18. Production Checklist

Learning Goal

Prepare a BrickflowUI app for serious use by a team.

UX Checklist

• Every page has a clear title.
• Every control has a label.
• Tables have useful column labels.
• Charts have titles and units.
• Empty states explain what happened.
• Loading states are visible where data may be delayed.
• Important actions are visually distinct.
• Risk states use consistent colors.
• The app works on laptop-sized screens.

State Checklist

• Inputs are controlled by state.
• Filters drive all relevant tables and charts.
• Row and chart clicks produce useful drilldowns.
• Multi-value controls use lists.
• Date controls use a consistent shape.
• State is not mutated in place.

Data Checklist

• Mock data shape matches production data shape.
• SQL result columns match component keys.
• Expensive queries are not repeated unnecessarily.
• Errors are surfaced clearly.
• Sensitive data is not displayed accidentally.

Deployment Checklist

• requirements.txt pins an appropriate version.
• app.yaml is present.
• Frontend assets are included in the installed package.
• Environment variables are documented.
• Databricks SQL access is tested.
• Auth expectations are clear.

Repository Checklist

Commit:

• source files
• docs
• examples
• tests
• brickflowui/frontend/dist

Do not commit:

• frontend/node_modules
• root dist
• site
• __pycache__
• .pytest_cache
• pytest-cache-files-*
• _tmp_cli_*
• throwaway sample_app_*

Release Checklist

Run:

python -m pytest -q
cd frontend
npm run build
cd ..
python -m mkdocs build
python -m build
python -m twine check dist/brickflowui-0.1.6*

Common Mistakes

• Treating examples as an afterthought.
• Shipping a new component without docs.
• Shipping docs without running MkDocs.
• Shipping frontend changes without checking the wheel.

Exercise

Take one example app and review it against this checklist. Write down the first five improvements you would make.

Checkpoint

You should be able to judge whether a BrickflowUI app is ready for team use.