Merge branch 'main' into issue-60-from-seurat

scverse · Sep 19, 2023 · f4784f3 · f4784f3
2 parents d713e7f + bffaf33
commit f4784f3
Show file tree

Hide file tree

Showing 49 changed files with 542 additions and 907 deletions.
diff --git a/.Rbuildignore b/.Rbuildignore
@@ -1,5 +1,4 @@
 ^LICENSE\.md$
-^doc$
 ^.*\.Rproj$
 ^\.Rproj\.user$
 ^\.github$
@@ -10,4 +9,6 @@
 ^_pkgdown\.yml$
 ^docs$
 ^pkgdown$
-^vignettes/features.Rmd$
+^vignettes/diagrams/*\.mmd$
+^vignettes/diagrams/*\.svg$
+^vignettes/design\.Rmd$
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -2,16 +2,41 @@ Package: anndataR
 Title: AnnData interoperability in R
 Version: 0.0.0.9000
 Authors@R: c(
-    person("Robrecht", "Cannoodt", , "[email protected]", role = c("aut", "cre"),
-           comment = c(ORCID = "0000-0003-3641-729X", github = "rcannood")),
-    person("Luke", "Zappia", , "[email protected]", role = "aut",
-           comment = c(ORCID = "0000-0001-7744-8565", github = "lazappi")),
-    person("Martin", "Morgan", , "[email protected]", role = "aut",
-           comment = c(ORCID = "0000-0002-5874-8148", github = "mtmorgan")),
-    person("Louise", "Deconinck", , "[email protected]", role = "aut",
-           comment = c(ORCID = "0000-0001-8100-6823", github = "LouiseDck")),
-    person("Danila", "Bredikhin", , "[email protected]", role = "aut",
-           comment = c(ORCID = "0000-0001-8089-6983", github = "gtca"))
+    person(
+        "Robrecht",
+        "Cannoodt", 
+        email = "[email protected]",
+        role = c("aut", "cre"),
+        comment = c(ORCID = "0000-0003-3641-729X", github = "rcannood")
+    ),
+    person(
+        "Luke",
+        "Zappia",
+        email = "[email protected]",
+        role = "aut",
+        comment = c(ORCID = "0000-0001-7744-8565", github = "lazappi")
+    ),
+    person(
+        "Martin",
+        "Morgan",
+        email = "[email protected]",
+        role = "aut",
+        comment = c(ORCID = "0000-0002-5874-8148", github = "mtmorgan")
+    ),
+    person(
+        "Louise",
+        "Deconinck",
+        email = "[email protected]",
+        role = "aut",
+        comment = c(ORCID = "0000-0001-8100-6823", github = "LouiseDck")
+    ),
+    person(
+        "Danila",
+        "Bredikhin",
+        email = "[email protected]",
+        role = "ctb",
+        comment = c(ORCID = "0000-0001-8089-6983", github = "gtca")
+    )
   )
 Description: Bring the power and flexibility of AnnData to the R
     ecosystem, allowing you to effortlessly manipulate and analyze your
@@ -29,6 +54,7 @@ Imports:
     R6
 Suggests:
     anndata,
+    BiocStyle,
     knitr,
     reticulate,
     rhdf5,

diff --git a/NAMESPACE b/NAMESPACE
@@ -2,7 +2,6 @@
 
 export(AnnData)
 export(InMemoryAnnData)
-export(dummy_data)
 export(from_Seurat)
 export(from_SingleCellExperiment)
 export(read_h5ad)

diff --git a/R/AbstractAnnData.R b/R/AbstractAnnData.R
@@ -57,7 +57,7 @@ AbstractAnnData <- R6::R6Class("AbstractAnnData", # nolint
     #'   computationally expensive.
     #' @param ... Optional arguments to print method.
     print = function(...) {
-      cat("AnnData object with n_obs \u00D7 n_vars = ", self$n_obs(), " × ", self$n_vars(), "\n", sep = "")
+      cat("AnnData object with n_obs \u00D7 n_vars = ", self$n_obs(), " \u00D7 ", self$n_vars(), "\n", sep = "")
 
       for (attribute in c(
         "obs",

diff --git a/R/AnnData.R b/R/AnnData.R
@@ -46,13 +46,12 @@
 #'
 #' adata
 AnnData <- function(
-  obs_names = NULL,
-  var_names = NULL,
-  X = NULL,
-  obs = NULL,
-  var = NULL,
-  layers = NULL
-) {
+    obs_names = NULL,
+    var_names = NULL,
+    X = NULL,
+    obs = NULL,
+    var = NULL,
+    layers = NULL) {
   InMemoryAnnData$new(
     obs_names = obs_names,
     var_names = var_names,

diff --git a/R/HDF5AnnData.R b/R/HDF5AnnData.R
@@ -15,10 +15,10 @@ HDF5AnnData <- R6::R6Class("HDF5AnnData", # nolint
     #' @field X The X slot
     X = function(value) {
       if (missing(value)) {
-        # trackstatus: class=HDF5AnnData, feature=get_X, status=wip
+        # trackstatus: class=HDF5AnnData, feature=get_X, status=done
         read_h5ad_element(private$.h5obj, "/X")
       } else {
-        # trackstatus: class=HDF5AnnData, feature=set_X, status=wip
+        # trackstatus: class=HDF5AnnData, feature=set_X, status=done
         value <- private$.validate_matrix(value, "X")
         write_h5ad_element(value, private$.h5obj, "/X")
       }
@@ -28,21 +28,21 @@ HDF5AnnData <- R6::R6Class("HDF5AnnData", # nolint
     #'   `obs` and `var`.
     layers = function(value) {
       if (missing(value)) {
-        # trackstatus: class=HDF5AnnData, feature=get_layers, status=wip
+        # trackstatus: class=HDF5AnnData, feature=get_layers, status=done
         read_h5ad_element(private$.h5obj, "layers")
       } else {
-        # trackstatus: class=HDF5AnnData, feature=set_layers, status=wip
+        # trackstatus: class=HDF5AnnData, feature=set_layers, status=done
         value <- private$.validate_layers(value)
         write_h5ad_element(value, private$.h5obj, "/layers")
       }
     },
     #' @field obs The obs slot
     obs = function(value) {
       if (missing(value)) {
-        # trackstatus: class=HDF5AnnData, feature=get_obs, status=wip
+        # trackstatus: class=HDF5AnnData, feature=get_obs, status=done
         read_h5ad_element(private$.h5obj, "/obs", include_index = FALSE)
       } else {
-        # trackstatus: class=HDF5AnnData, feature=set_obs, status=wip
+        # trackstatus: class=HDF5AnnData, feature=set_obs, status=done
         value <- private$.validate_obsvar_dataframe(value, "obs")
         write_h5ad_element(
           value,
@@ -55,10 +55,10 @@ HDF5AnnData <- R6::R6Class("HDF5AnnData", # nolint
     #' @field var The var slot
     var = function(value) {
       if (missing(value)) {
-        # trackstatus: class=HDF5AnnData, feature=get_var, status=wip
+        # trackstatus: class=HDF5AnnData, feature=get_var, status=done
         read_h5ad_element(private$.h5obj, "/var", include_index = FALSE)
       } else {
-        # trackstatus: class=HDF5AnnData, feature=set_var, status=wip
+        # trackstatus: class=HDF5AnnData, feature=set_var, status=done
         value <- private$.validate_obsvar_dataframe(value, "var")
         write_h5ad_element(
           value,
@@ -70,17 +70,16 @@ HDF5AnnData <- R6::R6Class("HDF5AnnData", # nolint
     },
     #' @field obs_names Names of observations
     obs_names = function(value) {
-      # TODO: directly write to and read from /obs/_index
       if (missing(value)) {
-        # trackstatus: class=HDF5AnnData, feature=get_obs_names, status=wip
+        # trackstatus: class=HDF5AnnData, feature=get_obs_names, status=done
         # obs names are cached to avoid reading all of obs whenever they are
         # accessed
         if (is.null(private$.obs_names)) {
           private$.obs_names <- read_h5ad_data_frame_index(private$.h5obj, "obs")
         }
         private$.obs_names
       } else {
-        # trackstatus: class=HDF5AnnData, feature=set_obs_names, status=wip
+        # trackstatus: class=HDF5AnnData, feature=set_obs_names, status=done
         value <- private$.validate_obsvar_names(value, "obs")
         write_h5ad_data_frame_index(value, private$.h5obj, "obs", "_index")
         private$.obs_names <- value
@@ -90,15 +89,15 @@ HDF5AnnData <- R6::R6Class("HDF5AnnData", # nolint
     var_names = function(value) {
       # TODO: directly write to and read from /var/_index
       if (missing(value)) {
-        # trackstatus: class=HDF5AnnData, feature=get_var_names, status=wip
+        # trackstatus: class=HDF5AnnData, feature=get_var_names, status=done
         # var names are cached to avoid reading all of var whenever they are
         # accessed
         if (is.null(private$.var_names)) {
           private$.var_names <- read_h5ad_data_frame_index(private$.h5obj, "var")
         }
         private$.var_names
       } else {
-        # trackstatus: class=HDF5AnnData, feature=set_var_names, status=wip
+        # trackstatus: class=HDF5AnnData, feature=set_var_names, status=done
         value <- private$.validate_obsvar_names(value, "var")
         write_h5ad_data_frame_index(value, private$.h5obj, "var", "_index")
         private$.var_names <- value

diff --git a/R/read_h5ad.R b/R/read_h5ad.R
@@ -0,0 +1,33 @@
+#' Read H5AD
+#'
+#' Read data from a H5AD file
+#'
+#' @param path Path to the H5AD file to read
+#' @param to The type of object to return. Must be one of: "SingleCellExperiment",
+#'   "Seurat", "HDF5AnnData", "InMemoryAnnData"
+#'
+#' @return The object specified by `to`
+#' @export
+#'
+#' @examples
+#' h5ad_file <- system.file("extdata", "example.h5ad", package = "anndataR")
+#' # Read the H5AD as a SingleCellExperiment object
+#' if (requireNamespace("SingleCellExperiment", quietly = TRUE)) {
+#'   sce <- read_h5ad(h5ad_file, to = "SingleCellExperiment")
+#' }
+#' # Read the H5AD as a Seurat object
+#' if (requireNamespace("SeuratObject", quietly = TRUE)) {
+#'   seurat <- read_h5ad(h5ad_file, to = "Seurat")
+#' }
+read_h5ad <- function(path, to = c("SingleCellExperiment", "Seurat", "HDF5AnnData", "InMemoryAnnData")) {
+  to <- match.arg(to)
+
+  adata <- HDF5AnnData$new(path)
+
+  switch(to,
+    "SingleCellExperiment" = to_SingleCellExperiment(adata),
+    "Seurat" = to_Seurat(adata),
+    "HDF5AnnData" = adata,
+    "InMemoryAnnData" = to_InMemoryAnnData(adata)
+  )
+}