Write support for anndata-zarr v3

.github/workflows/R-CMD-check-bioc.yaml

@@ -92,8 +92,8 @@ jobs:
           install.packages(c("BiocManager", "remotes"), quiet = TRUE)
           ## update=FALSE avoids calling old.packages(type="both") which
           ## triggers the R-devel bug in .available.both() (commit dd03406)
-          BiocManager::install(version = "3.23", ask = FALSE, update = FALSE)
-          cat("R_BIOC_VERSION=3.23\n", file = Sys.getenv("GITHUB_ENV"), append = TRUE)
+          BiocManager::install(version = "devel", ask = FALSE, update = FALSE)
+          cat("R_BIOC_VERSION=3.24\n", file = Sys.getenv("GITHUB_ENV"), append = TRUE)
 
       - name: Install pandoc
         uses: r-lib/actions/setup-pandoc@v2

DESCRIPTION

@@ -53,7 +53,7 @@ Suggests:
     knitr,
     processx,
     rhdf5 (>= 2.52.1),
-    Rarr (>= 1.11.12),
+    Rarr,
     rmarkdown,
     S4Vectors,
     Seurat,
@@ -72,5 +72,5 @@ Config/Needs/website: pkgdown, tibble, knitr, rprojroot, stringr, readr, purrr,
 Config/testthat/edition: 3
 Encoding: UTF-8
 Roxygen: list(markdown = TRUE, r6 = TRUE)
-RoxygenNote: 7.3.3
 Language: en-GB
+Config/roxygen2/version: 8.0.0

R/Rarr_utils.R

@@ -12,7 +12,7 @@ ZARR_METADATA_FILES <- c(".zarray", ".zattrs", ".zgroup", "zarr.json")
 #' @return `NULL`
 #'
 #' @noRd
-create_zarr_group <- function(store, name, version = "v2") {
+create_zarr_group <- function(store, name, version = .get_zarr_version()) {
   # Split "a/b/c" into c("a", "b", "c")
   split_name <- strsplit(name, split = "/", fixed = TRUE)[[1]]
   if (length(split_name) > 1) {
@@ -31,6 +31,7 @@ create_zarr_group <- function(store, name, version = "v2") {
     }
   }
   dir.create(file.path(store, split_name[1]), showWarnings = FALSE)
+  version <- paste0("v", version)
   switch(
     version,
     v2 = {
@@ -40,9 +41,12 @@ create_zarr_group <- function(store, name, version = "v2") {
       )
     },
     v3 = {
-      cli_abort("Currently only zarr v2 is supported!")
+      write(
+        "{\"zarr_format\": 3,\n\"node_type\": \"group\"}",
+        file = file.path(store, split_name[1], "zarr.json")
+      )
     },
-    cli_abort("Only zarr v2 is supported. Use version = 'v2'")
+    cli_abort("Incorrect Zarr version specified. Must be '2' or '3'.")
   )
 }
 
@@ -56,7 +60,7 @@ create_zarr_group <- function(store, name, version = "v2") {
 #' @return `NULL`
 #'
 #' @noRd
-create_zarr <- function(store, version = "v2") {
+create_zarr <- function(store, version = .get_zarr_version()) {
   prefix <- basename(store)
   dir <- gsub(paste0(prefix, "$"), "", store)
   create_zarr_group(store = dir, name = prefix, version = version)
@@ -103,3 +107,8 @@ zarr_path_exists <- function(store, target_path) {
     }
   }
 }
+
+.get_zarr_version <- function() {
+  vr <- getOption("anndataR.zarr_version")
+  if (is.null(vr)) 2 else vr
+}

R/ZarrAnnData.R

@@ -20,6 +20,7 @@
   cloneable = FALSE,
   private = list(
     .zarrobj = NULL,
+    .zarrversion = NULL,
     .compression = NULL,
     .readonly = NULL,
 
@@ -62,7 +63,8 @@
           write_zarr_element(
             private$.zarrobj,
             "X",
-            private$.compression
+            private$.compression,
+            private$.zarrversion
           )
       }
     },
@@ -87,7 +89,8 @@
           write_zarr_element(
             private$.zarrobj,
             "layers",
-            private$.compression
+            private$.compression,
+            private$.zarrversion
           )
       }
     },
@@ -114,7 +117,8 @@
           write_zarr_element(
             private$.zarrobj,
             "obsm",
-            private$.compression
+            private$.compression,
+            private$.zarrversion
           )
       }
     },
@@ -141,7 +145,8 @@
           write_zarr_element(
             private$.zarrobj,
             "varm",
-            private$.compression
+            private$.compression,
+            private$.zarrversion
           )
       }
     },
@@ -166,7 +171,8 @@
           write_zarr_element(
             private$.zarrobj,
             "obsp",
-            private$.compression
+            private$.compression,
+            private$.zarrversion
           )
       }
     },
@@ -191,7 +197,8 @@
           write_zarr_element(
             private$.zarrobj,
             "varp",
-            private$.compression
+            private$.compression,
+            private$.zarrversion
           )
       }
     },
@@ -209,7 +216,8 @@
           write_zarr_element(
             private$.zarrobj,
             "obs",
-            private$.compression
+            private$.compression,
+            private$.zarrversion
           )
       }
     },
@@ -227,7 +235,8 @@
           write_zarr_element(
             private$.zarrobj,
             "var",
-            private$.compression
+            private$.compression,
+            private$.zarrversion
           )
       }
     },
@@ -275,7 +284,8 @@
           write_zarr_element(
             private$.zarrobj,
             "uns",
-            private$.compression
+            private$.compression,
+            private$.zarrversion
           )
       }
     }
@@ -308,28 +318,29 @@
     #' create a new one. If any additional slot arguments are set an existing
     #' file will be overwritten.
     initialize = function(
-      file,
-      X = NULL,
-      obs = NULL,
-      var = NULL,
-      layers = NULL,
-      obsm = NULL,
-      varm = NULL,
-      obsp = NULL,
-      varp = NULL,
-      uns = NULL,
-      shape = NULL,
-      mode = c("a", "r", "r+", "w", "w-", "x"),
-      compression = c(
-        "none",
-        "gzip",
-        "blosc",
-        "zstd",
-        "lzma",
-        "bz2",
-        "zlib",
-        "lz4"
-      )
+    file,
+    X = NULL,
+    obs = NULL,
+    var = NULL,
+    layers = NULL,
+    obsm = NULL,
+    varm = NULL,
+    obsp = NULL,
+    varp = NULL,
+    uns = NULL,
+    shape = NULL,
+    mode = c("a", "r", "r+", "w", "w-", "x"),
+    compression = c(
+      "none",
+      "gzip",
+      "blosc",
+      "zstd",
+      "lzma",
+      "bz2",
+      "zlib",
+      "lz4"
+    ),
+    zarr_version = .get_zarr_version()
     ) {
       check_requires("ZarrAnnData", "Rarr", where = "Bioc")
 
@@ -338,6 +349,7 @@
 
       private$.compression <- compression
 
+      private$.zarrversion <- zarr_version
       is_readonly <- FALSE
 
       if (is.character(file)) {
@@ -520,7 +532,7 @@
 #'
 # nolint start: object_name_linter
 as_ZarrAnnData <- function(
-  # nolint end: object_name_linter
+    # nolint end: object_name_linter
   adata,
   file,
   compression = c(
@@ -555,6 +567,7 @@
     uns = adata$uns,
     shape = adata$shape(),
     mode = mode,
-    compression = compression
+    compression = compression,
+    zarr_version = .get_zarr_version()
   )
 }