Streamlit Non-Trivial Application Skeleton

This project demonstrates how to leverage the built-in power of Streamlit using best software engineering practices and tools. It provides a non-trivial Streamlit application skeleton, showcasing how to effectively utilize Streamlit's capabilities and address its limitations.

Key Features and Benefits

Architecture & Build

• 🏗️ Modular, layered architecture for better maintainability
• 🧩 Separation of concerns for improved testability
• ⚙️ Modern build tools to complete full CI builds in under 30 seconds
• 🧹 Automated formatter and linter
• 📊 Test coverage
• 🐞 Test execution and debugging available directly from VS Code
• 📦 Minimal number of external dependencies
• 🐳 Optimized Docker image for deploying application

Streamlit Features

• 🌐 Web routing using built-in multi-page navigation capabilities
• 🔄 Sample reusable, stateful components
• ✅ All pages and components tested with AppTest
• 📊 Vega-Altair example visualization
• 🗂️ BigQuery integration using the New York Taxi public dataset
• 🔒 Authentication skeleton, easily replaceable with OAuth
• 🔗 Application state sharing via URL
• 💾 Dataframe export buttons to CSV and XLS

TODO

• 🧪 Implement BigQuery integration tests
• 📈 Add more visualizations for integrated public dataset
• 🔐 Integration with external OAuth provider, see roadmap
• 📝 Add request logging
• 🔄 Redirect to the original page after login
• ⚖️ Describe load balancer strategies, for example: sticky session
• 🔄 Automatically update dependencies, see dependabot#10039
• 📝 Add more sophisticated type checking, see mypy

Modules

example.services

• Service layer for application logic
• Tested with Pytest and mocked infrastructure layer
• Shouldn't import Streamlit API besides @st.cache

example.infrastructure

• Infrastructure code, no application logic, no user interface
• Tested in realistic environment (Cloud, Test Containers, etc.)
• Shouldn't import Streamlit API
• Acts as anti-corruption layer, for example - expose Pandas DataFrame instead of underlying database API

example.ui.pages

• Application pages, no application logic
• Tested with Streamlit testing framework and mocked service layer
• Delegate shared UI components to example.ui.components
• Delegate application logic to example.service

example.ui.components

• Shared UI components
• Tested with Streamlit testing framework with small helper function wrappers
• Encapsulates Streamlit state management

example.utils

• Utility functions
• Tested within Pytest without mock
• Shouldn't import Streamlit API

Local development

Note

You need Google Cloud Platform account to access sample dataset!

Use VS Code with the following extensions:

• Python
• Ruff

Update environment:

uv sync

Execute the following commands to run the application locally:

streamlit run app.py

Check formatting:

ruff check

Execute tests with coverage:

pytest

Show outdated dependencies:

uv pip list --outdated