Hello Streamliters
I found Streamlit incredibly useful for data visualization and rapid application development. Thanks and kudos!
Because I’m going to build more serious application than a single page, where proper maintanability, extensibility and testability will be important factors I started looking for Streamlit software engineering best practices. How to build Streamlit application in a modular way using test driven development practices.
To share my experiments, I created project to demonstrate how to leverage Streamlit’s built-in power using best software engineering practices and tools:
-
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 execution and debugging available directly from VS Code
-
Minimal number of external dependencies
-
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
A few more ideas for the future:
- Test coverage
- Implement BigQuery integration tests
- More visualizations for integrated public dataset
- Integration with external OAuth provider
- Redirect to the original page after login
- Describe load balancer strategies (sticky sessions)
- Automated dependency updates with Dependabot/Renovate
- More sophisticated type checking
I would love to hear your opinions, questions, and overall impressions. It would be fantastic if my experiments could be useful to others!