api: round-trip injects implied default elements: `<type>` `<line>` `<voice>` `<staff>`

[webern]

Surfaced by the reworked round-trip classifier (PR #225; part of #212).

What

The api writer materializes elements the source left implicit, so the output has more of them than the input:

• <type> on notes that had none — extra <type> in lysuite/ly41d_StaffGroups_Nested.xml
• <line> under <clef> where the source omitted it — lysuite/ly12a_Clefs.xml
• <voice> where the source had none — mjbsuite/ChordDirectionPlacement.xml (expected: no <voice>; actual: <voice>1</voice>)
• <staff> where the source had none — lysuite/ly02e_Rests_NoType.xml

Related (same family, but never the sole blocker of a file): add:divisions (392) and add:attributes (359) — the writer also materializes implied containers.

Impact (reorder-free candidate files; 550 total)

• add:type 233, add:voice 29, add:line 27, add:staff 1

Decision

Per #214: decide per element whether the api should suppress these when the source omitted them, or whether injecting an implied default is acceptable and should be normalized away in comparison. This is the opposite direction from #219 (which tracks elements being dropped).

Repro

make dump-api-roundtrip && make classify-api-roundtrip
# build/api/classified.json: signatures add:type / add:line / add:voice / add:staff

References

• Part of api: expand corpus coverage of the api round-trip pass-list #208; feeds api: implement ranked fixes and grow the round-trip corpus baseline #213
• Relates to api: define divergence policy and guardrails for round-trip corpus coverage #214 (divergence policy) and investigate api dropping elements marked support=full on round-trip #219 (element drops)
• Surfaced by PR audit: classify api round-trip failures by measured divergence and rank a worklist #225