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
@@ -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)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -0,0 +1,127 @@
|
||||
// CursorDragController.swift
|
||||
// OSGKeyboard · Keyboard Extension
|
||||
//
|
||||
// Cursor-drag hint chrome and batched caret moves via textDocumentProxy.
|
||||
|
||||
import UIKit
|
||||
import OSGKeyboardShared
|
||||
|
||||
@MainActor
|
||||
final class CursorDragController {
|
||||
private let state: KeyboardState
|
||||
private let adjustTextPosition: (Int) -> Void
|
||||
private weak var parentView: UIView?
|
||||
|
||||
private var cursorDragHintLabel: UILabel?
|
||||
private var pendingHorizontalCursorSteps = 0
|
||||
private var pendingVerticalCursorSteps = 0
|
||||
private var cursorMoveFlushScheduled = false
|
||||
private let cursorLineHaptic = UIImpactFeedbackGenerator(style: .light)
|
||||
|
||||
private static let cursorVerticalChunkSize = 20
|
||||
|
||||
init(
|
||||
state: KeyboardState,
|
||||
adjustTextPosition: @escaping (Int) -> Void
|
||||
) {
|
||||
self.state = state
|
||||
self.adjustTextPosition = adjustTextPosition
|
||||
}
|
||||
|
||||
func install(on view: UIView) {
|
||||
parentView = view
|
||||
let hint = UILabel()
|
||||
hint.text = ExtL10n.string("keyboard.cursorDrag.centerHint")
|
||||
hint.font = .systemFont(ofSize: 22, weight: .medium)
|
||||
hint.textColor = UIColor.label.withAlphaComponent(0.10)
|
||||
hint.textAlignment = .center
|
||||
hint.numberOfLines = 1
|
||||
hint.adjustsFontSizeToFitWidth = true
|
||||
hint.minimumScaleFactor = 0.7
|
||||
hint.isUserInteractionEnabled = false
|
||||
hint.isHidden = true
|
||||
hint.alpha = 0
|
||||
view.addSubview(hint)
|
||||
cursorDragHintLabel = hint
|
||||
layoutChrome()
|
||||
}
|
||||
|
||||
func layoutChrome() {
|
||||
guard let view = parentView else { return }
|
||||
cursorDragHintLabel?.frame = view.bounds
|
||||
}
|
||||
|
||||
func setCursorDragActive(_ active: Bool) {
|
||||
state.cursorDragActive = active
|
||||
updateCursorDragWash(active: active)
|
||||
}
|
||||
|
||||
func moveCursorHorizontally(by steps: Int) {
|
||||
guard steps != 0 else { return }
|
||||
pendingHorizontalCursorSteps += steps
|
||||
scheduleCursorMoveFlush()
|
||||
}
|
||||
|
||||
func moveCursorVertically(by steps: Int) {
|
||||
guard steps != 0 else { return }
|
||||
pendingVerticalCursorSteps += steps
|
||||
scheduleCursorMoveFlush()
|
||||
}
|
||||
|
||||
private func updateCursorDragWash(active: Bool) {
|
||||
if active {
|
||||
cursorLineHaptic.prepare()
|
||||
}
|
||||
layoutChrome()
|
||||
guard let hint = cursorDragHintLabel else { return }
|
||||
if active {
|
||||
hint.isHidden = false
|
||||
UIView.animate(withDuration: 0.12) { hint.alpha = 1 }
|
||||
} else {
|
||||
UIView.animate(withDuration: 0.12, animations: { hint.alpha = 0 }) { [weak self] _ in
|
||||
guard let self, !self.state.cursorDragActive else { return }
|
||||
hint.isHidden = true
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
private func scheduleCursorMoveFlush() {
|
||||
guard !cursorMoveFlushScheduled else { return }
|
||||
cursorMoveFlushScheduled = true
|
||||
DispatchQueue.main.asyncAfter(deadline: .now() + 0.012) { [weak self] in
|
||||
guard let self else { return }
|
||||
self.cursorMoveFlushScheduled = false
|
||||
|
||||
let horizontal = self.pendingHorizontalCursorSteps
|
||||
let vertical = self.pendingVerticalCursorSteps
|
||||
self.pendingHorizontalCursorSteps = 0
|
||||
self.pendingVerticalCursorSteps = 0
|
||||
|
||||
if horizontal != 0 {
|
||||
OSGLog.keyboardExt.info("adjustTextPosition h=\(horizontal)")
|
||||
self.adjustTextPosition(horizontal)
|
||||
}
|
||||
|
||||
if vertical != 0 {
|
||||
self.applyVerticalCursorSteps(vertical)
|
||||
}
|
||||
|
||||
if self.pendingHorizontalCursorSteps != 0 || self.pendingVerticalCursorSteps != 0 {
|
||||
self.scheduleCursorMoveFlush()
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
private func applyVerticalCursorSteps(_ steps: Int) {
|
||||
let direction = steps > 0 ? 1 : -1
|
||||
var remaining = abs(steps)
|
||||
let chunk = Self.cursorVerticalChunkSize
|
||||
|
||||
while remaining > 0 {
|
||||
adjustTextPosition(direction * chunk)
|
||||
cursorLineHaptic.impactOccurred()
|
||||
cursorLineHaptic.prepare()
|
||||
remaining -= 1
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,124 @@
|
||||
// KeyboardConfigSync.swift
|
||||
// OSGKeyboard · Keyboard Extension
|
||||
//
|
||||
// App Group config hydration, Darwin observers, and onboarding mirroring.
|
||||
|
||||
import Foundation
|
||||
import OSGKeyboardShared
|
||||
|
||||
@MainActor
|
||||
final class KeyboardConfigSync {
|
||||
private let state: KeyboardState
|
||||
private let persistor: AppGroupPersistor
|
||||
private let onFlowSessionChanged: () -> Void
|
||||
|
||||
/// Grace period after a chip-side translation write during which the
|
||||
/// 1 Hz App Group poll must not overwrite `translationTargetLocaleId`.
|
||||
var translationConfigProtectedUntil: Date?
|
||||
|
||||
private var flowSessionDarwinObserver: FlowSessionDarwinObserver?
|
||||
private var transcriptionDarwinObserver: FlowSessionDarwinObserver?
|
||||
private var configDarwinObserver: FlowSessionDarwinObserver?
|
||||
|
||||
init(
|
||||
state: KeyboardState,
|
||||
persistor: AppGroupPersistor,
|
||||
onFlowSessionChanged: @escaping () -> Void
|
||||
) {
|
||||
self.state = state
|
||||
self.persistor = persistor
|
||||
self.onFlowSessionChanged = onFlowSessionChanged
|
||||
}
|
||||
|
||||
func installDarwinObservers() {
|
||||
flowSessionDarwinObserver = FlowSessionDarwinObserver { [weak self] in
|
||||
self?.onFlowSessionChanged()
|
||||
}
|
||||
transcriptionDarwinObserver = FlowSessionDarwinObserver(
|
||||
notificationName: FlowSessionDarwin.transcriptionNotificationName
|
||||
) { [weak self] in
|
||||
self?.onFlowSessionChanged()
|
||||
}
|
||||
configDarwinObserver = FlowSessionDarwinObserver(
|
||||
notificationName: AppGroupConfigDarwin.notificationName
|
||||
) { [weak self] in
|
||||
self?.refreshConfigFromAppGroup()
|
||||
}
|
||||
}
|
||||
|
||||
func loadPersistedConfig() -> AppGroupLoadResult {
|
||||
switch persistor.load(into: state) {
|
||||
case .loaded:
|
||||
OSGLog.keyboardExt.info(
|
||||
"config loaded — cursorDragNavigationEnabled=\(self.state.cursorDragNavigationEnabled)"
|
||||
)
|
||||
syncOnboardingStateFromAppGroup()
|
||||
return .loaded
|
||||
case .unavailable:
|
||||
state.phase = .error(
|
||||
.appGroupUnavailable,
|
||||
message: ExtL10n.string("keyboard.error.appGroupUnavailable")
|
||||
)
|
||||
return .unavailable
|
||||
}
|
||||
}
|
||||
|
||||
func refreshConfigFromAppGroup() {
|
||||
persistor.refreshRuntimeFlags(
|
||||
into: state,
|
||||
protectTranslationUntil: translationConfigProtectedUntil
|
||||
)
|
||||
}
|
||||
|
||||
func syncOnboardingStateFromAppGroup() {
|
||||
let store = AppGroupStore()
|
||||
state.hasCompletedOnboarding = store.hasCompletedOnboarding
|
||||
state.onboardingPage = store.onboardingPage
|
||||
}
|
||||
|
||||
func autoAdvancePastKeyboardSetupStepIfNeeded() {
|
||||
guard !state.hasCompletedOnboarding else { return }
|
||||
guard state.onboardingPage == 3 else { return }
|
||||
guard KeyboardSetupBridge.isReadyForOnboardingSkip else { return }
|
||||
let store = AppGroupStore()
|
||||
store.setOnboardingPage(4)
|
||||
state.onboardingPage = 4
|
||||
}
|
||||
|
||||
func advanceOnboarding() {
|
||||
let store = AppGroupStore()
|
||||
let nextPage = min(4, store.onboardingPage + 1)
|
||||
store.setOnboardingPage(nextPage)
|
||||
state.onboardingPage = nextPage
|
||||
}
|
||||
|
||||
func completeOnboarding() {
|
||||
let store = AppGroupStore()
|
||||
store.setHasCompletedOnboarding(true)
|
||||
store.setOnboardingPage(4)
|
||||
state.hasCompletedOnboarding = true
|
||||
state.onboardingPage = 4
|
||||
}
|
||||
|
||||
func persistLocale(_ id: String) {
|
||||
state.localeId = id
|
||||
persistor.persist(localeId: id)
|
||||
}
|
||||
|
||||
func persistEngineMode(_ mode: String) {
|
||||
state.engineMode = mode
|
||||
persistor.persist(engineMode: mode)
|
||||
}
|
||||
|
||||
func persistTranslationTargetLocaleId(_ id: String) {
|
||||
let resolved = TranslationLanguageCatalog.resolve(id).id
|
||||
state.translationTargetLocaleId = resolved
|
||||
translationConfigProtectedUntil = Date().addingTimeInterval(2.5)
|
||||
persistor.persist(translationTargetLocaleId: resolved)
|
||||
}
|
||||
|
||||
func persistMode(_ mode: KeyboardState.InputMode) {
|
||||
state.mode = mode
|
||||
persistor.persist(mode: mode)
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,412 @@
|
||||
// KeyboardFlowCoordinator.swift
|
||||
// OSGKeyboard · Keyboard Extension
|
||||
//
|
||||
// Flow session start, recording, watchdogs, and result delivery handling.
|
||||
|
||||
import UIKit
|
||||
import OSGKeyboardShared
|
||||
|
||||
@MainActor
|
||||
final class KeyboardFlowCoordinator {
|
||||
private enum FlowWatchdog {
|
||||
static let pollIntervalNs: UInt64 = 200_000_000
|
||||
/// Give the user time to manually open the host app when auto-jump fails.
|
||||
static let startTimeout: TimeInterval = 30
|
||||
|
||||
static func resultTimeout(engineMode: String) -> TimeInterval {
|
||||
FlowSessionKeys.keyboardResultTimeout(engineMode: engineMode)
|
||||
}
|
||||
}
|
||||
|
||||
private let state: KeyboardState
|
||||
private let textInserter: KeyboardTextInserter
|
||||
private let hasFullAccess: () -> Bool
|
||||
private let wakeLockView: () -> UIView?
|
||||
private let openHostApp: (String) -> Void
|
||||
private let detectAndStoreAppContext: () -> Void
|
||||
private let scheduleAutoClearError: () -> Void
|
||||
private let refreshConfigFromAppGroup: () -> Void
|
||||
|
||||
private var isPendingFlowStart = false
|
||||
private var flowStartDeadline: TimeInterval = 0
|
||||
private var isFlowRecording = false
|
||||
private var flowWatchdogTask: Task<Void, Never>?
|
||||
private var utteranceTimerTask: Task<Void, Never>?
|
||||
private var utteranceStartedAt: TimeInterval = 0
|
||||
private var wasFlowSessionActive = false
|
||||
private var flowSessionMonitorTask: Task<Void, Never>?
|
||||
private var isAwaitingFlowResult = false
|
||||
private var lastFlowAutoStartAttempt: TimeInterval = 0
|
||||
private static let flowAutoStartCooldown: TimeInterval = 20
|
||||
|
||||
init(
|
||||
state: KeyboardState,
|
||||
textInserter: KeyboardTextInserter,
|
||||
hasFullAccess: @escaping () -> Bool,
|
||||
wakeLockView: @escaping () -> UIView?,
|
||||
openHostApp: @escaping (String) -> Void,
|
||||
detectAndStoreAppContext: @escaping () -> Void,
|
||||
scheduleAutoClearError: @escaping () -> Void,
|
||||
refreshConfigFromAppGroup: @escaping () -> Void
|
||||
) {
|
||||
self.state = state
|
||||
self.textInserter = textInserter
|
||||
self.hasFullAccess = hasFullAccess
|
||||
self.wakeLockView = wakeLockView
|
||||
self.openHostApp = openHostApp
|
||||
self.detectAndStoreAppContext = detectAndStoreAppContext
|
||||
self.scheduleAutoClearError = scheduleAutoClearError
|
||||
self.refreshConfigFromAppGroup = refreshConfigFromAppGroup
|
||||
}
|
||||
|
||||
var preservesLifecycleOnDisappear: Bool {
|
||||
isPendingFlowStart || isFlowRecording || isAwaitingFlowResult
|
||||
}
|
||||
|
||||
func startSessionMonitor() {
|
||||
flowSessionMonitorTask?.cancel()
|
||||
flowSessionMonitorTask = Task { @MainActor [weak self] in
|
||||
while !Task.isCancelled {
|
||||
self?.refreshSessionState()
|
||||
try? await Task.sleep(nanoseconds: 1_000_000_000)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
func stopSessionMonitor() {
|
||||
flowSessionMonitorTask?.cancel()
|
||||
flowSessionMonitorTask = nil
|
||||
}
|
||||
|
||||
func refreshSessionState() {
|
||||
refreshConfigFromAppGroup()
|
||||
refreshFlowPartialIfNeeded()
|
||||
consumePendingFlowDeliveryIfNeeded()
|
||||
|
||||
let active = FlowSessionBridge.isSessionActive()
|
||||
state.flowSessionActive = active
|
||||
|
||||
if wasFlowSessionActive && !active && !isFlowRecording && !isPendingFlowStart {
|
||||
switch state.phase {
|
||||
case .recording, .processing:
|
||||
break
|
||||
default:
|
||||
showFlowSessionExpiredHint()
|
||||
}
|
||||
}
|
||||
wasFlowSessionActive = active
|
||||
|
||||
if !active {
|
||||
maybeAutoStartFlowSession()
|
||||
}
|
||||
}
|
||||
|
||||
func toggleRecording() {
|
||||
switch state.phase {
|
||||
case .recording:
|
||||
pressEnded()
|
||||
case .idle, .denied, .error:
|
||||
pressBegan()
|
||||
case .requestingPermissions, .processing:
|
||||
break
|
||||
}
|
||||
}
|
||||
|
||||
func pressBegan() {
|
||||
switch state.phase {
|
||||
case .idle, .denied, .error:
|
||||
break
|
||||
default:
|
||||
return
|
||||
}
|
||||
guard !state.micDisabled else { return }
|
||||
guard hasFullAccess() else {
|
||||
let msg = ExtL10n.string("keyboard.error.fullAccessRequired")
|
||||
state.phase = .error(.fullAccessRequired, message: msg)
|
||||
scheduleAutoClearError()
|
||||
return
|
||||
}
|
||||
guard AppGroup.isAvailable else {
|
||||
let msg = ExtL10n.string("keyboard.error.appGroupCommunication")
|
||||
state.phase = .error(.appGroupUnavailable, message: msg)
|
||||
scheduleAutoClearError()
|
||||
return
|
||||
}
|
||||
|
||||
detectAndStoreAppContext()
|
||||
|
||||
if FlowSessionBridge.isSessionActive() {
|
||||
startFlowRecording()
|
||||
} else {
|
||||
beginFlowStart()
|
||||
}
|
||||
}
|
||||
|
||||
func pressEnded() {
|
||||
if isPendingFlowStart {
|
||||
cancelPendingFlowStart()
|
||||
return
|
||||
}
|
||||
guard isFlowRecording else { return }
|
||||
|
||||
isFlowRecording = false
|
||||
stopUtteranceCountdown()
|
||||
ExtensionScreenWakeLock.release()
|
||||
FlowSessionBridge.setRecordingState(.stopped)
|
||||
state.phase = .processing
|
||||
state.lastTranscript = ExtL10n.string("keyboard.flow.transcribing")
|
||||
startFlowResultWatchdog()
|
||||
}
|
||||
|
||||
func beginFlowStart() {
|
||||
guard !isPendingFlowStart else { return }
|
||||
isPendingFlowStart = true
|
||||
isFlowRecording = false
|
||||
flowStartDeadline = Date().timeIntervalSince1970 + FlowWatchdog.startTimeout
|
||||
state.lastTranscript = ExtL10n.string("keyboard.flow.startingSession")
|
||||
state.phase = .processing
|
||||
openHostApp("startflow")
|
||||
startFlowStartWatchdog()
|
||||
debug("beginFlowStart")
|
||||
}
|
||||
|
||||
func handleHostAppOpenResult(path: String, success: Bool) {
|
||||
debug("openHostApp path=\(path) success=\(success)")
|
||||
guard !success else { return }
|
||||
|
||||
if path == "startflow", isPendingFlowStart {
|
||||
state.lastTranscript = ExtL10n.string("keyboard.flow.manualOpenHost")
|
||||
return
|
||||
}
|
||||
|
||||
showManualOpenHint(path: path)
|
||||
}
|
||||
|
||||
func cancelPipelineUnlessAwaitingResult() {
|
||||
guard !isAwaitingFlowResult else { return }
|
||||
if isFlowRecording || isPendingFlowStart {
|
||||
if isFlowRecording {
|
||||
FlowSessionBridge.setRecordingState(.aborted)
|
||||
ExtensionScreenWakeLock.release()
|
||||
}
|
||||
isFlowRecording = false
|
||||
isPendingFlowStart = false
|
||||
stopUtteranceCountdown()
|
||||
stopFlowWatchdog()
|
||||
state.level = 0
|
||||
}
|
||||
}
|
||||
|
||||
// MARK: - Private
|
||||
|
||||
private func consumePendingFlowDeliveryIfNeeded() {
|
||||
if isAwaitingFlowResult {
|
||||
if let delivery = FlowSessionBridge.consumeTranscriptionDelivery() {
|
||||
isAwaitingFlowResult = false
|
||||
stopFlowWatchdog()
|
||||
textInserter.handleFlowTranscript(delivery)
|
||||
return
|
||||
}
|
||||
if let error = FlowSessionBridge.consumeTranscriptionError() {
|
||||
isAwaitingFlowResult = false
|
||||
stopFlowWatchdog()
|
||||
state.phase = .error(
|
||||
.fromFlowTranscription(error),
|
||||
message: error.message
|
||||
)
|
||||
scheduleAutoClearError()
|
||||
return
|
||||
}
|
||||
}
|
||||
|
||||
if isPendingFlowStart, FlowSessionBridge.isSessionActive() {
|
||||
completeFlowStartHandoff()
|
||||
}
|
||||
}
|
||||
|
||||
private func maybeAutoStartFlowSession() {
|
||||
guard !FlowSessionBridge.isSessionActive() else { return }
|
||||
guard !isPendingFlowStart, !isFlowRecording, !isAwaitingFlowResult else { return }
|
||||
guard hasFullAccess(), AppGroup.isAvailable else { return }
|
||||
guard case .idle = state.phase else { return }
|
||||
|
||||
let now = Date().timeIntervalSince1970
|
||||
guard now - lastFlowAutoStartAttempt >= Self.flowAutoStartCooldown else { return }
|
||||
lastFlowAutoStartAttempt = now
|
||||
beginFlowStart()
|
||||
}
|
||||
|
||||
private func showFlowSessionExpiredHint() {
|
||||
let message = ExtL10n.string("keyboard.flow.sessionExpired")
|
||||
state.phase = .error(.flowSessionExpired, message: message)
|
||||
scheduleAutoClearError()
|
||||
}
|
||||
|
||||
private func showManualOpenHint(path: String) {
|
||||
let msg: String
|
||||
if !hasFullAccess() {
|
||||
msg = ExtL10n.string("keyboard.error.fullAccessForJump")
|
||||
} else if path == "settings" {
|
||||
msg = ExtL10n.string("keyboard.error.manualOpenSettings")
|
||||
} else if path == "startflow" {
|
||||
msg = ExtL10n.string("keyboard.error.manualOpenForFlow")
|
||||
} else {
|
||||
msg = ExtL10n.string("keyboard.error.manualOpenSettings")
|
||||
}
|
||||
state.phase = .error(.manualOpenRequired, message: msg)
|
||||
scheduleAutoClearError()
|
||||
}
|
||||
|
||||
private func startFlowRecording() {
|
||||
isPendingFlowStart = false
|
||||
flowStartDeadline = 0
|
||||
stopFlowWatchdog()
|
||||
|
||||
FlowSessionBridge.setTranscriptionLanguage(state.localeId)
|
||||
FlowSessionBridge.setRecordingState(.recording)
|
||||
isFlowRecording = true
|
||||
state.lastTranscript = ""
|
||||
state.phase = .recording
|
||||
if let view = wakeLockView() {
|
||||
ExtensionScreenWakeLock.acquire(from: view)
|
||||
}
|
||||
startUtteranceCountdown()
|
||||
startFlowLevelWatchdog()
|
||||
debug("startFlowRecording")
|
||||
}
|
||||
|
||||
private func startUtteranceCountdown() {
|
||||
utteranceStartedAt = Date().timeIntervalSince1970
|
||||
state.utteranceRemainingSeconds = Int(FlowSessionKeys.maxUtteranceDuration)
|
||||
utteranceTimerTask?.cancel()
|
||||
utteranceTimerTask = Task { @MainActor [weak self] in
|
||||
while let self, self.isFlowRecording, !Task.isCancelled {
|
||||
let elapsed = Date().timeIntervalSince1970 - self.utteranceStartedAt
|
||||
let remaining = max(0, Int(ceil(FlowSessionKeys.maxUtteranceDuration - elapsed)))
|
||||
self.state.utteranceRemainingSeconds = remaining
|
||||
if remaining <= 0 {
|
||||
self.pressEnded()
|
||||
return
|
||||
}
|
||||
try? await Task.sleep(nanoseconds: 200_000_000)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
private func stopUtteranceCountdown() {
|
||||
utteranceTimerTask?.cancel()
|
||||
utteranceTimerTask = nil
|
||||
state.utteranceRemainingSeconds = Int(FlowSessionKeys.maxUtteranceDuration)
|
||||
}
|
||||
|
||||
private func cancelPendingFlowStart() {
|
||||
isPendingFlowStart = false
|
||||
flowStartDeadline = 0
|
||||
stopFlowWatchdog()
|
||||
state.phase = .idle
|
||||
state.lastTranscript = ""
|
||||
}
|
||||
|
||||
private func startFlowStartWatchdog() {
|
||||
stopFlowWatchdog()
|
||||
flowWatchdogTask = Task { @MainActor [weak self] in
|
||||
while let self, !Task.isCancelled, self.isPendingFlowStart {
|
||||
if FlowSessionBridge.isSessionActive() {
|
||||
self.completeFlowStartHandoff()
|
||||
return
|
||||
}
|
||||
let now = Date().timeIntervalSince1970
|
||||
if self.flowStartDeadline > 0, now > self.flowStartDeadline {
|
||||
self.isPendingFlowStart = false
|
||||
self.flowStartDeadline = 0
|
||||
self.showManualOpenHint(path: "startflow")
|
||||
return
|
||||
}
|
||||
try? await Task.sleep(nanoseconds: FlowWatchdog.pollIntervalNs)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
private func completeFlowStartHandoff() {
|
||||
isPendingFlowStart = false
|
||||
flowStartDeadline = 0
|
||||
stopFlowWatchdog()
|
||||
state.lastTranscript = ""
|
||||
state.phase = .idle
|
||||
refreshSessionState()
|
||||
debug("completeFlowStartHandoff")
|
||||
}
|
||||
|
||||
private func startFlowLevelWatchdog() {
|
||||
stopFlowWatchdog()
|
||||
flowWatchdogTask = Task { @MainActor [weak self] in
|
||||
while let self, !Task.isCancelled, self.isFlowRecording {
|
||||
let levels = FlowSessionBridge.audioLevels()
|
||||
if let peak = levels.max(), peak > 0 {
|
||||
self.state.level = Double(peak)
|
||||
}
|
||||
self.refreshFlowPartialIfNeeded()
|
||||
try? await Task.sleep(nanoseconds: FlowWatchdog.pollIntervalNs)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
private func refreshFlowPartialIfNeeded() {
|
||||
guard isFlowRecording || isAwaitingFlowResult else { return }
|
||||
switch state.phase {
|
||||
case .recording, .processing:
|
||||
if let partial = FlowSessionBridge.transcriptionPartial() {
|
||||
state.lastTranscript = partial
|
||||
}
|
||||
default:
|
||||
break
|
||||
}
|
||||
}
|
||||
|
||||
private func startFlowResultWatchdog() {
|
||||
stopFlowWatchdog()
|
||||
isAwaitingFlowResult = true
|
||||
let startedAt = Date().timeIntervalSince1970
|
||||
let resultTimeout = FlowWatchdog.resultTimeout(engineMode: state.engineMode)
|
||||
flowWatchdogTask = Task { @MainActor [weak self] in
|
||||
while let self, !Task.isCancelled {
|
||||
if let delivery = FlowSessionBridge.consumeTranscriptionDelivery() {
|
||||
self.isAwaitingFlowResult = false
|
||||
self.stopFlowWatchdog()
|
||||
self.textInserter.handleFlowTranscript(delivery)
|
||||
return
|
||||
}
|
||||
if let error = FlowSessionBridge.consumeTranscriptionError() {
|
||||
self.isAwaitingFlowResult = false
|
||||
self.stopFlowWatchdog()
|
||||
self.state.phase = .error(
|
||||
.fromFlowTranscription(error),
|
||||
message: error.message
|
||||
)
|
||||
self.scheduleAutoClearError()
|
||||
return
|
||||
}
|
||||
self.refreshFlowPartialIfNeeded()
|
||||
let now = Date().timeIntervalSince1970
|
||||
if now - startedAt > resultTimeout {
|
||||
self.isAwaitingFlowResult = false
|
||||
self.stopFlowWatchdog()
|
||||
let msg = ExtL10n.string("keyboard.flow.resultTimeout")
|
||||
self.state.phase = .error(.flowResultTimeout, message: msg)
|
||||
self.scheduleAutoClearError()
|
||||
return
|
||||
}
|
||||
try? await Task.sleep(nanoseconds: FlowWatchdog.pollIntervalNs)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
private func stopFlowWatchdog() {
|
||||
flowWatchdogTask?.cancel()
|
||||
flowWatchdogTask = nil
|
||||
}
|
||||
|
||||
private func debug(_ message: String) {
|
||||
OSGLog.keyboardExt.info("\(message, privacy: .public)")
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,44 @@
|
||||
// KeyboardTextInserter.swift
|
||||
// OSGKeyboard · Keyboard Extension
|
||||
//
|
||||
// Inserts Flow transcripts from the host app and surfaces polish warnings
|
||||
// without re-running LLM polish in the extension.
|
||||
|
||||
import OSGKeyboardShared
|
||||
|
||||
@MainActor
|
||||
final class KeyboardTextInserter {
|
||||
private let state: KeyboardState
|
||||
private let insertText: (String) -> Void
|
||||
private let scheduleAutoClearError: () -> Void
|
||||
|
||||
init(
|
||||
state: KeyboardState,
|
||||
insertText: @escaping (String) -> Void,
|
||||
scheduleAutoClearError: @escaping () -> Void
|
||||
) {
|
||||
self.state = state
|
||||
self.insertText = insertText
|
||||
self.scheduleAutoClearError = scheduleAutoClearError
|
||||
}
|
||||
|
||||
func handleFlowTranscript(_ delivery: TranscriptionDelivery) {
|
||||
let trimmed = delivery.text.trimmingCharacters(in: .whitespacesAndNewlines)
|
||||
guard !trimmed.isEmpty else {
|
||||
state.phase = .idle
|
||||
state.level = 0
|
||||
return
|
||||
}
|
||||
// Host app already polished when configured; keyboard only inserts.
|
||||
insertText(trimmed)
|
||||
state.lastTranscript = ""
|
||||
state.level = 0
|
||||
if let warning = delivery.polishWarning {
|
||||
state.phase = .error(.polishDegraded(warning), message: warning)
|
||||
scheduleAutoClearError()
|
||||
} else {
|
||||
state.phase = .idle
|
||||
}
|
||||
OSGLog.keyboardExt.info("flow insert length=\(trimmed.count, privacy: .public)")
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user