Two (correlated) propositions :
-
The client.render() could take an optionnal mask describing the regions in the output image to render.
This could be of many forms, like a bitmask (of any size, to be scaled to the acutal rendering buffer size).
This could also be a list of 2D bounding box in screen coordinates or a list of circles (center + radius).
-
The client.getImage() could take a list of regions of interest to retrieve sub-image of the actual rendering
A nice and coherent API could be :
client.render(regions:List[Tuple[int,int,int,int]]=None)
client.getImage(regions:List[Tuple[int,int,int,int]]=None)
The client.render() proposition could be used to make faster computations.
The client.getImage() would maily be just an helper function, when combine with the client.render() proposition.
Two (correlated) propositions :
The client.render() could take an optionnal mask describing the regions in the output image to render.
This could be of many forms, like a bitmask (of any size, to be scaled to the acutal rendering buffer size).
This could also be a list of 2D bounding box in screen coordinates or a list of circles (center + radius).
The client.getImage() could take a list of regions of interest to retrieve sub-image of the actual rendering
A nice and coherent API could be :
client.render(regions:List[Tuple[int,int,int,int]]=None)
client.getImage(regions:List[Tuple[int,int,int,int]]=None)
The client.render() proposition could be used to make faster computations.
The client.getImage() would maily be just an helper function, when combine with the client.render() proposition.