fix(ui): derive completeness fields and scores from scoring payload

[jonathanjasare]

Fixes #24.

Summary

• Render all five field-checklist categories from completeness_score.category_fields_list instead of keeping template-owned field lists.
• Render category and overall score maxima from the scoring payload instead of hard-coded point totals.
• Normalise the total-score progress display against the payload's summed category maxima.
• Add a Jinja render regression using unique fields and maxima for every category.

Why

The scoring model already builds category_fields_list, category_details.*.max_points, and max_scores. The results template duplicated parts of that data, so schema or scoring changes could update backend totals while leaving stale field rows or point totals in the UI.

Verification

• python -m unittest -v tests.test_result_template: 2 passed.
• python -m pytest -q -k 'not TestFixtureEndToEnd': 180 applicable tests passed; 3 Torch-dependent fixture tests excluded.
• ruff check tests/test_result_template.py: passed.
• python -m compileall -q src tests/test_result_template.py: passed.
• git diff --check: passed.

Scope

Changed only:

• src/templates/result.html
• tests/test_result_template.py

No scoring algorithm, schema, dependency, configuration, documentation, or unrelated refactor is included.

Known limitation

Three TestFixtureEndToEnd tests were excluded from the applicable suite because they exercise Torch-dependent fixture processing rather than the changed template path.