xr_ai_tools.current_frame

Select current camera frames without invoking a VLM.

Classes

ImageFrame	One selected camera frame and its source metadata.
CurrentFrameRequest	Select the latest camera frame for one participant.
CurrentFrameTool	Return a participant's current frame without running visual inference.

Module Contents

class xr_ai_tools.current_frame.ImageFrame

One selected camera frame and its source metadata.

width: int

Frame width in pixels.

height: int

Frame height in pixels.

sequence: int

Source track sequence number.

participant_id: str

Participant that published the frame.

track_id: str = ''

Source video track identifier, when available.

class xr_ai_tools.current_frame.CurrentFrameRequest

Select the latest camera frame for one participant.

participant_id: str

Participant whose latest camera frame should be returned.

class xr_ai_tools.current_frame.CurrentFrameTool(

*,

endpoint: xr_ai_hub.ProcessorEndpoint,

images: xr_ai_tools.image.ImageRegistry,

frame_max_age_s: float = 2.0,

frame_timeout_s: float = 5.0,

)

Return a participant’s current frame without running visual inference.

release(participant_id: str) → None

Forget cached frame state and image handles after disconnect.