agent

type Agent struct {
	ID           string
	Instructions string
	ChatCtx      *llm.ChatContext
	Tools        []llm.Tool

	TurnDetection TurnDetectionMode
	TurnDetector  TurnDetector
	STT           stt.STT
	VAD           vad.VAD
	LLM           llm.LLM
	TTS           tts.TTS

	AllowInterruptions        bool
	MinConsecutiveSpeechDelay float64
	UseTTSAlignedTranscript   bool
	MinEndpointingDelay       float64
	MaxEndpointingDelay       float64
	// contains filtered or unexported fields
}

type AgentActivity struct {
	AgentIntf AgentInterface
	Agent     *Agent
	Session   *AgentSession
	// contains filtered or unexported fields
}

type AgentInterface interface {
	OnEnter()
	OnExit()
	OnUserTurnCompleted(ctx context.Context, chatCtx *llm.ChatContext, newMsg *llm.ChatMessage) error
	GetAgent() *Agent
	GetActivity() *AgentActivity
}

type AgentSession struct {
	Options AgentSessionOptions

	ChatCtx   *llm.ChatContext
	Agent     AgentInterface
	STT       stt.STT
	VAD       vad.VAD
	LLM       llm.LLM
	TTS       tts.TTS
	Tools     []llm.Tool
	Assistant *PipelineAgent
	Room      *lksdk.Room

	MetricsCollector *telemetry.UsageCollector

	UserState  UserState
	AgentState AgentState

	// Transcript attribution — set by RoomIO when tracks are established.
	RemoteUserIdentity string
	RemoteTrackSID     string
	AgentTrackSID      string

	// Event channels
	AgentStateChangedCh chan AgentStateChangedEvent
	UserStateChangedCh  chan UserStateChangedEvent
	// contains filtered or unexported fields
}

type AgentSessionOptions struct {
	AllowInterruptions            bool
	DiscardAudioIfUninterruptible bool
	MinInterruptionDuration       float64
	MinInterruptionWords          int
	MinEndpointingDelay           float64
	MaxEndpointingDelay           float64
	MaxToolSteps                  int
	UserAwayTimeout               float64
	FalseInterruptionTimeout      float64
	ResumeFalseInterruption       bool
	MinConsecutiveSpeechDelay     float64
	UseTTSAlignedTranscript       bool
	PreemptiveGeneration          bool
	AECWarmupDuration             float64
}

type AgentState string

type AgentStateChangedEvent struct {
	OldState AgentState
	NewState AgentState
}

type AgentTask[T any] struct {
	Agent
	Result chan T
	Err    chan error
}

type AudioConfig struct {
	Source      AudioSource
	Volume      float64
	Probability float64
}

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

type AudioSource interface{} // Can be string, BuiltinAudioClip, or <-chan *model.AudioFrame

type Avatar struct {
	State AvatarState
}

type AvatarIO interface {
	SendAvatarData(ctx context.Context, data []byte) error
}

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

type AvatarState string

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

type BuiltinAudioClip string

type ClientEventPayload struct {
	Type  string `json:"type"`
	State string `json:"state,omitempty"`
}

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

type CloseEvent struct {
	Reason    CloseReason
	Error     error
	CreatedAt time.Time
}

type CloseReason string

type ConversationItemAddedEvent struct {
	Item      llm.ChatItem
	CreatedAt time.Time
}

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

type DtmfEvent struct {
	Digit string
	Time  time.Time
}

type EndOfTurnInfo struct {
	SkipReply            bool
	NewTranscript        string
	TranscriptConfidence float64
	StartedSpeakingAt    *float64
	StoppedSpeakingAt    *float64
}

type EvaluationResult struct {
	Judgments map[string]string
}

type Event interface {
	GetType() string
}

type IVRActivity struct {
	AgentIntf AgentInterface
	Agent     *Agent
	// contains filtered or unexported fields
}

type InputDetails struct {
	Modality string
}

type LLMGenerationData struct {
	TextCh     chan string
	FunctionCh chan *llm.FunctionToolCall
	FullTextCh chan string // receives the complete assembled text when streaming is done
	Usage      *llm.CompletionUsage
}

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

type MetricsCollectedEvent struct {
	Metrics   telemetry.AgentMetrics
	CreatedAt time.Time
}

type MultimodalAgent struct {
	PublishAudio func(frame *model.AudioFrame) error
	// contains filtered or unexported fields
}

type PipelineAgent struct {
	LLM llm.LLM

	PublishAudio func(frame *model.AudioFrame) error
	// contains filtered or unexported fields
}

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

type Plugin interface {
	Title() string
	Version() string
	Package() string
	DownloadFiles() error
}

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

type RecognitionHooks interface {
	OnStartOfSpeech(ev *vad.VADEvent)
	OnEndOfSpeech(ev *vad.VADEvent)
	OnFinalTranscript(ev *stt.SpeechEvent)
}

type RecordingOptions struct {
	Audio      bool `json:"audio"`
	Traces     bool `json:"traces"`
	Logs       bool `json:"logs"`
	Transcript bool `json:"transcript"`
}

type RunAssert struct {
	ChatCtx *llm.ChatContext
	// contains filtered or unexported fields
}

type RunContext struct {
	Session      *AgentSession
	SpeechHandle *SpeechHandle
	FunctionCall *llm.FunctionCall
}

type RunResult struct {
	ChatCtx *llm.ChatContext
	Expect  *RunAssert
}

type SessionReport struct {
	RecordingOptions        RecordingOptions    `json:"recording_options"`
	JobID                   string              `json:"job_id"`
	RoomID                  string              `json:"room_id"`
	Room                    string              `json:"room"`
	Options                 AgentSessionOptions `json:"options"`
	Events                  []any               `json:"events"`
	ChatHistory             *llm.ChatContext    `json:"chat_history"`
	AudioRecordingPath      *string             `json:"audio_recording_path,omitempty"`
	AudioRecordingStartedAt *float64            `json:"audio_recording_started_at,omitempty"`
	Duration                *float64            `json:"duration,omitempty"`
	StartedAt               *float64            `json:"started_at,omitempty"`
	Timestamp               float64             `json:"timestamp"`
}

type SpeechCreatedEvent struct {
	UserInitiated bool
	Source        string // "say" or "generate_reply"
	SpeechHandle  *SpeechHandle
	CreatedAt     time.Time
}

type SpeechHandle struct {
	ID                 string
	AllowInterruptions bool
	InputDetails       InputDetails
	Priority           int
	CreatedAt          time.Time
	// contains filtered or unexported fields
}

type TTSGenerationData struct {
	AudioCh chan *model.AudioFrame
	TTFB    time.Duration
}

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

type TaskWaiter interface {
	WaitAny(ctx context.Context) (any, error)
}

type ToolExecutionOutput struct {
	FncCall    llm.FunctionCall
	FncCallOut *llm.FunctionCallOutput
	RawOutput  any
	RawError   error
}

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

type TranscriptionFilter struct {
	SpeakingRate float64
}

type TurnDetectionMode string

type TurnDetector interface {
	PredictEndOfTurn(ctx context.Context, chatCtx *llm.ChatContext) (float64, error)
}

type UserInputTranscribedEvent struct {
	Language   string
	Transcript string
	IsFinal    bool
	SpeakerID  string
	CreatedAt  time.Time
}

type UserState string

type UserStateChangedEvent struct {
	OldState UserState
	NewState UserState
}

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