04.03 — Tests scénarios (pytest + allure)#
15 fichiers test_*.py du dossier tests/scenarios/ couvrent le DSL, l’orchestration et l’acteur Playwright. Tests dynamiques.
Listing#
tests/scenarios/
├── conftest.py # FakeDriver, RecordingPOM, make_*
├── test_railway_and_action_chain.py # create_act, drive_page, Result, ChainRunner
├── test_match_page.py # match_page / when (premières / dernières branches, exceptions)
├── test_validate.py # validate, assert_that, otherwise, then, chain_validations
├── test_invariants_properties.py # hypothesis (cf. 06-hypothesis-properties)
├── test_drivers_pool.py # WebDriversPool (acquire, warmup, shutdown, semaphore leaks)
├── test_screenshotter.py # Screenshotter (single shot, burst, healthcheck, threadsafety)
├── test_driver_builder.py # DriverBuilder (profile tmp dir, dispose)
├── test_watcher.py # Watcher (start, stop, callback errors swallowed, dedup cache)
├── test_playwright_driver_actor.py # acteur PlaywrightDriver (sync_playwright mocké, sans navigateur)
├── test_playwright_adapter.py # smoke navigateur réel (skip si pas de Chromium installé)
├── test_test_suite.py # TestSuite (parallel, saturation, only/exclude, transient_errors)
├── test_test_cycle_and_bootstrap.py # TestCycle, mode fail-fast vs wait-for-all, bootstrap
├── test_loggers_and_reports.py # PrintLogger, FileLogger, pretty_print, JSON
├── test_docx_tests_proofs.py # generate_docx_proof (parse logs, insert images)
└── test_env.py # smoke test : versions, Python 3.14+
Exemple de test interne#
@allure.epic("Railway / action chain")
@allure.feature("Action chain")
@allure.tag("act", "handlers", "happy-path")
@allure.severity(allure.severity_level.CRITICAL)
@allure.label("layer", "unit")
@allure.title("A successful act reports success; the success handler fires once")
def test_success_act_fires_success_handler() -> None:
pom = RecordingPOM()
calls = {"success": 0, "failure": 0}
chain = (
create_act(pom, lambda p: p.step("click"))
.failure(lambda _: calls.__setitem__("failure", calls["failure"] + 1))
.success(lambda: calls.__setitem__("success", calls["success"] + 1))
.execute()
)
assert chain.is_ok()
assert pom.calls == ["click"]
assert calls == {"success": 1, "failure": 0}
| Assertion | Cible |
|---|
chain.is_ok() | API publique |
pom.calls == ["click"] | comportement observable du POM (recording) |
calls == {"success": 1, "failure": 0} | bon handler appelé |
Annotations Allure#
| Catégorie | Valeurs typiques |
|---|
epic | “Railway / action chain”, “Invariants”, “Orchestration”, “Watcher”, “Drivers pool”, “Reports”, “Match page” |
feature | “Action chain”, “Validate”, “TestSuite”, “TestCycle”, “Bootstrap”, “Pretty print”, “JSON”, “DOCX proofs” |
tag | “act”, “handlers”, “happy-path”, “error-handling”, “retry”, “transient”, “smoke”, “skip”, “saturation”, “parallel” |
severity | CRITICAL > NORMAL > MINOR > TRIVIAL |
label("layer", ...) | “unit”, “integration” |
test_test_suite.py#
- Tests passent en mode séquentiel (
max_workers=1). - Tests passent en mode parallèle (
max_workers=N). - Filtrage
--only / --exclude. - Mutex
--only + --exclude. - Saturation : tests clonés avec
[COPY N]. - Retries sur transient errors.
- Setup failure → SKIPPED si tous les attempts ratent dès le setup.
- Test
skipped=True s’arrête directement. - Garde-fous (noms uniques, IDs uniques).
take_screenshot appelé sur fail si autoscreen_on_fail.- Comptage exact des tentatives sur
max_retries (un test couvre le bord de la boucle).
test_match_page.py#
| Cas | Assertion |
|---|
| Première branche matche → exécutée | chain.is_ok(), branches suivantes non évaluées |
| Deuxième branche matche → exécutée | id. |
| Aucune branche ne matche | NoMatchingBranchError dans le Fail |
Branche matche, mais son then fail | Fail propagé |
condition lève une Exception ordinaire | Traité comme False, branches suivantes essayées |
condition lève une exception du tuple raised_exceptions | Re-raise |
Branche then=[] matche | Ok(None) |
match_page au milieu d’un test_chain | Court-circuite si fail |
test_drivers_pool.py#
Note : utilisation d’un FakeDriverFactory qui peut être configuré pour lever / bloquer / dormir, et qui compte ses appels.
