liwei1dao
/
deepvoice_client
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
Browse Source

add

weicu
wolfplus 2 years ago
parent
a6ab7812d1
commit
02cc915997
7 changed files with 281 additions and 268 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 154
      ios/Runner/AppDelegate.swift
  2. 41
      ios/Runner/AudioSessionManager.swift
  3. 95
      ios/Runner/AzureAsrHelper.swift
  4. 85
      ios/Runner/AzureTtsHelper.swift
  5. 150
      ios/Runner/VoiceInteractionService.swift
  6. 4
      lib/data/services/speech_impl/azure_asr_service.dart
  7. 20
      macos/Podfile.lock

154
ios/Runner/AppDelegate.swift
View File

@ -188,10 +188,7 @@ import AVFoundation
)
// 初始化Azure ASR辅助类
azureAsrHelper = AzureAsrHelper { [weak self] eventName, eventData in
// 推送事件到Flutter
self?.sendAsrEvent(eventName: eventName, eventData: eventData)
}
azureAsrHelper = AzureAsrHelper()
// 设置ASR方法处理器
azureAsrChannel.setMethodCallHandler { [weak self] (call, result) in
@ -246,7 +243,7 @@ import AVFoundation
)
// 创建ASR流处理器
let asrStreamHandler = AsrStreamHandler()
let asrStreamHandler = AsrStreamHandler(azureAsrHelper: azureAsrHelper as! AzureAsrHelper)
azureAsrStreamChannel.setStreamHandler(asrStreamHandler)
}
@ -260,10 +257,7 @@ import AVFoundation
)
// 初始化Azure TTS辅助类
azureTtsHelper = AzureTtsHelper { [weak self] eventName, eventData in
// 推送事件到Flutter
self?.sendTtsEvent(eventName: eventName, eventData: eventData)
}
azureTtsHelper = AzureTtsHelper()
// 设置TTS方法处理器
azureTtsChannel.setMethodCallHandler { [weak self] (call, result) in
@ -343,7 +337,7 @@ import AVFoundation
)
// 创建TTS流处理器
let ttsStreamHandler = TtsStreamHandler()
let ttsStreamHandler = TtsStreamHandler(azureTtsHelper: azureTtsHelper as! AzureTtsHelper)
azureTtsStreamChannel.setStreamHandler(ttsStreamHandler)
}
@ -487,44 +481,6 @@ import AVFoundation
let voiceInteractionStreamHandler = VoiceInteractionStreamHandler(voiceInteractionService: voiceInteractionService!)
voiceInteractionStreamChannel.setStreamHandler(voiceInteractionStreamHandler)
}
// 存储ASR事件
private var _cachedAsrEvent: [String: Any]?
// 发送ASR事件
@available(iOS 13.0, *)
private func sendAsrEvent(eventName: String, eventData: [String: Any]) {
// 调整格式：将event改为type，并将data的内容提升到顶层
var eventMap = eventData
eventMap["type"] = eventName
_cachedAsrEvent = eventMap
NotificationCenter.default.post(
name: Notification.Name("AzureAsrEvent"),
object: nil,
userInfo: eventMap
)
}
// 存储TTS事件
private var _cachedTtsEvent: [String: Any]?
// 发送TTS事件
@available(iOS 13.0, *)
private func sendTtsEvent(eventName: String, eventData: [String: Any]) {
// 调整格式：将event改为type，并将data的内容提升到顶层
var eventMap = eventData
eventMap["type"] = eventName
_cachedTtsEvent = eventMap
NotificationCenter.default.post(
name: Notification.Name("AzureTtsEvent"),
object: nil,
userInfo: eventMap
)
}
}
// 蓝牙事件流处理器
@ -596,92 +552,90 @@ class MediaButtonStreamHandler: NSObject, FlutterStreamHandler, MediaButtonEvent
}
}
// ASR事件处理器协议
@available(iOS 13.0, *)
protocol AsrEventHandler: AnyObject {
func sendEvent(_ event: [String: Any])
}
// ASR事件流处理器
@available(iOS 13.0, *)
class AsrStreamHandler: NSObject, FlutterStreamHandler {
class AsrStreamHandler: NSObject, FlutterStreamHandler, AsrEventHandler {
private let azureAsrHelper: AzureAsrHelper
private var eventSink: FlutterEventSink?
private var notificationObserver: Any?
deinit {
if let observer = notificationObserver {
NotificationCenter.default.removeObserver(observer)
init(azureAsrHelper: AzureAsrHelper) {
self.azureAsrHelper = azureAsrHelper
super.init()
// 设置事件处理器
azureAsrHelper.setEventHandler { [weak self] event in
self?.sendEvent(event)
}
}
func onListen(withArguments arguments: Any?, eventSink events: @escaping FlutterEventSink) -> FlutterError? {
NSLog("AsrStreamHandler: 开始监听ASR事件")
self.eventSink = events
// 注册通知观察者
notificationObserver = NotificationCenter.default.addObserver(
forName: Notification.Name("AzureAsrEvent"),
object: nil,
queue: .main) { [weak self] notification in
if let userInfo = notification.userInfo as? [String: Any], let eventSink = self?.eventSink {
// 确保在主线程上发送事件
DispatchQueue.main.async {
eventSink(userInfo)
}
}
}
return nil
}
func onCancel(withArguments arguments: Any?) -> FlutterError? {
NSLog("AsrStreamHandler: 取消监听ASR事件")
eventSink = nil
// 移除通知观察者
if let observer = notificationObserver {
NotificationCenter.default.removeObserver(observer)
notificationObserver = nil
}
return nil
}
// 实现AsrEventHandler协议
func sendEvent(_ event: [String: Any]) {
NSLog("AsrStreamHandler: 收到事件 \(event)")
DispatchQueue.main.async { [weak self] in
self?.eventSink?(event)
}
}
}
// TTS事件处理器协议
@available(iOS 13.0, *)
protocol TtsEventHandler: AnyObject {
func sendEvent(_ event: [String: Any])
}
// TTS事件流处理器
@available(iOS 13.0, *)
class TtsStreamHandler: NSObject, FlutterStreamHandler {
class TtsStreamHandler: NSObject, FlutterStreamHandler, TtsEventHandler {
private let azureTtsHelper: AzureTtsHelper
private var eventSink: FlutterEventSink?
private var notificationObserver: Any?
deinit {
if let observer = notificationObserver {
NotificationCenter.default.removeObserver(observer)
init(azureTtsHelper: AzureTtsHelper) {
self.azureTtsHelper = azureTtsHelper
super.init()
// 设置事件处理器
azureTtsHelper.setEventHandler { [weak self] event in
self?.sendEvent(event)
}
}
func onListen(withArguments arguments: Any?, eventSink events: @escaping FlutterEventSink) -> FlutterError? {
NSLog("TtsStreamHandler: 开始监听TTS事件")
self.eventSink = events
// 注册通知观察者
notificationObserver = NotificationCenter.default.addObserver(
forName: Notification.Name("AzureTtsEvent"),
object: nil,
queue: .main) { [weak self] notification in
if let userInfo = notification.userInfo as? [String: Any], let eventSink = self?.eventSink {
// 确保在主线程上发送事件
DispatchQueue.main.async {
eventSink(userInfo)
}
}
}
return nil
}
func onCancel(withArguments arguments: Any?) -> FlutterError? {
NSLog("TtsStreamHandler: 取消监听TTS事件")
eventSink = nil
// 移除通知观察者
if let observer = notificationObserver {
NotificationCenter.default.removeObserver(observer)
notificationObserver = nil
}
return nil
}
// 实现TtsEventHandler协议
func sendEvent(_ event: [String: Any]) {
NSLog("TtsStreamHandler: 收到事件 \(event)")
DispatchQueue.main.async { [weak self] in
self?.eventSink?(event)
}
}
}
// 火山AI事件流处理器

41
ios/Runner/AudioSessionManager.swift
View File

@ -419,40 +419,6 @@ import UIKit
// MARK: - 蓝牙媒体按钮配置
/// 配置用于获取蓝牙媒体按钮响应的音频会话
/// 此函数专门针对需要接收蓝牙耳机上媒体按钮事件的场景进行优化
/// - Returns: 配置是否成功
@objc func configureForBluetoothMediaButtons() -> Bool {
do {
// 配置音频会话以支持蓝牙媒体按钮
// 使用playback类别确保可以接收远程控制事件
// 使用spokenAudio模式优化语音交互
try configureSession(
category: .playback,
mode: .default, // 使用default模式，避免与spokenAudio可能的冲突
options: [
.allowBluetooth, // 允许通过蓝牙SCO连接进行输入和输出
.allowBluetoothA2DP, // 允许通过蓝牙A2DP连接进行高质量音频输出
.mixWithOthers // 允许与其他音频混合，避免中断
],
force: true // 强制重新配置，确保设置生效
)
NSLog("[AudioSessionManager] 已配置蓝牙媒体按钮响应的音频会话")
return true
} catch {
NSLog("[AudioSessionManager] 配置蓝牙媒体按钮响应的音频会话失败: \(error.localizedDescription)")
return false
}
}
// MARK: - Azure语音服务配置
/// 配置用于语音交互(同时支持ASR和TTS)的音频会话
/// 此函数综合优化语音识别和语音合成，适用于需要双向交互的场景
@ -466,11 +432,10 @@ import UIKit
category: .playAndRecord,
mode: .spokenAudio,
options: [
.allowBluetoothA2DP, // 允许通过蓝牙A2DP连接进行高质量音频输出
// .allowBluetooth, // 允许通过蓝牙SCO连接进行输入和输出
.allowBluetoothA2DP, // 允许通过蓝牙A2DP连接进行高质量音频输出
// .defaultToSpeaker, // 默认使用扬声器输出
// .duckOthers, // 降低其他应用的音量
// .interruptSpokenAudioAndMixWithOthers // 中断其他语音音频并混合
// .mixWithOthers
],
force: force // 强制重新配置，确保设置生效
)
@ -478,7 +443,7 @@ import UIKit
// 设置首选输入设备
// try configurePreferredInput()
// NSLog("[AudioSessionManager] 已配置语音交互(ASR+TTS)的音频会话")
NSLog("[AudioSessionManager] 已配置语音交互(ASR+TTS)的音频会话")
return true
} catch {
NSLog("[AudioSessionManager] 配置语音交互(ASR+TTS)的音频会话失败: \(error.localizedDescription)")

95
ios/Runner/AzureAsrHelper.swift
View File

@ -8,7 +8,7 @@ class AzureAsrHelper: NSObject {
// MARK: - 属性
/// 事件处理回调
private var eventHandler: (String, [String: Any]) -> Void
private var eventHandler: (([String: Any]) -> Void)?
/// 语音配置信息
private var speechSubscriptionKey: String = ""
@ -33,14 +33,14 @@ class AzureAsrHelper: NSObject {
// MARK: - 初始化
init(eventHandler: @escaping (String, [String: Any]) -> Void) {
self.eventHandler = eventHandler
override init() {
super.init()
}
deinit {
dispose()
}
// MARK: - ASR Service 接口实现
@ -56,10 +56,15 @@ class AzureAsrHelper: NSObject {
// 检查配置是否为空
if speechSubscriptionKey.isEmpty || serviceRegion.isEmpty {
print("[AzureAsrHelper] 错误: Azure 配置信息不完整")
eventHandler("error", ["message": "Azure 配置信息不完整"])
eventHandler?(["type": "error", "message": "Azure 配置信息不完整"])
return false
}
// 检查语言参数
if let languages = supportedLanguages, languages.isEmpty {
print("[AzureAsrHelper] 警告: 传入的支持语言列表为空，将使用默认语言")
}
// 释放之前的资源
dispose()
@ -134,7 +139,7 @@ class AzureAsrHelper: NSObject {
return true
} catch {
print("[AzureAsrHelper] 错误: 创建识别器失败: \(error.localizedDescription)")
eventHandler("error", ["message": "创建识别器失败: \(error.localizedDescription)"])
eventHandler?(["type": "error", "message": "创建识别器失败: \(error.localizedDescription)"])
return false
}
}
@ -158,7 +163,7 @@ class AzureAsrHelper: NSObject {
if event.result.reason == SPXResultReason.recognizedSpeech {
let detectedLanguage = self.getDetectedLanguage(from: event.result)
print("[AzureAsrHelper] 识别结果: \(event.result.text ?? ""), 语言: \(detectedLanguage)")
self.eventHandler("result", [
self.eventHandler?(["type": "result",
"text": event.result.text ?? "",
"detectedLanguage": detectedLanguage
])
@ -172,7 +177,7 @@ class AzureAsrHelper: NSObject {
if event.result.reason == SPXResultReason.recognizingSpeech {
let detectedLanguage = self.getDetectedLanguage(from: event.result)
// print("[AzureAsrHelper] 识别中: \(event.result.text ?? ""), 语言: \(detectedLanguage)")
self.eventHandler("recognizing", [
self.eventHandler?(["type": "recognizing",
"text": event.result.text ?? "",
"detectedLanguage": detectedLanguage
])
@ -185,7 +190,7 @@ class AzureAsrHelper: NSObject {
print("[AzureAsrHelper] 识别会话已开始")
self._isContinuousRecognitionActive = true
self.eventHandler("sessionStarted", [:])
self.eventHandler?(["type": "sessionStarted"])
}
recognizer.addSessionStoppedEventHandler { [weak self] _, _ in
@ -193,7 +198,7 @@ class AzureAsrHelper: NSObject {
print("[AzureAsrHelper] 识别会话已结束")
self._isContinuousRecognitionActive = false
self.eventHandler("sessionStopped", [:])
self.eventHandler?(["type": "sessionStopped"])
}
// 取消事件
@ -205,10 +210,7 @@ class AzureAsrHelper: NSObject {
print("[AzureAsrHelper] 识别取消: \(errorDetails)")
self.eventHandler("canceled", [
"reason": reason,
"errorDetails": errorDetails
])
self.eventHandler?(["type": "canceled", "reason": reason, "errorDetails": errorDetails])
self._isContinuousRecognitionActive = false
}
@ -217,9 +219,17 @@ class AzureAsrHelper: NSObject {
/// 执行一次性语音识别
/// - Returns: 是否成功启动识别
func recognizeOnce() -> Bool {
// 检查是否初始化
if !isInitialized {
print("[AzureAsrHelper] 错误: 语音服务未初始化")
eventHandler("error", ["message": "语音服务未初始化"])
eventHandler?(["type": "error", "message": "语音服务未初始化"])
return false
}
// 检查参数有效性
if speechSubscriptionKey.isEmpty || serviceRegion.isEmpty {
print("[AzureAsrHelper] 错误: Azure 配置信息不完整")
eventHandler?(["type": "error", "message": "Azure 配置信息不完整"])
return false
}
@ -235,7 +245,7 @@ class AzureAsrHelper: NSObject {
do {
// 通知会话开始
eventHandler("sessionStarted", [:])
eventHandler?(["type": "sessionStarted"])
// 执行识别
try recognizer?.recognizeOnceAsync { [weak self] result in
@ -243,21 +253,21 @@ class AzureAsrHelper: NSObject {
if result.reason == SPXResultReason.recognizedSpeech {
let detectedLanguage = self.getDetectedLanguage(from: result)
self.eventHandler("result", [
self.eventHandler?(["type": "result",
"text": result.text ?? "",
"detectedLanguage": detectedLanguage
])
} else if result.reason == SPXResultReason.noMatch {
print("[AzureAsrHelper] 无匹配结果")
self.eventHandler("noMatch", [:])
self.eventHandler?(["type": "noMatch"])
} else if result.reason == SPXResultReason.canceled {
do {
let details = try SPXCancellationDetails(fromCanceledRecognitionResult: result)
let errorDetails = details.errorDetails ?? "未知错误"
self.eventHandler("error", ["message": "识别取消: \(errorDetails)"])
self.eventHandler?(["type": "error", "message": "识别取消: \(errorDetails)"])
} catch {
print("[AzureAsrHelper] 错误: 获取取消详情失败: \(error.localizedDescription)")
self.eventHandler("error", ["message": "识别取消，无法获取详细原因"])
self.eventHandler?(["type": "error", "message": "识别取消，无法获取详细原因"])
}
}
}
@ -265,7 +275,7 @@ class AzureAsrHelper: NSObject {
return true
} catch {
print("[AzureAsrHelper] 错误: 识别异常: \(error.localizedDescription)")
eventHandler("error", ["message": "识别异常: \(error.localizedDescription)"])
eventHandler?(["type": "error", "message": "识别异常: \(error.localizedDescription)"])
return false
}
}
@ -273,9 +283,17 @@ class AzureAsrHelper: NSObject {
/// 开始连续语音识别
/// - Returns: 是否成功启动识别
func startContinuousRecognition() -> Bool {
// 检查是否初始化
if !isInitialized {
print("[AzureAsrHelper] 错误: 语音服务未初始化")
eventHandler("error", ["message": "语音服务未初始化"])
eventHandler?(["type": "error", "message": "语音服务未初始化"])
return false
}
// 检查参数有效性
if speechSubscriptionKey.isEmpty || serviceRegion.isEmpty {
print("[AzureAsrHelper] 错误: Azure 配置信息不完整")
eventHandler?(["type": "error", "message": "Azure 配置信息不完整"])
return false
}
@ -289,7 +307,7 @@ class AzureAsrHelper: NSObject {
if recognizer == nil {
print("[AzureAsrHelper] 尝试重新创建识别器...")
if !createRecognizerAndSetupCallbacks() {
eventHandler("error", ["message": "无法创建语音识别器"])
eventHandler?(["type": "error", "message": "无法创建语音识别器"])
return false
}
}
@ -309,7 +327,7 @@ class AzureAsrHelper: NSObject {
print("[AzureAsrHelper] 错误: 开始连续识别失败: \(error.localizedDescription)")
// 发送错误通知
eventHandler("error", ["message": "开始连续识别失败: \(error.localizedDescription)"])
eventHandler?(["type": "error", "message": "开始连续识别失败: \(error.localizedDescription)"])
return false
}
@ -319,6 +337,13 @@ class AzureAsrHelper: NSObject {
/// 停止连续语音识别
/// - Returns: 是否成功停止识别
func stopContinuousRecognition() -> Bool {
// 检查是否初始化
if !isInitialized {
print("[AzureAsrHelper] 错误: 语音服务未初始化")
eventHandler?(["type": "error", "message": "语音服务未初始化"])
return false
}
if !_isContinuousRecognitionActive {
print("[AzureAsrHelper] 未进行连续识别，忽略停止请求")
return true
@ -330,7 +355,7 @@ class AzureAsrHelper: NSObject {
guard let recognizer = recognizer else {
print("[AzureAsrHelper] 警告: 识别器为空，但状态显示活跃")
_isContinuousRecognitionActive = false
eventHandler("sessionStopped", [:])
eventHandler?(["type": "sessionStopped"])
return true
}
@ -355,7 +380,7 @@ class AzureAsrHelper: NSObject {
// 主线程回调通知结果
DispatchQueue.main.async {
print("[AzureAsrHelper] 连续识别已停止")
self.eventHandler("sessionStopped", [:])
self.eventHandler?(["type": "sessionStopped"])
}
} catch {
// 主线程回调通知错误
@ -363,7 +388,7 @@ class AzureAsrHelper: NSObject {
print("[AzureAsrHelper] 错误: 停止连续识别失败: \(error.localizedDescription)")
// 停止失败，恢复状态
self._isContinuousRecognitionActive = true
self.eventHandler("error", ["error": "停止识别失败: \(error.localizedDescription)"])
self.eventHandler?(["type": "error", "message": "停止识别失败: \(error.localizedDescription)"])
}
}
}
@ -380,9 +405,17 @@ class AzureAsrHelper: NSObject {
/// 释放资源
func dispose() {
// 停止所有识别操作
// 尝试停止所有识别操作
if _isContinuousRecognitionActive {
stopContinuousRecognition()
do {
if let recognizer = recognizer {
try recognizer.stopContinuousRecognition()
print("[AzureAsrHelper] 连续识别已停止")
}
} catch {
print("[AzureAsrHelper] 警告: 停止连续识别失败: \(error.localizedDescription)")
}
_isContinuousRecognitionActive = false
}
// 释放识别器
@ -392,7 +425,6 @@ class AzureAsrHelper: NSObject {
// 重置状态
isInitialized = false
_isContinuousRecognitionActive = false
print("[AzureAsrHelper] 资源已释放")
}
@ -411,4 +443,9 @@ class AzureAsrHelper: NSObject {
return currentLanguage
}
}
/// 设置事件处理器
@objc func setEventHandler(_ handler: @escaping ([String: Any]) -> Void) {
self.eventHandler = handler
}
}

85
ios/Runner/AzureTtsHelper.swift
View File

@ -8,7 +8,7 @@ class AzureTtsHelper: NSObject {
// MARK: - 属性
/// 事件处理回调
private var eventHandler: (String, [String: Any]) -> Void
private var eventHandler: (([String: Any]) -> Void)?
/// 语音配置信息
private var speechSubscriptionKey: String = ""
@ -54,8 +54,7 @@ class AzureTtsHelper: NSObject {
// MARK: - 初始化
init(eventHandler: @escaping (String, [String: Any]) -> Void) {
self.eventHandler = eventHandler
override init() {
super.init()
}
@ -63,6 +62,11 @@ class AzureTtsHelper: NSObject {
dispose()
}
/// 设置事件处理器
@objc func setEventHandler(_ handler: @escaping ([String: Any]) -> Void) {
self.eventHandler = handler
}
// MARK: - TTS 接口实现
/// 初始化语音合成服务
@ -77,7 +81,7 @@ class AzureTtsHelper: NSObject {
// 检查配置是否为空
if speechSubscriptionKey.isEmpty || serviceRegion.isEmpty {
print("[AzureTtsHelper] 错误: Azure 配置信息不完整")
eventHandler("error", ["error": "Azure 配置信息不完整"])
eventHandler?(["type": "error", "message": "Azure 配置信息不完整"])
return false
}
@ -114,7 +118,7 @@ class AzureTtsHelper: NSObject {
return true
} catch {
print("[AzureTtsHelper] 错误: 初始化语音合成服务失败: \(error.localizedDescription)")
eventHandler("error", ["error": "初始化语音合成服务失败: \(error.localizedDescription)"])
eventHandler?(["type": "error", "message": "初始化语音合成服务失败: \(error.localizedDescription)"])
return false
}
}
@ -139,15 +143,23 @@ class AzureTtsHelper: NSObject {
/// - Parameter voiceName: 语音名称 (如 "zh-CN-XiaoxiaoNeural")
/// - Returns: 设置是否成功
func setVoice(voiceName: String) -> Bool {
// 检查是否初始化
if !isInitialized {
print("[AzureTtsHelper] 错误: TTS 引擎尚未初始化")
eventHandler("error", ["error": "TTS 引擎尚未初始化"])
eventHandler?(["type": "error", "message": "TTS 引擎尚未初始化"])
return false
}
// 检查参数有效性
if speechSubscriptionKey.isEmpty || serviceRegion.isEmpty {
print("[AzureTtsHelper] 错误: Azure 配置信息不完整")
eventHandler?(["type": "error", "message": "Azure 配置信息不完整"])
return false
}
if voiceName.isEmpty {
print("[AzureTtsHelper] 错误: 声音名称为空")
eventHandler("error", ["error": "声音名称不能为空"])
eventHandler?(["type": "error", "message": "声音名称不能为空"])
return false
}
@ -175,9 +187,17 @@ class AzureTtsHelper: NSObject {
/// - volume: 音量，范围 0 到 100，默认为 100
/// - Returns: 是否设置成功
func setSpeechParams(rate: Int = 0, pitch: Int = 0, volume: Int = 100) -> Bool {
// 检查是否初始化
if !isInitialized {
print("[AzureTtsHelper] 错误: TTS 引擎尚未初始化")
eventHandler("error", ["error": "TTS 引擎尚未初始化"])
eventHandler?(["type": "error", "message": "TTS 引擎尚未初始化"])
return false
}
// 检查参数有效性
if speechSubscriptionKey.isEmpty || serviceRegion.isEmpty {
print("[AzureTtsHelper] 错误: Azure 配置信息不完整")
eventHandler?(["type": "error", "message": "Azure 配置信息不完整"])
return false
}
@ -194,9 +214,17 @@ class AzureTtsHelper: NSObject {
/// - Parameter text: 要合成的文本
/// - Returns: 操作是否成功启动
func speakText(text: String) -> Bool {
// 检查是否初始化
if !isInitialized {
print("[AzureTtsHelper] 错误: TTS 引擎尚未初始化")
eventHandler("error", ["error": "TTS 引擎尚未初始化"])
eventHandler?(["type": "error", "message": "TTS 引擎尚未初始化"])
return false
}
// 检查参数有效性
if speechSubscriptionKey.isEmpty || serviceRegion.isEmpty {
print("[AzureTtsHelper] 错误: Azure 配置信息不完整")
eventHandler?(["type": "error", "message": "Azure 配置信息不完整"])
return false
}
@ -223,12 +251,12 @@ class AzureTtsHelper: NSObject {
private func speakSsmlInternal(text: String) -> Bool {
guard let synthesizer = synthesizer else {
print("[AzureTtsHelper] 错误: 合成器未初始化")
eventHandler("error", ["error": "合成器未初始化"])
eventHandler?(["type": "error", "message": "合成器未初始化"])
return false
}
_isSpeaking = true
eventHandler("started", [:])
eventHandler?(["type": "started"])
// Task {
do {
@ -239,7 +267,7 @@ class AzureTtsHelper: NSObject {
print("[AzureTtsHelper] 错误: 语音合成失败: \(error.localizedDescription)")
DispatchQueue.main.async {
self._isSpeaking = false
self.eventHandler("error", ["error": "语音合成失败: \(error.localizedDescription)"])
self.eventHandler?(["type": "error", "message": "语音合成失败: \(error.localizedDescription)"])
}
}
// }
@ -250,7 +278,15 @@ class AzureTtsHelper: NSObject {
/// 停止当前语音合成
/// - Returns: 操作是否成功
func stopSpeaking() -> Bool {
if !isInitialized || !_isSpeaking {
// 检查是否初始化
if !isInitialized {
print("[AzureTtsHelper] 错误: TTS 引擎尚未初始化")
eventHandler?(["type": "error", "message": "TTS 引擎尚未初始化"])
return false
}
if !_isSpeaking {
print("[AzureTtsHelper] 未进行语音播放，忽略停止请求")
return true
}
@ -258,12 +294,12 @@ class AzureTtsHelper: NSObject {
do {
try synthesizer?.stopSpeaking()
_isSpeaking = false
eventHandler("canceled", [:])
eventHandler?(["type": "canceled"])
print("[AzureTtsHelper] 已停止语音合成")
return true
} catch {
print("[AzureTtsHelper] 错误: 停止语音合成失败: \(error.localizedDescription)")
eventHandler("error", ["error": "停止语音合成失败: \(error.localizedDescription)"])
eventHandler?(["type": "error", "message": "停止语音合成失败: \(error.localizedDescription)"])
return false
}
}
@ -276,9 +312,17 @@ class AzureTtsHelper: NSObject {
/// 释放资源
func dispose() {
// 停止合成
// 尝试停止所有语音合成
if _isSpeaking {
stopSpeaking()
do {
if let synthesizer = synthesizer {
try synthesizer.stopSpeaking()
print("[AzureTtsHelper] 语音合成已停止")
}
} catch {
print("[AzureTtsHelper] 警告: 停止语音合成失败: \(error.localizedDescription)")
}
_isSpeaking = false
}
// 释放合成器
@ -287,7 +331,6 @@ class AzureTtsHelper: NSObject {
// 重置状态
isInitialized = false
_isSpeaking = false
isAudioSessionConfigured = false
print("[AzureTtsHelper] 资源已释放")
@ -310,7 +353,7 @@ class AzureTtsHelper: NSObject {
print("[AzureTtsHelper] 语音合成完成: 音频持续时间: \(e.result.audioDuration)")
DispatchQueue.main.async {
self._isSpeaking = false
self.eventHandler("completed", [:])
self.eventHandler?(["type": "completed"])
}
}
@ -330,14 +373,14 @@ class AzureTtsHelper: NSObject {
DispatchQueue.main.async {
self._isSpeaking = false
self.eventHandler("error", ["error": "语音合成取消: \(cancellationDetails.errorDetails ?? "未知错误")"])
self.eventHandler?(["type": "error", "message": "语音合成取消: \(cancellationDetails.errorDetails ?? "未知错误")"])
}
} catch {
print("[AzureTtsHelper] 获取取消详情时出错: \(error)")
DispatchQueue.main.async {
self._isSpeaking = false
self.eventHandler("error", ["error": "语音合成被取消"])
self.eventHandler?(["type": "error", "message": "语音合成被取消"])
}
}
}

150
ios/Runner/VoiceInteractionService.swift
View File

@ -162,15 +162,21 @@ class VoiceInteractionService: NSObject {
NSLog("%@: 初始化服务组件", TAG)
// 初始化Azure ASR
azureAsrHelper = AzureAsrHelper { [weak self] eventName, eventData in
// 处理ASR事件
self?.handleAsrEvent(eventName: eventName, eventData: eventData)
azureAsrHelper = AzureAsrHelper()
if let helper = azureAsrHelper {
helper.setEventHandler({ [weak self] eventData in
// 处理ASR事件
self?.handleAsrEvent(eventData: eventData)
})
}
// 初始化Azure TTS
azureTtsHelper = AzureTtsHelper { [weak self] eventName, eventData in
// 处理TTS事件
self?.handleTtsEvent(eventName: eventName, eventData: eventData)
azureTtsHelper = AzureTtsHelper()
if let helper = azureTtsHelper {
helper.setEventHandler({ [weak self] eventData in
// 处理TTS事件
self?.handleTtsEvent(eventData: eventData)
})
}
// 初始化火山AI服务
@ -237,69 +243,97 @@ class VoiceInteractionService: NSObject {
/**
* 处理ASR事件
*/
private func handleAsrEvent(eventName: String, eventData: [String: Any]) {
// NSLog("%@: 处理ASR事件: %@", TAG, eventName)
switch eventName {
case "recognizing":
if let text = eventData["text"] as? String, !text.isEmpty {
hasSpeechDetected = true
stopCurrentTTS()
updateLastActivityTime()
}
private func handleAsrEvent(eventData: [String: Any]) {
if let type = eventData["type"] as? String {
// NSLog("%@: 处理ASR事件: %@", TAG, type)
case "result":
if let text = eventData["text"] as? String, !text.isEmpty {
switch type {
case "recognizing":
if let text = eventData["text"] as? String, !text.isEmpty {
hasSpeechDetected = true
stopCurrentTTS()
updateLastActivityTime()
} else if let data = eventData["data"] as? [String: Any],
let text = data["text"] as? String,
!text.isEmpty {
hasSpeechDetected = true
stopCurrentTTS()
updateLastActivityTime()
}
case "result":
var text: String?
// 直接读取text或从data中读取
if let directText = eventData["text"] as? String, !directText.isEmpty {
text = directText
} else if let data = eventData["data"] as? [String: Any],
let dataText = data["text"] as? String,
!dataText.isEmpty {
text = dataText
}
if let finalText = text {
updateLastActivityTime()
// 保存用户消息
lastUserMessage = finalText
processWithVolcanoAI(finalText)
}
// 重置状态，继续识别
hasSpeechDetected = false
case "sessionStarted":
updateLastActivityTime()
// 保存用户消息
lastUserMessage = text
processWithVolcanoAI(text)
case "sessionStopped":
isRecognitionActive = false
case "canceled":
isRecognitionActive = false
case "error":
isRecognitionActive = false
if let message = eventData["message"] as? String {
NSLog("%@: ASR错误: %@", TAG, message)
}
playNotification("语音识别出错")
default:
break
}
// 重置状态，继续识别
hasSpeechDetected = false
case "sessionStarted":
updateLastActivityTime()
case "sessionStopped":
isRecognitionActive = false
case "canceled":
isRecognitionActive = false
case "error":
isRecognitionActive = false
playNotification("语音识别出错")
default:
break
}
}
/**
* 处理TTS事件
*/
private func handleTtsEvent(eventName: String, eventData: [String: Any]) {
// NSLog("%@: 处理TTS事件: %@", TAG, eventName)
switch eventName {
case "started":
isTtsSpeaking = true
private func handleTtsEvent(eventData: [String: Any]) {
if let type = eventData["type"] as? String {
// NSLog("%@: 处理TTS事件: %@", TAG, type)
case "completed":
isTtsSpeaking = false
updateLastActivityTime()
case "canceled":
isTtsSpeaking = false
case "error":
isTtsSpeaking = false
NSLog("%@: TTS错误: %@", TAG, eventData["error"] as? String ?? "未知错误")
default:
break
switch type {
case "started":
isTtsSpeaking = true
case "completed":
isTtsSpeaking = false
updateLastActivityTime()
case "canceled":
isTtsSpeaking = false
case "error":
isTtsSpeaking = false
if let message = eventData["message"] as? String {
NSLog("%@: TTS错误: %@", TAG, message)
}
default:
break
}
}
}

4
lib/data/services/speech_impl/azure_asr_service.dart
View File

@ -36,7 +36,6 @@ class AzureAsrService extends GetxService implements AsrService {
AzureAsrService() {
_loadConfig();
_setupEventChannel();
}
/// 从环境变量加载配置
@ -80,7 +79,8 @@ class AzureAsrService extends GetxService implements AsrService {
'region': _serviceRegion,
'supportedLanguages': _supportedLanguages,
});
_setupEventChannel();
_isInitialized = result;
Logger.info('Azure 语音识别服务初始化${result ? '成功' : '失败'}');
return result;

20
macos/Podfile.lock
View File

@ -1,9 +1,4 @@
PODS:
- app_settings (5.1.1):
- FlutterMacOS
- audio_service (0.0.1):
- Flutter
- FlutterMacOS
- audio_session (0.0.1):
- FlutterMacOS
- device_info_plus (0.0.1):
@ -32,16 +27,11 @@ PODS:
- Flutter
- FlutterMacOS
- Try
- sqflite_darwin (0.0.4):
- Flutter
- FlutterMacOS
- Try (2.1.1)
- window_size (0.0.2):
- FlutterMacOS
DEPENDENCIES:
- app_settings (from `Flutter/ephemeral/.symlinks/plugins/app_settings/macos`)
- audio_service (from `Flutter/ephemeral/.symlinks/plugins/audio_service/darwin`)
- audio_session (from `Flutter/ephemeral/.symlinks/plugins/audio_session/macos`)
- device_info_plus (from `Flutter/ephemeral/.symlinks/plugins/device_info_plus/macos`)
- file_selector_macos (from `Flutter/ephemeral/.symlinks/plugins/file_selector_macos/macos`)
@ -54,7 +44,6 @@ DEPENDENCIES:
- record_darwin (from `Flutter/ephemeral/.symlinks/plugins/record_darwin/macos`)
- share_plus (from `Flutter/ephemeral/.symlinks/plugins/share_plus/macos`)
- speech_to_text (from `Flutter/ephemeral/.symlinks/plugins/speech_to_text/darwin`)
- sqflite_darwin (from `Flutter/ephemeral/.symlinks/plugins/sqflite_darwin/darwin`)
- window_size (from `Flutter/ephemeral/.symlinks/plugins/window_size/macos`)
SPEC REPOS:
@ -62,10 +51,6 @@ SPEC REPOS:
- Try
EXTERNAL SOURCES:
app_settings:
:path: Flutter/ephemeral/.symlinks/plugins/app_settings/macos
audio_service:
:path: Flutter/ephemeral/.symlinks/plugins/audio_service/darwin
audio_session:
:path: Flutter/ephemeral/.symlinks/plugins/audio_session/macos
device_info_plus:
@ -90,14 +75,10 @@ EXTERNAL SOURCES:
:path: Flutter/ephemeral/.symlinks/plugins/share_plus/macos
speech_to_text:
:path: Flutter/ephemeral/.symlinks/plugins/speech_to_text/darwin
sqflite_darwin:
:path: Flutter/ephemeral/.symlinks/plugins/sqflite_darwin/darwin
window_size:
:path: Flutter/ephemeral/.symlinks/plugins/window_size/macos
SPEC CHECKSUMS:
app_settings: cd21e176b56f8172043640ade81322a98896bff4
audio_service: aa99a6ba2ae7565996015322b0bb024e1d25c6fd
audio_session: eaca2512cf2b39212d724f35d11f46180ad3a33e
device_info_plus: a56e6e74dbbd2bb92f2da12c64ddd4f67a749041
file_selector_macos: 6280b52b459ae6c590af5d78fc35c7267a3c4b31
@ -110,7 +91,6 @@ SPEC CHECKSUMS:
record_darwin: 30509266ae213af8afdb09a8ae7467cb64c1377e
share_plus: 510bf0af1a42cd602274b4629920c9649c52f4cc
speech_to_text: 9dc43a5df3cbc2813f8c7cc9bd0fbf94268ed7ac
sqflite_darwin: 20b2a3a3b70e43edae938624ce550a3cbf66a3d0
Try: 5ef669ae832617b3cee58cb2c6f99fb767a4ff96
window_size: 4bd15034e6e3d0720fd77928a7c42e5492cfece9

Write Preview
Loading…
Cancel
Save