add some docs

docs/src/UserGuide/cache.md

@@ -0,0 +1,18 @@
+# Caching YAXArrays
+
+For some applications like interactive plotting of large datasets it can not be avoided that the same data must be accessed several times. In these cases it can be useful to store recently accessed data in a cache. In YAXArrays this can be easily achieved using the `cache` function. For example, if we open a large dataset from a remote source and want to keep data in a cache of size 500MB one can use:
+
+````julia
+using YAXArrays, Zarr
+ds = open_dataset("path/to/source")
+cachesize = 500 #MB
+cache(ds,maxsize = cachesize)
+````
+
+The above will wrap every array in the dataset into its own cache, where the 500MB are distributed equally across datasets. 
+Alternatively individual caches can be applied to single `YAXArray`s
+
+````julia
+yax = ds.avariable
+cache(yax,maxsize = 1000)
+````

src/Cubes/Cubes.jl

@@ -17,7 +17,7 @@ using Tables: istable, schema, columns
 using DimensionalData: DimensionalData as DD, AbstractDimArray, NoName
 import DimensionalData: name
 
-export concatenatecubes, caxes, subsetcube, readcubedata, renameaxis!, YAXArray, setchunks
+export concatenatecubes, caxes, subsetcube, readcubedata, renameaxis!, YAXArray, setchunks, cache
 
 """
 This function calculates a subset of a cube's data

test/Datasets/datasets.jl

@@ -411,6 +411,10 @@ end
 end
 
 @testset "Caching" begin
+    using YAXArrays.Cubes.DiskArrays.TestTypes
+    using YAXArrays.Cubes: DiskArrays
+    using Test
+    import DimensionalData as DD
     a = AccessCountDiskArray(reshape(1:100, 5, 20), chunksize=(2, 10))
     ds = Dataset(; ar=YAXArray((DD.X(1:5), DD.Y(1:20)), a))
     dscached = DiskArrays.cache(ds)