multimodal

type AudioConfig struct {
	MaxDuration    float64       `json:"max_duration"`  // Max duration in seconds
	MaxFileSize    int64         `json:"max_file_size"` // Max file size in bytes
	SampleRate     int           `json:"sample_rate"`   // Target sample rate
	AllowedFormats []AudioFormat `json:"allowed_formats"`
}

type AudioFormat string

type Capability string

type Content struct {
	Type     ContentType `json:"type"`
	Text     string      `json:"text,omitempty"`
	ImageURL string      `json:"image_url,omitempty"`
	AudioURL string      `json:"audio_url,omitempty"`
	VideoURL string      `json:"video_url,omitempty"`

	// Base64 编码数据( URL 的选项)
	Data      string `json:"data,omitempty"`
	MediaType string `json:"media_type,omitempty"` // e.g., "image/png", "audio/mp3"

	// 元数据
	FileName   string            `json:"file_name,omitempty"`
	FileSize   int64             `json:"file_size,omitempty"`
	Duration   float64           `json:"duration,omitempty"` // For audio/video in seconds
	Dimensions *ImageDimensions  `json:"dimensions,omitempty"`
	Metadata   map[string]string `json:"metadata,omitempty"`
}

type ContentType string

type ImageDimensions struct {
	Width  int `json:"width"`
	Height int `json:"height"`
}

type ImageFormat string

type MultimodalMessage struct {
	Role     string    `json:"role"`
	Contents []Content `json:"contents"`
}

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

type MultimodalRequest struct {
	llm.ChatRequest
	MultimodalMessages []MultimodalMessage `json:"multimodal_messages,omitempty"`
}

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

type ResolutionPreset string

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

type VisionConfig struct {
	Resolution     ResolutionPreset `json:"resolution"`
	MaxImageSize   int64            `json:"max_image_size"` // Max file size in bytes
	MaxDimension   int              `json:"max_dimension"`  // Max width/height
	AllowedFormats []ImageFormat    `json:"allowed_formats"`
}