msal_streamlit_t2 1.1.5

Streamlit Authentication library based on MSAL.JS

OpenID Connect (OIDC) authentication component for Streamlit

About

This Streamlit component enables client-side authentication using Azure AD work and school accounts (AAD), Microsoft personal accounts (MSA) and social identity providers like Facebook, Google, LinkedIn, Microsoft accounts, etc. through Azure AD B2C service. The component is achieving this by applying the Microsoft MSAL JS Library inside of a React project. Since the component is based on MSAL, it can be configured to support any provider that supports the OpenID Connect Authorization Code Flow (PKCE). For more information on MSAL, consult the Github project and its offical documentation.

Usage

Below is a sample Python snippet displaying how to apply the component. Visually, the component gives rise to a single button in the Streamlit Dashboard with a text that depends on whether an active login session exists. The auth and cache parameters are entirely equivalent to the properties mentioned in the Github documentation. The login_request and logout_request parameters are covered here.

import streamlit as st
from msal_streamlit_t2 import msal_authentication


login_token = msal_authentication(
    auth={
        "clientId": "aaaaaaa-bbbb-cccc-dddd-eeeeeeeeeee",
        "authority": "https://login.microsoftonline.com/aaaaaaa-bbbb-cccc-dddd-eeeeeeeeeee",
        "redirectUri": "/",
        "postLogoutRedirectUri": "/"
    }, # Corresponds to the 'auth' configuration for an MSAL Instance
    cache={
        "cacheLocation": "sessionStorage", # or 'localStorage' for sharing sessions between tabs 
        "storeAuthStateInCookie": False
    }, # Corresponds to the 'cache' configuration for an MSAL Instance
    login_request={
        "scopes": ["aaaaaaa-bbbb-cccc-dddd-eeeeeeeeeee/.default"]
    }, # Optional
    logout_request={}, # Optional
    login_button_text="Login", # Optional, defaults to "Login"
    logout_button_text="Logout", # Optional, defaults to "Logout"
    class_name="css_button_class_selector", # Optional, defaults to None. Corresponds to HTML class.
    html_id="html_id_for_button", # Optional, defaults to None. Corresponds to HTML id.
    key="your_streamlit_key" # Optional if only a single instance is needed
    hide_button=hide_button, # Optional, sometimes you can hide the button in some pages
    hide_button_text="loading...", # Optional, display some text when button is hidden
)
st.write("Recevied login token:", login_token)

or wrap it in a function and reuse it in other pages:

import streamlit as st
from msal_streamlit_t2 import msal_authentication


def app_msal(hide_button=True):
    return msal_authentication(
        auth={
            ..........


##page1.py:

app_msal()

##page2.py

app_msal(False)

A minimal sample project using the library can be found here. Note that it is Dockerized.

The component currently expects for the user to go through a popup based login flow. Further flows may be supported at a later time. As discussed here, the protocolMode parameter in auth can be used to configure OIDC providers that differ from Azure AD.

Inspiration

Inspired by official Streamlit template, this tutorial (Github) and the official Streamlit NPM component-lib. Speciall thanks to msal_streamlit_authentication, which is a project hasn't been updated for a long time. This library is basically an updated version of that.