PyScript and pyscope: Running Python Interactively in the Browser

The Problem: Python in the Browser Without a Server

Web pages typically run JavaScript. If you want to execute Python code perhaps for a data analysis widget a calculator or a learning tool you usually send data to a backend PyScript changes that by embedding a Python runtime directly in HTML

How PyScript Works

The Pyodide Runtime

PyScript relies on Pyodide a Python to WebAssembly compiler that brings a curated subset of the Python standard library and many third party packages into the browser When the browser loads a py script tag Pyodide downloads compiles and starts up before your code runs This start up can take several seconds on slow connections which is the first trade off to consider

Two Way Interop With pyscope

The most distinctive feature in PyScripts interop layer is the pyscope object It lets Python code invoke JavaScript functions through a simple call syntax

import numpy as np data = np.array([1, 2, 3, 4, 5]) pyscope.call('displayResult', data.mean())

On the JavaScript side define a function with the given name that receives the value and updates the page as needed

function displayResult(value) { // This function is called by PyScript update the DOM as needed }

A Worked Example Browser Based Statistics

Suppose you want a page where a statistical result is computed entirely in Python without a network request The Python code below uses NumPy to compute the mean and standard deviation of a small array then passes both values to JavaScript via pyscope call

import numpy as np data = np.array([10, 20, 30, 40, 50]) pyscope.call('showStats', data.mean(), data.std())

Define the JavaScript receiver function to display the results

function showStats(mean, std) { // This function receives the two values from PyScript }

When the page loads Pyodide downloads the Python block executes and the showStats function is called You can open the file in a browser and verify the values appear in the console or a designated element

Trade Offs and Limitations

Start Up Latency

Because PyScript must load Pyodide a multi megabyte WebAssembly module before any Python runs initial page load can be slower than a pure JavaScript equivalent On fast connections this is often imperceptible for a single script but repeated loads or slow networks will notice the delay

Standard Library and Package Constraints

The Python library available in PyScript is exactly what Pyodide packages Pure C extensions modules that rely on system level headers and some rarely used standard library modules are unavailable If your code imports csv or json those work because Pyodide ships them but a package like pandas may require an extra download depending on the PyScript version

Security and File System Access

Browser security policies restrict direct file system access Reading a local file from a file picker into Python requires explicit permission and writing back to the client filesystem is generally not allowed without a user gesture and the pyfilesystem2 shim which PyScript may or may not include depending on the build

Verifying Your Setup

1. Create a new HTML file and add a single py script tag with print Hello Open it in a browser Hello should appear in the DOM or console confirming the runtime loads
2. Try importing numpy add import numpy as np arr = np.array([1, 2, 3]) pyscope call log arr sum and define a log JavaScript function that console logs the result Use the browsers developer tools to confirm the import succeeds and the call executes

These steps match the verification hints from the PyScript documentation and should take under a minute on a typical connection

Closing Is PyScript Your Tool?

PyScript is compelling when you need Python semantics in a web page quick data wrangling educational examples or interactive widgets that stay entirely in the browser Its less suited for heavy numerical loops file system heavy tools or situations where every millisecond of start up matters If your use case fits the interactive script profile give the verification steps a try if the initial load or package limitations feel like blockers a JavaScript first approach with a Python to JS compiler might be worth exploring