Documentation
¶
Overview ¶
Package vlm provides image annotation (image -> scene summary + numbered element list) via a side-call to a vision model through the inference gateway. Gated by config.VisionConfig. The gateway also serves local models (e.g. Ollama), so offline annotation is a gateway concern.
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type GatewayAnnotator ¶
type GatewayAnnotator struct {
// contains filtered or unexported fields
}
GatewayAnnotator annotates images with a one-off side-call to a vision-capable provider/model routed through the inference gateway - the same mechanism conversation-title generation uses.
func NewGatewayAnnotator ¶
func NewGatewayAnnotator(client sdk.Client, cfg *config.Config) *GatewayAnnotator
NewGatewayAnnotator creates a gateway-backed annotator.
func (*GatewayAnnotator) AnnotateImage ¶
func (a *GatewayAnnotator) AnnotateImage(ctx context.Context, img agentdomain.ImageAttachment, opts agentdomain.AnnotateOptions) (*agentdomain.ImageAnnotation, error)
AnnotateImage sends the image to the configured vision model and parses the JSON reply.
Source Files
Click to show internal directories.
Click to hide internal directories.