Struct VisionSignal 

pub struct VisionSignal {
    pub recent_image_count: usize,
    pub has_image_in_latest_turn: bool,
    pub image_producing_tools: Vec<String>,
}

Signal derived from image content in the conversation.

Detects image blocks in user messages and tool results (e.g. screenshots) to determine whether a vision-capable model is needed.

Fields

recent_image_count: usize

Number of image blocks found in the recent window.

has_image_in_latest_turn: bool

Whether the latest user turn contains an image.

image_producing_tools: Vec<String>

Tool names that produced images (e.g. “browse”, “browse_script”).

Implementations

impl VisionSignal

pub fn extract(messages: &[Message], window: usize) -> VisionSignal

Extract vision signal from the last window messages.

pub fn requires_vision(&self) -> bool

Whether any image content requires a vision-capable model.

pub fn normalized(&self) -> f64

Normalize to [0, 1] for scoring.

• 0 images → 0.0 (no effect)
• 1 image → 0.7
• 2+ images → 0.9 → approaching 1.0

Trait Implementations

impl Clone for VisionSignal

fn clone(&self) -> VisionSignal

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for VisionSignal

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.

impl Default for VisionSignal

fn default() -> VisionSignal

Returns the “default value” for a type.