Summary

The web content provides a detailed guide on packaging a Streamlit application into a standalone executable for Windows or Linux using PyInstaller, allowing users to run the app without installing any dependencies.

Abstract

The article outlines a process for converting a Streamlit application into a standalone executable that can be run on Windows or Linux with a double-click. The author, using Ubuntu 18.04 and PyCharm, demonstrates creating a Python project with a virtual environment, installing necessary packages like Streamlit and Pandas, and writing a simple Streamlit application that displays a Pandas DataFrame. The guide includes steps to install PyInstaller, create a wrapper script, modify the Streamlit CLI script, and configure PyInstaller hooks and spec files to ensure all Streamlit dependencies are included. The final executable is built using PyInstaller with a custom hook directory and spec file, and the Streamlit configuration file is placed alongside the executable in the distribution folder. The process aims to simplify application deployment for non-technical users or those without the necessary environment set up.

Opinions

The author believes that packaging Streamlit applications as standalone executables is convenient for non-technical users or those without the proper environment.
The use of PyInstaller is presented as an effective solution for bundling all dependencies into a single executable file.
The author emphasizes the importance of modifying the Streamlit CLI script and PyInstaller's spec file to correctly handle Streamlit's specific requirements.
The guide suggests that creating a Streamlit application that can be executed without installation is beneficial for ease of use and accessibility.
The author's approach to including Streamlit files and configuration in the build process indicates a preference for thoroughness and attention to detail in the deployment process.

Make Streamlit as Standalone App Execution

Hello, I will present how can we use streamlit as One App that could be executed on Windows or Linux without installing anything (Just with a double click). So it will be very convinient for one who don’t have the envirenment, or for no Technical peapole …

Let’s first create a sample python app.

I Will use Ubuntu 18.04 as OS, and Pycharm to edit code.

Let’s create our project called : sample_exe and use a virtualenv as the virtual environement on which we install all our necessay pakcages for the application. I use python3.9 in my example

After creating: Here is my code structure

Let’s create a simple streamlit program that show a pandas dataframe.

For That we have to install pandas and streamlit.

pip install streamlit

pip install pandas

Modify your main.py by this code:

import streamlit as st
import pandas as pd

if __name__ == '__main__':
    st.header("Hello world")
    d = {'col1': [1, 2], 'col2': [3, 4]}
    df = pd.DataFrame(data=d)
    print(df)
    st.dataframe(df)

You can display the result by executing :

streamlit run main.py

Now let’s install pyinstaller, a way allow us to make all depandency packaged in the same folder

pip install pyinstaller==4.10.0

Create a wrap script we can called it run_main.py

import streamlit.web.cli

if __name__ == '__main__':
    streamlit.web.cli._main_run_clExplicit('main.py', 'streamlit run')

in the venv folder search the cli.py script

for me it is in :

/sample_exe/venv/lib/python3.9/site-packages/streamlit/web/cli.py

add the following function to that script:

def _main_run_clExplicit(file, command_line, args=[], flag_options={}):
    import streamlit
    streamlit._is_running_with_streamlit = True
    bootstrap.run(file, command_line, args, flag_options)

Now, in the project main directory (sample_exe), let’s create two folders and two files inside each of them

Create ./hooks/hook-streamlit.py
Create ./.streamlit/config.toml

hook-streamlit.py

from PyInstaller.utils.hooks import copy_metadata
datas = copy_metadata('streamlit')

config.toml

[global]
developmentMode = false

[server]
port = 8501

Here is our actual tree

==> Execute this command to create a run_main file as an executer. This will take a couple of minutes (Building Time). But this is not working yet, we have to work more to make streamlit files detected natiely.

pyinstaller --onefile --additional-hooks-dir=./hooks run_main.py --clean

Let’s see now our new Tree project (Refresh project if needed)

We see that we have a new created file called run_main.spec

This file is responsible for the execution file. In order to able charging files of streamlit, we have to do some modifications, like that.

We will modify some things

1- pathex=[] , datas = [], hiddenimports=[]

2- Make a recursion limit

The run_main.spec file will seem like that

# -*- mode: python ; coding: utf-8 -*-

import sys
sys.setrecursionlimit(sys.getrecursionlimit() * 5)

block_cipher = None


a = Analysis(['run_main.py'],
             pathex=['.'],
             binaries=[],
             datas=[
                 (
                     "venv/lib/python3.9/site-packages/altair/vegalite/v4/schema/",
                     "./altair/vegalite/v4/schema"
                 ),
                 (
                     "venv/lib/python3.9/site-packages/streamlit/static",
                     "./streamlit/static"
                 )
            ],
             hiddenimports=['streamlit','pandas'],
             hookspath=['./hooks'],
             hooksconfig={},
             runtime_hooks=[],
             excludes=[],
             win_no_prefer_redirects=False,
             win_private_assemblies=False,
             cipher=block_cipher,
             noarchive=False)
pyz = PYZ(a.pure, a.zipped_data,
             cipher=block_cipher)

exe = EXE(pyz,
          a.scripts,
          a.binaries,
          a.zipfiles,
          a.datas,
          [],
          name='run_main',
          debug=False,
          bootloader_ignore_signals=False,
          strip=False,
          upx=True,
          upx_exclude=[],
          runtime_tmpdir=None,
          console=True,
          disable_windowed_traceback=False,
          target_arch=None,
          codesign_identity=None,
          entitlements_file=None )

Finally, execute, to rebuild project with the new spec.

pyinstaller --onefile --additional-hooks-dir=./hooks run_main.spec --clean

Now copy the .streamlit folder and main.py into dist folder

Go to the dist folder and execute :

./run_main