Annual Daylight Simulation (MassesToZones component) struggles with "donut" shaped BREPS

Working from example file 06, from honeybee, the annual daylight analysis…

Everything works fine when i reference masses that are relatively simple extrusions of a boundary curve, such as the attached. And i get realistic annual results.

Strangely, when i insert a void in the center, making a courtyard, the simulation still runs, but the results are very low, and not realistic. My theory here is that 1) i am doing something wrong, or that 2) the “MassesToZones” component was not intended to parse “donut” shapes, and is having trouble.

What is troubling, is that the windows preview correctly, so the geometry looks right,

I was able to model “donut” shapes by modeling the space without using the MassesToZones component, but in this case it appears that i cannot take advantage of the glazing-by-ratio components features, which appear very useful.

Suggestions appreciated!

I have uploaded the rhino and gh files here

file too large, uploading individually…

I guess you are using old version of Honeybee. I tested an example with the new version and it just works fine (attached!). Update the components in the example file and try again.

Also you can use importRAD component to see how the geometry has been exported to Radiance. (164 KB)

Indeed, that worked…

I downloaded the latest version of HB from Github and the donut geometry now simulates perfectly.