Param in Panel#

Panel and other projects in the HoloViz ecosystem all build on Param. Param provides a framework to add validation, documentation and interactivity to a project. It is similar to more modern projects such as Pydantic but focuses primarily on providing APIs that make it easy to express complex dependencies, reactivity and UI interactions.

To get a better understanding of Param itself check out the user guide in the Param documentation but here we will focus on the use of Param in Panel and how it powers all reactive and callback based APIs.

What is a Parameter?#

Parameters are objects that express the semantics of a value of an attribute on an object, e.g. the value of Widget, i.e. whether it is a number or a string. Generally Parameters are broader than Python types because they deal with semantics not the specific representation of a value. A Parameter object also acts as a reference to the underlying value, which is incredibly useful when trying to add interactivity to a UI and forms the basis of most of the reactive APIs in Panel.

To make this more concrete let’s look at a parameter definition, in this case for the Panel TextInput widget value parameter:

class TextInput(Widget):


    value = param.String(default='', allow_None=True, doc="""
        Initial or entered text value updated when <enter> key is pressed.""")

A Parameter must be defined inside a class and that class must be a subclass of Parameterized. We can also see that it is of type param.String, defines a default of '' and allows None value. It also has a docstring which gives us some information about it. All Panel components are configurable via parameters.


The difference between a parameter value and the parameter itself is a central concept to really understand how Panel works, so make sure you pay attention to the next part.

Let’s start by working with a TextInput widget:

import panel as pn

text_input = pn.widgets.TextInput(value='A string!')


We can access the current value of the widget:

A string!

But we can also access the Parameter instance for value:

<param.parameterized.String object at 0x10e264d60>

Not only does the parameter hold the various metadata and validation options for the parameter but in many scenarios it can be used as a stand-in, proxy or reference for the actual value which can then be resolved interactively whenever the value changes.


The concept of a reference is crucial to using Panel effectively, as much of the reactivity in Panel is built on top of the idea of reference binding. The other piece (i.e. the binding) is that parameters can be given a reference, which they will internally resolve and update whenever the reference updates. In param itself this has to be declared by declaring allow_ref=True on the parameter, in Panel effectively all parameters automatically enable this functionality and also allow support for nested_refs=True where possible.

Let us explore this with a simple example:

import param

class TextFormatter(param.Parameterized):

    text = param.String(allow_refs=True)

    def __str__(self):
        return f'**{self.text}**'

text_input = pn.widgets.TextInput(value='A string!')

TextFormatter(name='TextFormatter00116', text='A string!')