Integrating streamlit and jupyter notebook's IPython.display

[Shawn_D_Souza]

I was trying to implement a simplified click-based Machine Learning builder utilizing Pycaret. One of the biggest issues that I faced was utilizing IPython.display modules, more specifically: display, HTML, clear_output, update_display.

What would be the best workaround to show the contents from the display module on streamlit? A bonus question would be how to auto-refresh the display to update the values every time a new model is updated.

[randyzwitch]

Hi @Shawn_D_Souza, welcome to the Streamlit community!

This idea sounds similar to what I did here for another user:

Is there a way to incorporate pivottablejs in streamlit

It’s a little awkward because the pivot table dynamically resizes, but this works on my computer import streamlit as st import streamlit.components.v1 as components from pivottablejs import pivot_ui import pandas as pd iris = pd.read_csv( "https://raw.githubusercontent.com/mwaskom/seaborn-data/master/iris.csv" ) t = pivot_ui(iris) with open(t.src) as t: components.html(t.read(), width=900, height=1000, scrolling=True) It wouldn’t be much to either add this to the pivottablejs Python…

The basic idea is to find where in your Python object you can get the HTML from, then display it. Additionally, we are trying to evaluate how we might automatically do something like this, so that all Jupyter widgets are displayed

Best,
Randy