validation: scan early lines for Feature: in .feature test files

[hdamker]

What type of PR is this?

bug

What this PR does / why we need it:

P-007 / P-024 (validation/engines/python_checks/test_checks.py, _extract_feature_version) read only the first physical line of each .feature file. When line 1 is a Gherkin comment or a feature-level tag, the version regex finds no match and P-024 fires as a hard error.

Both styles are valid Gherkin and consistent with documentation/API-Testing-Guidelines.md in camaraproject/Commonalities. A sweep across the 49 upstream API repositories with a code/Test_definitions/ directory shows 25 files in 9 repositories with this layout (DeviceIdentifier, DeviceStatus, DeviceReachabilityStatus, DeviceRoamingStatus, ConnectedNetworkType, ApplicationEndpointDiscovery, ApplicationEndpointRegistration, ApplicationProfiles, OptimalEdgeDiscovery).

The release-time mechanical transformer is unaffected — _apply_regex runs re.subn on full file content, so it correctly transforms these files regardless of Feature-line position. The validation rule was the only thing flagging them.

This PR:

• Rewrites _extract_feature_version() to scan the first 50 lines for a line beginning with Feature: (skipping preceding comment, tag, and blank lines) and to return the actual Feature-line number alongside the version.
• Reports P-007 / P-024 findings at the actual Feature-line number rather than a hardcoded line=1.
• Splits P-024 into two sub-cases with distinct messages: (1) no Feature: line in the scan window, (2) Feature line present but carrying no wip / vwip / v{version} token.
• Drops implementation-coupling "snapshot transformation T1b" phrasing from user-facing messages.
• Revises the P-024 rule comment block in validation/rules/python-rules.yaml.

No release-automation changes. No regression-fixture recapture (the regression/r4.2-broken-spec-test-files regression-expected.yaml has no P-007/P-024 entries).

Which issue(s) this PR fixes:

Fixes #248

Special notes for reviewers:

Verified locally:

• 36/36 unit tests pass (python3 -m pytest validation/tests/test_python_checks_test.py -v), including 8 new cases covering comment/tag/blank-line preambles, the actual line number reported in findings, the two distinct P-024 sub-cases, and the 50-line scan cap.
• Direct check on the three real DeviceIdentifier feature files now resolves ('vwip', 2) for each — the original false positive is gone.
• Sweep across all 9 affected upstream API repositories returns the expected version + line number for every one of the 25 files.

Tag-move scope: rule-metadata tier (validation/engines/python_checks/**, validation/rules/python-rules.yaml, validation/tests/**). Validation Regression canary green is sufficient — no full ReleaseTest E2E required.

Changelog input

 release-note
P-007 / P-024 now correctly handle `.feature` files where the `Feature:` line is preceded by Gherkin comments, feature-level tags, or blank lines. Findings report the actual Feature-line number instead of always line 1.

Additional documentation

This section can be blank.

docs

[Kevsy]

LGTM