voice

type ChatFunc func(ctx context.Context, req llm.ChatRequest) (*llm.ChatResponse, error)

type Handler struct {
	// contains filtered or unexported fields
}

type Service interface {
	// CreateSession creates a new voice session.
	CreateSession(ctx context.Context, userID string, config *VoiceConfig) (*Session, error)
	// GetSession gets a session by ID.
	GetSession(ctx context.Context, sessionID string) (*Session, error)
	// UpdateSessionState updates the session state.
	UpdateSessionState(ctx context.Context, sessionID string, state SessionState) error
	// CloseSession closes a session.
	CloseSession(ctx context.Context, sessionID string) error
	// Transcribe transcribes audio to text.
	Transcribe(ctx context.Context, req *TranscribeRequest, audio []byte) (*TranscribeResponse, error)
	// Synthesize synthesizes text to speech.
	Synthesize(ctx context.Context, req *SynthesizeRequest) ([]byte, string, error)
	// SynthesizeStream synthesizes text and calls back with each audio chunk.
	// This enables true streaming: the first chunk arrives while later chunks are still being synthesized.
	SynthesizeStream(ctx context.Context, req *SynthesizeRequest, callback func(audio []byte, contentType string) error) error
	// SpeakLocally plays text through local audio output (blocking until done).
	// Returns false if the current provider doesn't support local playback.
	SpeakLocally(ctx context.Context, text string) (bool, error)
	// StopSpeaking stops any currently running local speech.
	StopSpeaking()
	// SetSTTService replaces the STT service at runtime.
	SetSTTService(svc stt.Service)
	// SetChatFunc sets the LLM chat function used by ProcessVoiceInput.
	SetChatFunc(fn ChatFunc)
	// ProcessVoiceInput processes voice input and returns a response.
	ProcessVoiceInput(ctx context.Context, sessionID string, text string) (string, error)
}

type ServiceConfig struct {
	STTService stt.Service
	TTSService tts.Service
}

type Session struct {
	// ID is the session identifier.
	ID string `json:"id"`
	// UserID is the user identifier.
	UserID string `json:"user_id"`
	// State is the current session state.
	State SessionState `json:"state"`
	// Language is the session language.
	Language string `json:"language"`
	// Voice is the TTS voice ID.
	Voice string `json:"voice"`
	// CreatedAt is when the session was created.
	CreatedAt time.Time `json:"created_at"`
	// LastActivity is the last activity time.
	LastActivity time.Time `json:"last_activity"`
	// ConversationID is the linked conversation ID.
	ConversationID string `json:"conversation_id,omitempty"`
}

type SessionState string

type SynthesizeRequest struct {
	// Text is the text to synthesize.
	Text string `json:"text"`
	// Format is the output format.
	Format string `json:"format,omitempty"`
}

type TranscribeRequest struct {
	// SessionID is the session identifier.
	SessionID string `json:"session_id,omitempty"`
	// Language is the language code.
	Language string `json:"language,omitempty"`
	// Format is the audio format.
	Format string `json:"format"`
}

type TranscribeResponse struct {
	// Text is the transcribed text.
	Text string `json:"text"`
	// Language is the detected language.
	Language string `json:"language,omitempty"`
	// Duration is the audio duration.
	Duration float64 `json:"duration,omitempty"`
	// Confidence is the confidence score.
	Confidence float64 `json:"confidence,omitempty"`
}

type VoiceConfig struct {
	// Language is the default language.
	Language string `json:"language"`
	// Voice is the default TTS voice.
	Voice string `json:"voice"`
	// WakeWord is the wake word (optional).
	WakeWord string `json:"wake_word,omitempty"`
	// WakeWordEnabled indicates if wake word detection is enabled.
	WakeWordEnabled bool `json:"wake_word_enabled"`
	// ContinuousListening indicates if continuous listening is enabled.
	ContinuousListening bool `json:"continuous_listening"`
	// AutoPlayResponse indicates if responses should auto-play.
	AutoPlayResponse bool `json:"auto_play_response"`
}

type VoiceMessage struct {
	// ID is the message identifier.
	ID string `json:"id"`
	// SessionID is the session identifier.
	SessionID string `json:"session_id"`
	// Role is the message role (user or assistant).
	Role string `json:"role"`
	// Text is the transcribed or generated text.
	Text string `json:"text"`
	// AudioURL is the URL to the audio file (if stored).
	AudioURL string `json:"audio_url,omitempty"`
	// Duration is the audio duration in seconds.
	Duration float64 `json:"duration,omitempty"`
	// Timestamp is when the message was created.
	Timestamp time.Time `json:"timestamp"`
}

type WSHandler struct {
	// contains filtered or unexported fields
}

type WebSocketMessage struct {
	// Type is the message type.
	Type string `json:"type"`
	// Data is the message data.
	Data interface{} `json:"data,omitempty"`
	// Error is the error message (if any).
	Error string `json:"error,omitempty"`
}