A4 Tests & code quality

The application repository hosts the combined checks of both projects. One command, run from the application repository with the engine checkout beside it, verifies both together:

make quality

It runs in Docker: the catalog, architecture and lint checks; the application's Python and browser tests, without an installed engine and with a disposable PostGIS database; the engine's tests in a separate interpreter; then the coverage sanity and report steps, after a check for new duplicated blocks. CI's Required quality job uses the same runner. A4.4 Running the quality checks gives the faster commands, the native setup, CI, the evidence files and troubleshooting.

Where each part is described

Suite or topic What it covers Page
Application Python Flask blueprints and queries, the independent bundle reader, import and publication into PostGIS, the engine import contract, scheduling, the documentation portal and exports A4.1 Backend tests
Browser JavaScript Map adapters and lifecycle, filters, popups, review pages and progress presentation A4.2 JavaScript tests
Engine Python Source-neutral matching, input integrations, profiles, validation and atomic bundle writing E6 Engine tests
Code quality The dashboard, module criticality and how to read the evidence A4.3 Code quality
Quality checks Commands, CI, how the evidence reaches the site, coverage, the module catalog and the clone baseline A4.4 Running the quality checks

The suites meet only through versioned data. The application's tests read result bundles with engine imports forbidden, and its engine import contract runs the engine CLI as a subprocess. Pipeline progress is tested the same way: the engine checks its stage catalog and events, the application checks plan merging and state transitions, and the browser checks a generic tree renderer, including a stage the frontend does not know. Algorithm tests do not repeat UI assertions.

Related pages

Data update in progress
Elapsed: -- ETA: -- Phase: idle