Figured out the problem with the fill colour alpha.
At first I thought I was on the wrong track. I thought maybe I wasn’t using the per-pixel transparency option, which is selected by setting the options mask to SRCALPHA. But when I looked at the code to initialize the surface I found:
alpha_surf = pygame.Surface(screen.get_size(), pygame.SRCALPHA)
It’s already using SRCALPHA. Womp.
But then I realized the second parameter in the constructor is flags not masks so I tried changing to the following line…
alpha_surf = pygame.Surface(screen.get_size(), masks=pygame.SRCALPHA)
…and it worked!
Wute.
This seems really inefficient, though.
I started off doing a fill because I was going to need to create a mask for reaction-diffusion. That’s no longer the case but calculating the area using a flood fill seemed like it would be easier than managing all the line segments.
Maybe it’s not as bad as I think, though.
All I need to do is travel down one primary vein collecting all the segment start points. Then collect the segment endpoint. Then jump to the next primary vein and add that endpoint. Then crawl up the vein adding points along the way.
With a sequence of points I can get the area from a library.
polygon = Polygon(points)
return polygon.area
Easy peasy.
For kicks (or rather for a quick sanity check) I printed both sets of numbers out and I can see that they are comparable.
32021 32498 32093 30792 28991 25827 21864 18999 16850 14984 14410
33811 34149 33606 32130 30181 26880 22892 19806 17558 15769 15176
Okay, so now I’m gonna try to do Poisson Disc sampling to get a bunch of points on the wing that are separated by a minimum distance.
Start with dart-throwing algorithm?