Skip to content

Add TiledDiskArray #283

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
wants to merge 2 commits into
base: main
Choose a base branch
from
Draft

Add TiledDiskArray #283

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
69725cd
Move AbstractChunkTiledArray into separate file
felixcremer Oct 24, 2025
ced4c47
Initial implementation of TiledDiskArray
felixcremer Oct 24, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions src/DiskArrays.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,6 +34,7 @@ include("cat.jl")
include("generator.jl")
include("zip.jl")
include("show.jl")
include("chunktiledarray.jl")
include("cached.jl")
include("pad.jl")

Expand Down
31 changes: 2 additions & 29 deletions src/cached.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,34 +1,7 @@
import Mmap

"""
ChunkTiledDiskArray <: AbstractDiskArray

And abstract supertype for disk arrays that have fast indexing
of tiled chunks already stored as separate arrays, such as [`CachedDiskArray`](@ref).
"""
abstract type ChunkTiledDiskArray{T,N} <: AbstractDiskArray{T,N} end

Base.size(a::ChunkTiledDiskArray) = arraysize_from_chunksize.(eachchunk(a).chunks)

function readblock!(A::ChunkTiledDiskArray{T,N}, data, I...) where {T,N}
chunks = eachchunk(A)
chunk_indices = findchunk.(chunks.chunks, I)
data_offset = OffsetArray(data, map(i -> first(i) - 1, I)...)
foreach(CartesianIndices(chunk_indices)) do ci
chunkindex = ChunkIndex(ci; offset=true)
chunk = A[chunkindex]
# Find the overlapping indices
inner_indices = map(axes(chunk), axes(data_offset)) do ax1, ax2
max(first(ax1), first(ax2)):min(last(ax1), last(ax2))
end
for ii in CartesianIndices(inner_indices)
data_offset[ii] = chunk[ii]
end
end
end

"""
CachedDiskArray <: ChunkTiledDiskArray
CachedDiskArray <: AbstractChunkTiledDiskArray

CachedDiskArray(A::AbstractArray; maxsize=1000, mmap=false)

Expand All @@ -40,7 +13,7 @@ to temproray files.

Can also be called with `cache`, which can be extended for wrapper array types.
"""
struct CachedDiskArray{T,N,A<:AbstractArray{T,N},C} <: ChunkTiledDiskArray{T,N}
struct CachedDiskArray{T,N,A<:AbstractArray{T,N},C} <: AbstractChunkTiledDiskArray{T,N}
parent::A
cache::C
mmap::Bool
Expand Down
60 changes: 60 additions & 0 deletions src/chunktiledarray.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,60 @@
"""
AbstractChunkTiledDiskArray <: AbstractDiskArray

An abstract supertype for disk arrays that have fast indexing
of tiled chunks already stored as separate arrays, such as [`CachedDiskArray`](@ref).
"""
abstract type AbstractChunkTiledDiskArray{T,N} <: AbstractDiskArray{T,N} end

Base.size(a::AbstractChunkTiledDiskArray) = arraysize_from_chunksize.(eachchunk(a).chunks)

function readblock!(A::AbstractChunkTiledDiskArray{T,N}, data, I...) where {T,N}
chunks = eachchunk(A)
chunk_indices = findchunk.(chunks.chunks, I)
data_offset = OffsetArray(data, map(i -> first(i) - 1, I)...)
foreach(CartesianIndices(chunk_indices)) do ci
@show ci
Copy link
Collaborator

@meggart meggart Oct 28, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

remove

chunkindex = ChunkIndex(ci; offset=true)
@show chunkindex
Copy link
Collaborator

@meggart meggart Oct 28, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

remove

chunk = A[chunkindex]
# Find the overlapping indices
inner_indices = map(axes(chunk), axes(data_offset)) do ax1, ax2
max(first(ax1), first(ax2)):min(last(ax1), last(ax2))
end
for ii in CartesianIndices(inner_indices)
data_offset[ii] = chunk[ii]
end
end
end

"""
TiledDiskArray <: AbstractChunkTiledDiskArray

Construct an array from a collection of tiles.
This needs a function to find the tile given a tile position and the overall size of the array.
"""
struct TiledDiskArray{T,N} <: AbstractChunkTiledDiskArray{T,N}
tilefunction
tilenum
tilesizes
Comment on lines +37 to +39
Copy link
Collaborator

@meggart meggart Oct 28, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you use this definition instead:

struct TiledDiskArray{T,N,F} <: AbstractChunkTiledDiskArray{T,N}
    tilefunction::F
    tilenum::NTuple{N,Int}
    tilesizes::NTuple{N,Int}
end

Copy link
Member

@asinghvi17 asinghvi17 Oct 28, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Might be worth having the tile definition be GridChunks instead - for example the Copernicus DEM has irregular sizes

Copy link
Contributor Author

@felixcremer felixcremer Oct 28, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do you have a link to that?

Copy link
Member

@asinghvi17 asinghvi17 Oct 28, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

https://copernicus-dem-30m.s3.amazonaws.com/readme.html

Copy link
Collaborator

@rafaqz rafaqz Oct 29, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is tilesizes just tilesize? What does the pluralisation mean?

end


Base.size(A::TiledDiskArray) = A.tilenum .* A.tilesizes
eachchunk(A::TiledDiskArray) = DiskArrays.GridChunks(size(A), A.tilesizes)
haschunks(A::TiledDiskArray) = Chunked()

function Base.getindex(A::TiledDiskArray, i::ChunkIndex{N,OffsetChunks}) where {N}
tile = _getchunk(A,i)
inds = eachchunk(A)[i.I]
wrapchunk(tile, inds)
end

Base.getindex(A::TiledDiskArray, i::ChunkIndex{N,OneBasedChunks}) where {N} =
_getchunk(A, i)



function _getchunk(A::TiledDiskArray, i::ChunkIndex)
A.tilefunction(i.I.I...)
end
6 changes: 6 additions & 0 deletions test/runtests.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1157,3 +1157,9 @@ end
@test DiskArrays.haschunks(a_chunked_2) isa DiskArrays.Chunked
@test size(DiskArrays.eachchunk(a_chunked_2)) == (5,5)
end

@testset "TiledChunkArray" begin
tiles = [fill(10x+y, 10,10) for x in 1:9, y in 1:9]
a = DiskArrays.TiledDiskArray((x,y) -> tiles[x,y], (90,90))
@test a[1,1] == 11
end
Loading