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.
This commit is contained in:
Rocky
2026-07-06 00:00:19 +08:00
parent cfbfb542cc
commit 537a68552a
76 changed files with 3456 additions and 121086 deletions
@@ -34,7 +34,6 @@ public struct AppGroupPersistor {
// Both engines always polish; ignore legacy off/transcribe modeId.
state.mode = .polish
state.engineMode = store.engineMode
state.localASRBackend = store.localASRBackend
// v0.2.1 follow-up: only the target locale is persisted
// `translationEnabled` is derived from it. Hydrate once at
// startup; `refreshRuntimeFlags` keeps the chip in sync while
@@ -42,16 +41,10 @@ public struct AppGroupPersistor {
state.translationTargetLocaleId = store.translationTargetLocaleId
state.handednessPreference = store.handednessPreference
state.cursorDragNavigationEnabled = store.cursorDragNavigationEnabled
state.localModeCloudPolishEnabled = store.localModeCloudPolishEnabled
state.micDisabled = store.isCloudAPIKeyMissingForVoiceInput
state.micDisabledHint = store.isCloudAPIKeyMissingForVoiceInput
? ExtL10n.string("keyboard.mic.disabled.missingApiKey")
: ""
// v0.2.0: iOS `SpeechAnalyzer` is always ready; mirror that
// into the State flags so downstream consumers see the same
// shape they did when the previous Qwen3 stack reported "ready".
state.localModelsReady = true
state.localModelsLoaded = false
#if DEBUG
// Print a masked view of the live App Group config so we can see
@@ -74,7 +67,6 @@ public struct AppGroupPersistor {
model = \(store.model)
modeId = \(store.modeId)
localeId = \(store.localeId)
localASRBackend = \(store.localASRBackend.rawValue)
""")
#endif
return .loaded
@@ -93,8 +85,6 @@ public struct AppGroupPersistor {
guard AppGroup.isAvailable else { return }
let store = AppGroupStore()
state.engineMode = store.engineMode
state.localASRBackend = store.localASRBackend
state.localModeCloudPolishEnabled = store.localModeCloudPolishEnabled
let shouldProtectTranslation = protectTranslationUntil.map { Date() < $0 } ?? false
if !shouldProtectTranslation {
state.translationTargetLocaleId = store.translationTargetLocaleId
@@ -105,11 +95,6 @@ public struct AppGroupPersistor {
state.micDisabledHint = store.isCloudAPIKeyMissingForVoiceInput
? ExtL10n.string("keyboard.mic.disabled.missingApiKey")
: ""
// v0.2.0: iOS `SpeechAnalyzer` is always ready. Keep these
// toggles here so the keyboard UI doesn't flicker if the host
// app briefly clears them while refactoring.
state.localModelsReady = true
state.localModelsLoaded = false
}
/// Persist `mode` to the App Group store.
@@ -130,12 +115,6 @@ public struct AppGroupPersistor {
AppGroupStore().setEngineMode(engineMode)
}
/// Persist `localASRBackend` to the App Group store.
public func persist(localASRBackend: LocalASRBackend) {
guard AppGroup.isAvailable else { return }
AppGroupStore().setLocalASRBackend(localASRBackend)
}
/// v0.2.1: persist translation target locale id (e.g. `"en"`,
/// `"ja"`, or `TranslationLanguageCatalog.offLocaleId`). The
/// chip / picker call this through `KeyboardState.setTranslationTargetLocaleId`.
@@ -149,4 +128,4 @@ public struct AppGroupPersistor {
guard AppGroup.isAvailable else { return }
AppGroupStore().setTranslationTargetLocaleId(translationTargetLocaleId)
}
}
}