Demo metadata, build vignettes on CI

Conflicts:
	README.Rmd
	README.md
	vignettes/Introduction.Rmd

Author: multimeric

.github/workflows/check-bioc.yml

@@ -219,8 +219,8 @@ jobs:
           _R_CHECK_CRAN_INCOMING_: false
         run: |
           rcmdcheck::rcmdcheck(
-              args = c("--no-build-vignettes", "--no-manual", "--timings"),
-              build_args = c("--no-manual", "--no-resave-data"),
+              args = c(""--no-manual", "--timings"),
+              build_args = c("--no-resave-data"),
               error_on = "warning",
               check_dir = "check"
           )

DESCRIPTION

@@ -105,7 +105,8 @@ Suggests:
     spelling,
     forcats,
     ggplot2,
-    tidySingleCellExperiment
+    tidySingleCellExperiment,
+    rprojroot
 Biarch: true
 biocViews: 
     AssayDomain,

NAMESPACE

@@ -1,6 +1,8 @@
 # Generated by roxygen2: do not edit by hand
 
 S3method(as.sparse,DelayedMatrix)
+export(DATABASE_URL)
+export(SAMPLE_DATABASE_URL)
 export(get_SingleCellExperiment)
 export(get_metadata)
 export(get_seurat)

R/counts.R

@@ -13,14 +13,14 @@ assay_map <- c(
     cpm = "cpm"
 )
 
-#' Base URL pointing to the count data
+#' Base URL pointing to the count data at the current version
 COUNTS_URL <- single_line_str(
-    "https://swift.rc.nectar.org.au/v1/
-    AUTH_06d6e008e3e642da99d806ba3ea629c5/harmonised-human-atlas"
+    "https://object-store.rc.nectar.org.au/v1/
+    AUTH_06d6e008e3e642da99d806ba3ea629c5/cellxgene-0.2.1-hdf5"
 )
 #' Current version of the counts. This will be incremented when a newer
 #' version is released
-COUNTS_VERSION <- "0.2"
+COUNTS_VERSION <- "0.2.1"
 
 #' @inherit get_single_cell_experiment
 #' @inheritDotParams get_single_cell_experiment
@@ -199,10 +199,10 @@ group_to_sce <- function(i, df, dir_prefix, features) {
 
     file.exists(sce_path) |>
         assert_that(
-            msg = "Your cache does not contain a file you
+            msg = "Your cache does not contain a file {sce_path} you
                             attempted to query. Please provide the repository
                             parameter so that files can be synchronised from the
-                            internet"
+                            internet" |> glue()
         )
 
     sce <- loadHDF5SummarizedExperiment(sce_path)

R/dev.R

@@ -184,3 +184,64 @@ dir_to_anndata <- function(src, dest){
             }, .progress = "Converting files")
     }, env = zellkonverter::zellkonverterAnnDataEnv())
 }
+
+#' Makes a "downsampled" metadata file that only contains the minimal data
+#' needed to run the vignette.
+#' @param output Character scalar. Path to the output file.
+#' @return NULL
+#' @keywords internal
+downsample_metadata <- function(output = "sample_meta.parquet"){
+    metadata <- get_metadata()
+    
+    # Make a table of rows per dataset
+    dataset_sizes <- metadata |>
+        dplyr::group_by(.data$file_id_db) |>
+        summarise(n = dplyr::n()) |> 
+        dplyr::collect()
+    
+    # For each of the 3 examples, we select the minimal file_id_db that will
+    # satisfy the corresponding filters
+    example_a_all <- metadata |> 
+        dplyr::filter(
+            .data$ethnicity == "African" &
+            stringr::str_like(.data$assay, "%10x%") &
+            .data$tissue == "lung parenchyma" &
+            stringr::str_like(.data$cell_type, "%CD4%")
+        ) |>
+        dplyr::pull(.data$file_id_db)
+    example_a_minimal <- dataset_sizes |>
+        dplyr::filter(.data$file_id_db %in% example_a_all) |>
+        dplyr::slice_head(n=5) |>
+        dplyr::pull(.data$file_id_db)
+    
+    example_b_all <- metadata |> 
+        dplyr::filter(.data$cell_type_harmonised == "cd14 mono") |>
+        dplyr::pull(.data$file_id_db)
+    example_b_minimal <- dataset_sizes |>
+        dplyr::filter(.data$file_id_db %in% example_b_all) |>
+        dplyr::slice_head(n=1) |>
+        dplyr::pull(.data$file_id_db)
+    
+    example_c_all <- metadata |> 
+        dplyr::filter(.data$cell_type_harmonised == "nk") |>
+        dplyr::pull(.data$file_id_db)
+    example_c_minimal <- dataset_sizes |>
+        dplyr::filter(.data$file_id_db %in% example_c_all) |>
+        dplyr::slice_head(n=1) |>
+        dplyr::pull(.data$file_id_db)
+    
+    # The final dataset is the union of all the selected file IDs
+    minimal_file_ids <- union(
+        example_a_minimal,
+        example_b_minimal
+    ) |>
+        union(example_c_minimal)
+
+    metadata |>
+        dplyr::filter(.data$file_id_db %in% minimal_file_ids) |>
+        dplyr::arrange(.data$file_id_db, .data$sample_) |>
+        dplyr::collect() |>
+        arrow::write_parquet("sample_meta.parquet")
+    
+    NULL
+}

R/metadata.R

@@ -8,11 +8,22 @@ cache <- rlang::env(
     metadata_table = rlang::env()
 )
 
+#' URL pointing to the full metadata file
+#' @export
 DATABASE_URL <- single_line_str(
     "https://object-store.rc.nectar.org.au/v1/
     AUTH_06d6e008e3e642da99d806ba3ea629c5/metadata/metadata.0.2.3.parquet"
 )
 
+#' URL pointing to the sample metadata file, which is smaller and for test,
+#' demonstration, and vignette purposes only
+#' @export
+SAMPLE_DATABASE_URL <- single_line_str(
+    "https://object-store.rc.nectar.org.au/v1/
+    AUTH_06d6e008e3e642da99d806ba3ea629c5/metadata/
+    sample_metadata.0.2.3.parquet"
+)
+
 #' Gets the Curated Atlas metadata as a data frame.
 #'
 #' Downloads a parquet database of the Human Cell Atlas metadata to a local
@@ -24,7 +35,7 @@ DATABASE_URL <- single_line_str(
 #'   to the location of the parquet database.
 #' @param cache_directory Optional character vector of length 1. A file path on
 #'   your local system to a directory (not a file) that will be used to store
-#'   metadata.parquet
+#'   `metadata.parquet`
 #' @param use_cache Optional logical scalar. If `TRUE` (the default), and this
 #'   function has been called before with the same parameters, then a cached
 #'   reference to the table will be returned. If `FALSE`, a new connection will