perf(asr): speed up local Flow dictation and land CLM/keyboard refactor

Reduce perceived latency from key release to final text:
- Adaptive chunking: 2.5s first chunk + 5s follow-ups so short
  utterances start on-device recognition while still recording.
- Session-level ASR warmup and audio-format cache reuse to remove
  per-utterance cold-start of SpeechAnalyzer.
- Mirror live pipelined partials to the keyboard transcript line via
  a new flow.transcriptionPartial App Group key + Darwin ping.

Also commits the accumulated custom language model, Flow session,
keyboard extension restructure, and Xiaomi MiMo provider work in
progress on this branch.

OSGKeyboardShared/Utilities/OSGLog.swift

@@ -0,0 +1,17 @@
+// OSGLog.swift
+// OSGKeyboard · Shared
+//
+// Unified os.Logger categories for cross-target diagnostics. Filter in
+// Console.app with subsystem `com.osgkeyboard.ios`.
+
+import os
+
+public enum OSGLog {
+    private static let subsystem = "com.osgkeyboard.ios"
+
+    public static let flow = Logger(subsystem: subsystem, category: "flow")
+    public static let clm = Logger(subsystem: subsystem, category: "clm")
+    public static let config = Logger(subsystem: subsystem, category: "config")
+    public static let asr = Logger(subsystem: subsystem, category: "asr")
+    public static let keyboardExt = Logger(subsystem: subsystem, category: "keyboardExt")
+}