) ``` 1. Definition of the custom chart function as usual. #### Validate dashboards without defining `CapturedCallable` functions It is possible to validate a dashboard configuration without importing or executing some `CapturedCallable` functions. You can use this method when you want to check if the dashboard configuration is valid, but you don't want to import or execute the custom functions until run-time, which may be in a sandboxed environment. This is useful when the custom functions are not available at validation time, or when they originate from untrusted sources (e.g. when a large language model is used to generate that code). !!! note When starting the server If you validate your dashboard configuration with a `CapturedCallable` function that is undefined, you will not be able to start the server. Hence when running the dashboard, you will need to recreate a dashboard object with all the `CapturedCallable` functions defined. !!! example "Validating dashboards without executing `CapturedCallable` functions" ```python import vizro.models as vm import vizro.plotly.express as px from vizro import Vizro from vizro.managers import data_manager data_manager["iris"] = px.data.iris() dashboard_config = { "title": "Test dashboard", "pages": [ { "title": "Page 1", "components": [ { "type": "ag_grid", "figure": {"_target_": "llm_generated_grid", "data_frame": "iris"}, }, ], } ], } dashboard = vm.Dashboard.model_validate( dashboard_config, context={ "allow_undefined_captured_callable": [ "llm_generated_grid", ] }, ) app = Vizro().build(dashboard) # (1)! try: app.run() # (2)! except ValueError as exc: print(exc) ``` 1. The dashboard configuration contains a `CapturedCallable` function that is undefined but allowed by `allow_undefined_captured_callable`. The app can still be built without raising any errors. 1. However, it is not possible to run the app without undefined `CapturedCallable`s. This raises an error. ## Dash customizations Since Vizro is built using Dash, it is possible to use [Dash callbacks](https://dash.plotly.com/basic-callbacks) directly in any Vizro dashboard. This enables you to code beneath the Vizro layer and control Dash directly, which is especially useful when working with callbacks Here is an example showing a Dash callback within Vizro, enabling an interaction between data points in a scatter plot and the content of a text card: !!! example "Dash callback example" === "app.py" ```{.python pycafe-link hl_lines="6-13 18 20"} from dash import callback, Input, Output import vizro.models as vm import vizro.plotly.express as px from vizro import Vizro @callback( Output("card_id", "children"), Input("source_chart", "clickData") ) def update_card(click_data): if click_data is None: return "Click on the graph to select a data point." return f"Clicked species: '{click_data['points'][0]['customdata'][0]}'" page = vm.Page( title="Example: Dash callback within Vizro", components=[ vm.Graph(id="source_chart", figure=px.scatter(px.data.iris(), x="sepal_width", y="sepal_length", color="species", custom_data="species")), vm.Card(id="card_id", text="Click on the graph to apply filter interaction."), ] ) dashboard = vm.Dashboard(pages=[page]) Vizro().build(dashboard).run() ``` ## CSS customizations Vizro is opinionated about visual formatting, and some elements, such as the layout of the navigation and controls, are fixed. You can customize some settings such as background colors, fonts, and other styles via CSS overrides. For more information, see our documentation on [customizing CSS](custom-css.md) ## React.js customizations It is possible to create custom React.js components and add them directly to any Vizro dashboard so enabling you to code beneath both the Vizro and Dash layers and control React.js directly For more information, see the documentation on [using React.js components with Dash](https://dash.plotly.com/plugins)