write/update to disk utilities

[Artur-man]

We can implement methods like create_zarr and create_zarr_group until they are available in Rarr
Huber-group-EMBL/Rarr#18

[Artur-man]

Slowly starting to introduce methods to write/update zarr stores, groups, atttibutes and SpatialElements.

Will work on this branch, and will keep a track of stuff below:

• create zarr stores
• create zarr groups
• write zattrs
• create in-memory metadata

image	label	shape	point	table
❌	❌	❌	✔	❌

• create in-memory spatialelement

image	label	shape	point	table
❌	❌	❌	✔	❌

• write metadata to json

image	label	shape	point	table
❌	❌	❌	✔	❌

• write spatialelement
  https://spatialdata.scverse.org/en/latest/api/SpatialData.html#spatialdata.SpatialData.write_element

image	label	shape	point	table
❌	❌	❌	✔	❌

• write SpatialData itself https://spatialdata.scverse.org/en/latest/api/SpatialData.html#spatialdata.SpatialData.write
• add to vignette and examples

[Artur-man]

@HelenaLC

Is checking for dask index necessary here ? is it failing for any of the Spatialdata.data examples without ?

SpatialData/R/PointFrame.R

Lines 108 to 123 in 1c14379

	#' @rdname PointFrame
	#' @importFrom dplyr mutate filter select
	#' @export
	setMethod("[", c("PointFrame", "numeric", "numeric"), \(x, i, j, ...) {
	.i <- `__null_dask_index__` <- NULL # R CMD check
	i <- seq_len(nrow(x))[i]
	x@data <- data(x) |>
	mutate(.i=1+`__null_dask_index__`) |>
	filter(.i %in% i) |>
	select(-.i)
	# make sure this is kept in any case
	ndi <- "__null_dask_index__"
	ndi <- match(ndi, names(x@data), nomatch=0)
	x@data <- x@data[, c(j, ndi)]
	return(x)
	})

Did also a quick check on indexing, the regular indexing seem also be doing stuff lazily.

> library(SpatialData.data)
> x <- SpatialData.data::MouseBrainMERFISH()
> object.size(x@points$single_molecule[1:100000,]@data)
488 bytes
> object.size(x@points$single_molecule@data[1:100000,])
488 bytes