Your approach is fine, but maybe easier is to use ImageView's canvas grid, and
then you can use write_to_png(cg, filename).
--Tim
On Tuesday, September 09, 2014 07:16:04 PM Yuri Vishnevsky wrote:
> Hi all,
>
> I'm writing code to take a bunch of images and concatenate them into one
> big image (you can see the attached picture for an example of what this can
> look like).
>
> The code I wrote to generate the above picture is fairly hideous; I arrived
> at it after trying a number of approaches that seemed more natural and then
> giving up in despair.
>
> The code I initially wanted to write looked something like this:
>
>
> grid = Char[
> 'A' 'B' 'C' 'D' 'E';
> 'F' 'G' 'H' 'I' 'J';
> 'K' 'L' 'M' 'N' 'O';
> 'P' 'Q' 'R' 'S' 'T';
> 'U' 'V' 'W' 'X' 'Y';
> 'Z' 0 0 0 0 ;
> ]
>
> arr = map(c -> c == 0 ? zeros(Ufixed8, 64, 64) : convert(Array,
> imread("$prefix/sdf/$c.png")), grid)
> imwrite(arr, "$prefix/atlas.png")
>
> This approach fails in imwrite because the individual elements in the array
> are themselves arrays, rather than numbers. I couldn't figure out a way to
> flatten the resulting array-of-arrays into a single matrix and ended up
> writing code that preallocates a big matrix of the right output size and
> assigns to calculated subranges in a loop.
>
> My background is in user interface design and web programming, and right
> now I'm fairly new to writing matrix-manipulation code. If there's a
> natural way to express this computation in Julia I'd love to know.
>
> Cheers,
> Yuri
