ImGui key names aren't canonicalized: LeftCtrl+Key1 shows as Ctrl+KEY1 and ModCtrl+S as MODCTRL+S, so neither equals Ctrl+1 / Ctrl+S #224

What's wrong

The comment above the modifier switch in Note.CanonicalizeKey (Keybinding/Models/MusicalTypes.cs, ~L66–120) says it accepts ImGui names. It only covers ImGui's left/right modifier keys (LeftCtrl, LeftSuper, …). Hexa.NET.ImGui, which ktsu's ImGui stack uses, spells two other groups of keys differently. Neither group is folded.

• Main-row digits are ImGuiKey.Key0 … Key9.
• Modifier flags used with ImGui's chord API (IsKeyChordPressed(ImGuiKey.ModCtrl | ImGuiKey.S)) are ModCtrl, ModShift, ModAlt and ModSuper.

These names were confirmed by reflecting over ImGuiKey in Hexa.NET.ImGui 2.2.9.

Repro, using a console app that references Keybinding.csproj at 89f14b9:

static void Show(string a, string b) { Chord x = Chord.Parse(a), y = Chord.Parse(b); Console.WriteLine($"{a} -> {x} equals {b}? {x.Equals(y)}"); }
Show("LeftCtrl+Key1", "Ctrl+1");
Show("ModCtrl+S", "Ctrl+S");
Show("ModShift+ModAlt+S", "Shift+Alt+S");
Show("ModSuper+S", "Meta+S");

LeftCtrl+Key1     -> Ctrl+KEY1          equals Ctrl+1?      False
ModCtrl+S         -> MODCTRL+S          equals Ctrl+S?      False
ModShift+ModAlt+S -> MODALT+MODSHIFT+S  equals Shift+Alt+S? False
ModSuper+S        -> MODSUPER+S         equals Meta+S?      False

GLFW has the same gap. #209 accepted GLFW's snake_case modifier names (LEFT_CONTROL, LEFT_SUPER), but GLFW's other multi-word keys are not folded:

• LEFT_CONTROL+PAGE_UP shows as Ctrl+PAGE_UP, and Equals(Ctrl+PageUp) is false.
• CAPS_LOCK and PRINT_SCREEN behave the same way.
• By reading the code, PAGE_DOWN, NUM_LOCK and SCROLL_LOCK would too.

Why it matters

An ImGui host that builds chords from ImGuiKey.ToString() never matches stored bindings such as Ctrl+1 or Ctrl+S, and it shows unreadable names in the UI. This is the #205/#218 symptom again, reached through a framework the code says it supports.

Relation to #218 / PR #219

#218 covers the browser and Silk.NET spellings (KeyS, Digit1), not these. PR #219's test NoteConstructor_OtherKeyNamesStayDistinct has [DataRow("Key1")], which asserts that Key1 stays KEY1. Merging that test as written would lock this bug in. The fix here should flip that row to expect 1, while Keypad1/Numpad1 stay distinct.

Suggested fix / acceptance criteria

• Fold KEY<digit> (exactly four characters, last one 0–9) to the digit.
  • LeftCtrl+Key1 equals Ctrl+1.
• Map the ImGui modifier flags:
  • MODCTRL → CTRL
  • MODSHIFT → SHIFT
  • MODALT → ALT
  • MODSUPER → META
• Map the GLFW multi-word names to the canonical names already used for other frameworks: PAGE_UP, PAGE_DOWN, CAPS_LOCK, NUM_LOCK, SCROLL_LOCK, PRINT_SCREEN.
• Add tests for each of the inputs above.