Streamlit Cloud

Learn to deploy our application on Streamlit Cloud.

We'll cover the following

Deployment options

Now that we’ve gone through the steps of developing an application, it’s time to deploy it so users can interact with it.

There are several options to deploy applications, including the following:

In this lesson, we learn to deploy applications using Streamlit Cloud.

Streamlit Cloud

Connect to the service

When we first access Streamlit Cloud, we see the page shown in the image above.

You need a Github account where our repository (repo) will be situated to deploy on Streamlit Cloud. This is the advisable way to sign up for the service.

The first time we do this, we go through the process of connecting our GitHub account to the service.

The repo should contain the Python script file and a requirements.txt file containing packages needed to build and run our application successfully.

Streamlit Cloud’s Community Tier allows the deployment of just one private repo and unlimited public repos.

Required files

As a demonstration, we use the following script for the application to be deployed. This script sets up a basic photo editor application that lets users upload an image and perform basic manipulation on it:

import streamlit as st
import numpy as np
from PIL import Image
st.title("Super Basic Photo Editor")
def display_figures(img1, img2):
    fig1_left, fig2_right = st.columns(2)
    with fig1_left:
        st.markdown("### Original image")
        st.image(img1)
    with fig2_right:
        st.markdown("### Processed image")
        st.image(img2)
image_file = st.sidebar.file_uploader("Upload an image file", type=["jpg","png","tif"])
if image_file is not None:
    input_image = Image.open(image_file)
    effect = st.sidebar.selectbox("Select the type of image manipulation.",
         ["Rotate Image", "Flip Image", "Resize"])
        
    if effect == "Rotate Image":
        angle = st.sidebar.select_slider("Select the angle of rotation", [0, 90, 180, 270, 360])
        output_image = input_image.rotate(angle)
    if effect == "Flip Image":
        mirror_choice = st.sidebar.radio("Choose  how to flip the image",
                             ["Horizontally", "Vertically"])
        if mirror_choice == "Horizontally":
            output_image = input_image.transpose(Image.FLIP_LEFT_RIGHT)
        else:
            output_image = input_image.transpose(Image.FLIP_TOP_BOTTOM)
    if effect == "Resize":
        scale_factor = st.sidebar.slider("Select the percentage size", 5, 10, 6)
        output_image = input_image.resize((input_image.width//scale_factor, input_image.height//scale_factor))
    if st.button("Process image"):
	    display_figures(input_image, output_image)

We need to set up a requirements.txt file with the following entries:

  • gitdb==4.0.7
  • GitPython==3.1.17
  • gunicorn==20.1.0
  • Markdown==3.3.4
  • streamlit==1.0.0
  • pip==21.3

Replace the library versions with appropriate values as needed.

Deploy the application

Once the repo is in place, we go to our Streamlit Cloud account and click on the “New app” button.

We specify the repo and script file for oour project in the new application window that comes up.

If we have any access credentials needed for the application to run, such as API keys, we need to pass those in as a toml configuration file. This can be done from the deployment page’s “Advanced settings” tab.

With everything in place, we click “Deploy” and let it build the project.

Once completed successfully, we can now access the application from anywhere.