style: apply spotless to the enable_thinking and calibrate changes

[bernardladenthin]

Summary

• main fails mvn verify at spotless:check. Five files carry format violations, from both feat!: send enable_thinking only when it was actually set #199 (the isChatTemplateEnableThinking → getChatTemplateEnableThinking rename pushed lines past the wrap point) and feat: write the calibration report as JSON and YAML #202 (the new renderers).
• Formatting only. No logic, no assertion, no test changed — the diff is 12 insertions and 8 deletions of line breaks.
• Produced by mvn spotless:apply, not by hand. My own hand-wrap of the same line was rejected by spotless:check on the next run, which is the argument for letting the tool decide the layout.

Why it got through twice

spotless:check binds to a phase mvn test never reaches. #202 was validated with a full reactor test run and a complete PIT pass at mutationThreshold 100 across all three modules — and still shipped unformatted, because none of that runs the goal.

This is the same failure mode as the SpotBugs breakage in java-llama.cpp#409 this week: a gate bound past test, a local loop that stops at test, and a green-looking change that reddens main. More tests do not close it; running the goal does.

Test plan

• mvn -o -f srcmorph/pom.xml spotless:check exits 0 (was exit 1 naming five files)
• Diff reviewed line by line — every hunk is a line break, nothing else
• CI is green on this branch — no: main pins the unpublished net.ladenthin:llama:5.2.0, so every srcmorph PR still fails at dependency resolution. Unrelated to this change.

Related issues / PRs

Fixes formatting introduced by #199 and #202.

Checklist

• I have read CONTRIBUTING.md and CODE_OF_CONDUCT.md
• My commits follow Conventional Commits
• No security-sensitive changes

🤖 Generated with Claude Code

https://claude.ai/code/session_01AnNYn8W1xuVxVJtyL34GyH

---

Generated by Claude Code

[claude]

Code Review Summary

✅ Approved — This PR contains formatting-only changes with no logic, assertions, or test modifications.

Review Findings

Code Quality: ✅ No issues

• All changes are pure line-break adjustments to comply with Spotless configuration
• No logic changes, no new bugs introduced
• Formatting applied by mvn spotless:apply tool, ensuring consistency

Security: ✅ No issues

• No sensitive data exposure
• No security-relevant code paths modified

Architecture & Design: ✅ Compliant

• Changes maintain the existing code structure and design patterns
• Complies with project conventions in CODE_WRITING_GUIDE.md

Context Validation

The root-cause analysis in the PR description is correct: spotless:check binds to the verify phase, which mvn test never reaches. This explains how #199 and #202 shipped with formatting violations despite passing comprehensive test suites (including full PIT mutation testing at 100% threshold). Running the formatting goal, not additional tests, is what catches these violations — this design pattern needs to be addressed in CI (noted in the PR as a parallel issue to SpotBugs in java-llama.cpp#409).

Files Reviewed

All 5 files contain only line wrapping/joining adjustments:

• AiModelDefinition.java — field initialization joined to single line
• LlamaCppJniConfig.java — field initialization joined to single line
• CalibrationReport.java — append chains split across multiple lines for length compliance
• AiModelDefinitionSupportTest.java — assertion split for length compliance
• CalibrationReportTest.java — assertion split for length compliance

✅ Ready to merge once CI gates pass.