How to get a decent understanding of param.Parameterized classes?

Param
1

Hello All!

I am trying to make a simple dashboard, where I have some filters and a button, which when pressed should show the DataFrame with the filtered data.

Something in line with this, for instance:
image

I feel that such a dashboard is a good use case for param.Parameterized classes, where @param.depends('button') would trigger the Data to update and therefore I wished to get a decent understanding of how it all works.

However, I feel somehow stuck in getting started. Probably because I am not sure what would be the best practices to use pn.widgets.MultiChoice within param.Parameterized.

If anyone can advice, thank you in advance!

2

The following code seems to work, although I am not fully sure why.
I am still skeptical that I am doing it correctly and I will not run into issues later on.

import panel as pn
import param
import pandas as pd
import random

df = pd.DataFrame(
    data={
        'label': [random.choice(list('abc')) for _ in range(100)],
        'value': random.sample(range(1, 1000), 100)
    }
)

LABELS = list(df['label'].unique())

class FilterData(param.Parameterized):
    max_value = df['value'].max()
    values = param.Range(default=(0, max_value), bounds=(0, max_value), step=max_value // 10)
    labels = pn.widgets.MultiChoice(value=LABELS, name="Filter by label", options=LABELS)
    button = param.Action(lambda x: x.param.trigger('button'), label='Fetch data')

    @param.depends('button')
    def output_panel(self):
        df_panel = pn.widgets.Tabulator(df[(df['value'] >=  self.values[0]) & (df['value'] <=  self.values[1]) & (df['label'].isin(self.labels.value))].head(10))
        return df_panel

    def dashboard(self):
        dashboard = pn.Row(
            pn.Column(
                self.param['values'],
                self.labels,
                self.param['button'],
                width=300
            ),
            self.output_panel
        )
        return dashboard

filter = FilterData()
filter.dashboard().show()

image

Also, I am not sure why the MultiChoice component goes into the button…
I don’t think I can mark my own answer as the solution :slight_smile:.

3

You are almost there, though some small tips:

  • You should almost never use a pn.widgets in the class declaration, as all objects will share it.
  • Instead, you should use param.Parameter and pn.Param either inside the class or outside on the object. I have chosen the first in my example below.
  • Try to look at pn.viewable.Viewer, which is almost the same as a param.Parameterized class except it has a __panel__ method which removes the need for calling your dashboard method. (I haven’t done that in my example below).
class FilterData(param.Parameterized):
    max_value = df['value'].max()
    values = param.Range(default=(0, max_value), bounds=(0, max_value), step=max_value // 10)
    labels = param.ListSelector(default=LABELS, objects=LABELS)
    button = param.Action(lambda x: x.param.trigger('button'), label='Fetch data')

    @param.depends('button')
    def output_panel(self):
        df_panel = pn.widgets.Tabulator(df[(df['value'] >=  self.values[0]) & (df['value'] <=  self.values[1]) & (df['label'].isin(self.labels))].head(10))
        return df_panel

    def dashboard(self):
        sidebar = pn.Param(self.param, widgets={"labels": pn.widgets.MultiChoice}, width=300)
      
        dashboard = pn.Row(
            sidebar,
            self.output_panel
        )
        return dashboard
2 Likes
4

Thank you for the tips, @Hoxbro
For some reason I can’t mark your answer as solution - I don’t even see the solution checkbox.

2 Likes
5

I believe that marking a post as a solution has not been enabled for the param category? I am not an discourse admin and cannot do it.

Can you do it @droumis? See GitHub - discourse/discourse-solved: Allow accepted answers on topics. If you can do it please enable it across all categories. Thanks.

1 Like
6

done and marked

2 Likes