[m-messer]

Moves off baseevalutionfunctionlayer:main-3.8 (Python 3.8 runtime, deprecated; the base layer has stopped building the 3.8 image).

• app/Dockerfile: main-3.12, the same recipe as survey. yum install git becomes dnf install git findutils; also adds PYTHONPATH=/app, pytest, and CMD ["app.handler"]
• Renamed evaluation_test.py / preview_test.py to *_tests.py. The base layer healthcheck imports evaluation_tests, which is why prod healthcheck currently reports tests_passed: false
• sympy 1.10.1 changed to 1.12.1, because 1.10.1's parse_expr misbehaves under Python 3.12's new tokenizer. SymbolicEqual needed the same bump (feedback-messages #80)
• Workflows now match survey's: adds test-lint.yml (runs on PRs, Python 3.12), the staging test job runs on 3.12, and run-database-tests: true on production deploy
• Added pytest.ini

Testing (branch deployed to staging; not tested locally):

• PR CI on 3.12: 55 passed, 3 skipped, 1008 subtests passed
• Staging endpoint vs prod: eval and preview output identical across 6 cases, including 3! and -∞
• Database tests on staging: 500/500 passed
• Note: healthcheck now returns 503 on staging. With the files renamed, the base layer actually runs the full suite (1008 subtests) inside the Lambda, which takes longer than API Gateway's 30s limit. Prod answers fast only because it can't import the tests. Nothing in the platform calls healthcheck
• TEST_API_ENDPOINT repo secret set to the comparePhysicalQuantities-staging URL so pre_production_tests.yml works

🤖 Generated with Claude Code