Westfall-io / windstorm / 39885264012

83%

push

github

:sparkles: Open element selection to Verification cases (#95)

* :sparkles: open element selection to Verification cases (issue #95)

Selecting a model case was hardcoded to Analysis* element types. Rikki's
sensor-verification pipeline drives windstorm with VerificationCaseDefinition
elements, which the adapter's metadata-resolution branch mis-tagged as
ToolVariable (not ToolExecution), so verification cases were never
windstorm-visible.

Broaden the adapter's "windstorm-visible case" set (is_analysis -> is_case) to
include VerificationCaseDefinition/Usage alongside Analysis* and actions, and
add the verification types to the qualifiedName pass. Selection was already
parameterized by --element_type, so verification cases are now selectable and
walked end-to-end.

Scope is intentionally Analysis + Verification only (no general broadening).

Verified with a real-jar export of a minimal `verification def` fixture
(validated by windtrader-java, 324KB export) committed under
tests/mocks/verification/. 62/62 tests pass, fresh-venv-proven.

* :robot: Format code with black

* :bug: address review-1: real verification render regression test + adapter hardening

Independent review of PR #99 flagged a blocking gap: the new tests passed
unmodified on main, so the behavior change shipped untested (only the
is_case->toolName broadening in export_adapter actually implements #95, and it
had zero coverage).

- Add test_verification_case_renders_offline: a full offline render through the
  committed real-jar verification fixture selecting via
  element_type="VerificationCaseDefinition", asserting deltaT/deltaE -> 5.0/4.0.
  Proven to FAIL on main's adapter ("Copying all files... no changes") and PASS
  with the fix — a genuine regression test.
- Tighten test_verification_case_metadata_tagged_tool_execution to match the
  case by @type AND declaredName (was first-by-name).
- Adapter hardening (review NB-1/3/4): shared _CASE_TYPES constant (no drift,
  single-site type ... (continued)

8 of 9 new or added lines in 1 file covered. (88.89%)

820 of 993 relevant lines covered (82.58%)

0.83 hits per line