feat(editor): chord inspector — author chord name + fingering (E1)

[byrongamatos]

What

Adds a Chord section to the arrangement-editor inspector, shown whenever the selection resolves to a chord (≥2 notes sharing a save-time group). A charter can author the shared chord template inline:

• Name and Display name (text inputs)
• Per-string fingering — one <select> per sounding string (—/open/1–4), width-aware via lanes() so 7/8-string charts work
• Arpeggio toggle

All controls are inline DOM — no window.prompt() (it's a silent no-op in the bundled Electron).

How

• Every edit goes through a new EditChordTemplateCmd undo command (exec/rollback, like EditAnchorFretWidthCmd). It finds-or-creates the matching arr.chord_templates entry keyed by the width-L fret pattern.
• The fret key is resolved from the full save-time group (time.toFixed(4)), exactly as reconstructChords() groups — so a partial selection (e.g. rectangle-selecting 2 of a 3-note chord) still authors the chord's real key and the metadata survives the save-time rebuild via relinkChordTemplate.
• displayName/arp now round-trip end to end (E0 deferred them):
  • relinkChordTemplate carries displayName (fallback to name) and arp.
  • routes.py wire writer emits templates[].displayName + arp; XML writer emits the chordTemplate displayName attribute.
  • routes.py read side (_song_to_dict + the two GP import serializers) surfaces displayName/arp off the core ChordTemplate, so reload repopulates the inspector.
  • Core lib/song.py already models display_name/arpeggio and arrangement_from_wire already reads displayName/arp, so the round-trip is lossless with no core change.

Notes / limitations

• Template-level arp is wire-only. RS-XML arpeggio lives on <handShape>, not <chordTemplate>, so template arp persists/reloads through the sloppak wire but is not emitted as a chordTemplate XML attribute — it's reconstructed from the handshape layer. Bridging template arp → <handShape> is future work.
• Same-fret chords share one template (first-wins), so renaming one occurrence affects all occurrences of that fret pattern. Acceptable for E1; a per-_chordId split is future (E2-adjacent) work.

Verify

• node --check screen.js — parses.
• node tests/chord_relink.test.js — 6/6 pass, including new displayName/arp carry + fallback + blank assertions.
• python -m pytest tests/ — 67 pass (the one failure, test_populates_empty_notation_only_arrangement, is pre-existing on main and unrelated to chord templates; test_xml_export.py/test_arrangement_xml_alignment.py 57/57 green).

Closes #4. Part of got-feedBack/feedBack#334.

🤖 Generated with Claude Code