R CMD check fixes

Author: multimeric

R/unharmonised.R

@@ -4,9 +4,13 @@
 #' make sense for these to live in the main metadata table. This function is a
 #' utility that allows easy fetching of this data if necessary.
 #'
-#' @param dataset_ids A character vector, where each entry is a dataset ID
+#' @param dataset_id A character vector, where each entry is a dataset ID
 #'   obtained from the `$file_id` column of the table returned from
 #'   [get_metadata()]
+#' @param cells An optional character vector of cell IDs. If provided, only
+#'   metadata for those cells will be returned.
+#' @param conn An optional DuckDB connection object. If provided, it will re-use
+#'   the existing connection instead of opening a new one.
 #' @param remote_url Optional character vector of length 1. An HTTP URL pointing
 #'   to the root URL under which all the unharmonised dataset files are located.
 #' @param cache_directory Optional character vector of length 1. A file path on
@@ -17,12 +21,13 @@
 #' @importFrom DBI dbConnect
 #' @importFrom duckdb duckdb
 #' @importFrom dplyr tbl filter
+#' @importFrom rlang .data
 #' @return A named list, where each name is a dataset file ID, and each value is
 #'   a "lazy data frame", ie a `tbl`.
 #' @examples
 #' dataset = "838ea006-2369-4e2c-b426-b2a744a2b02b"
 #' harmonised_meta = get_metadata() |> dplyr::filter(file_id == dataset) |> dplyr::collect()
-#' unharmonised_meta = get_unharmonised_metadata_list(dataset)
+#' unharmonised_meta = get_unharmonised_dataset(dataset)
 #' unharmonised_tbl = dplyr::collect(unharmonised_meta[[dataset]])
 #' dplyr::left_join(harmonised_meta, unharmonised_tbl, by=c("file_id", "cell_"))
 get_unharmonised_dataset = function(
@@ -41,7 +46,7 @@ get_unharmonised_dataset = function(
             progress(type = "down", con = stderr())
         )
     tbl(conn, local_path) |>
-        filter(cell_ %in% cells)
+        filter(.data$cell_ %in% cells)
 }
 
 #' Returns unharmonised metadata for a metadata query
@@ -54,16 +59,17 @@ get_unharmonised_dataset = function(
 #'  * `unharmonised`: a nested tibble, with one row per cell in the input `metadata`, containing unharmonised metadata
 #' @export
 #' @importFrom dplyr group_by summarise filter collect
+#' @importFrom rlang .data
 #' @examples
 #' harmonised <- get_metadata() |> dplyr::filter(tissue == "kidney blood vessel")
 #' unharmonised <- get_unharmonised_metadata(harmonised)
 get_unharmonised_metadata = function(metadata, ...){
     args = list(...)
     metadata |>
         collect() |>
-        group_by(file_id) |>
+        group_by(.data$file_id) |>
         summarise(
-            unharmonised = list(dataset_id=file_id[[1]], cells=cell_, conn=dbplyr::remote_con(metadata)) |>
+            unharmonised = list(dataset_id=.data$file_id[[1]], cells=.data$cell_, conn=dbplyr::remote_con(metadata)) |>
                 c(args) |> 
                 do.call(get_unharmonised_dataset, args=_) |> 
                 list()

README.Rmd

@@ -279,26 +279,22 @@ knitr::include_graphics("man/figures/HLA_A_tissue_plot.png")
 
 Various metadata fields are *not* common between datasets, so it does not
 make sense for these to live in the main metadata table. However, we can
-obtain it using the `get_unharmonised_metadata()` function.
-
-Note how this table has additional columns that are not in the normal metadata:
+obtain it using the `get_unharmonised_metadata()` function. This function
+returns a data frame with one row per dataset, including the `unharmonised` 
+column which contains unharmnised metadata as a nested data frame.
 
 ```{r}
-dataset = "838ea006-2369-4e2c-b426-b2a744a2b02b"
-unharmonised_meta = get_unharmonised_metadata(dataset)
-unharmonised_tbl = dplyr::collect(unharmonised_meta[[dataset]])
-unharmonised_tbl
+harmonised <- get_metadata() |> dplyr::filter(tissue == "kidney blood vessel")
+unharmonised <- get_unharmonised_metadata(harmonised)
+unharmonised
 ```
 
-If we have metadata from the normal metadata table that is from a single dataset,
-we can even join this additional metadata into one big data frame:
+Notice that the columns differ between each dataset's data frame:
+
 ```{r}
-harmonised_meta = get_metadata() |> dplyr::filter(file_id == dataset) |> dplyr::collect()
-dplyr::left_join(harmonised_meta, unharmonised_tbl, by=c("file_id", "cell_"))
+dplyr::pull(unharmonised, unharmonised) |> head(2)
 ```
 
-
-
 # Cell metadata
 
 Dataset-specific columns (definitions available at cellxgene.cziscience.com)

man/get_unharmonised_dataset.Rd

@@ -157,18 +157,29 @@ test_that("get_SingleCellExperiment() assigns the right cell ID to each cell", {
     )
 })
 
-test_that("get_unharmonised_metadata works with one ID", {
+test_that("get_unharmonised_dataset works with one ID", {
     dataset_id = "838ea006-2369-4e2c-b426-b2a744a2b02b"
-    unharmonised_meta = get_unharmonised_metadata(dataset_id)
-    unharmonised_tbl = unharmonised_meta[[dataset_id]]
-    
-    expect_type(unharmonised_meta, "list")
-    expect_s3_class(unharmonised_tbl, "tbl")
+    unharmonised_meta = get_unharmonised_dataset(dataset_id)
+
+    expect_s3_class(unharmonised_meta, "tbl")
 })
 
-test_that("get_unharmonised_metadata works with multiple IDs", {
-    dataset_ids = c("838ea006-2369-4e2c-b426-b2a744a2b02b", "83b9cb97-9ee4-404d-8cdf-ccede8235356")
-    unharmonised_meta = get_unharmonised_metadata(dataset_ids)
+test_that("get_unharmonised_metadata() returns the appropriate data", {
+    harmonised <- get_metadata() |> dplyr::filter(tissue == "kidney blood vessel")
+    unharmonised <- get_unharmonised_metadata(harmonised)
 
-    expect_equal(names(unharmonised_meta), dataset_ids)
-})
+    unharmonised |> is.data.frame() |> expect_true()
+    expect_setequal(colnames(unharmonised), c("file_id", "unharmonised"))
+    
+    # The number of cells in both harmonised and unharmonised should be the same
+    expect_equal(
+        dplyr::collect(harmonised) |> nrow(),
+        unharmonised$unharmonised |> purrr::map_int(function(df) dplyr::tally(df) |> dplyr::pull(n)) |> sum()
+    )
+    
+    # The number of datasets in both harmonised and unharmonised should be the same
+    expect_equal(
+        harmonised |> dplyr::group_by(file_id) |> dplyr::n_groups(),
+        nrow(unharmonised)
+    )
+})

man/get_unharmonised_metadata.Rd

@@ -297,22 +297,20 @@ knitr::include_graphics("../man/figures/HLA_A_tissue_plot.png")
 
 Various metadata fields are *not* common between datasets, so it does not
 make sense for these to live in the main metadata table. However, we can
-obtain it using the `get_unharmonised_metadata()` function.
-
-Note how this table has additional columns that are not in the normal metadata:
+obtain it using the `get_unharmonised_metadata()` function. This function
+returns a data frame with one row per dataset, including the `unharmonised` 
+column which contains unharmnised metadata as a nested data frame.
 
 ```{r}
-dataset = "838ea006-2369-4e2c-b426-b2a744a2b02b"
-unharmonised_meta = get_unharmonised_metadata(dataset)
-unharmonised_tbl = dplyr::collect(unharmonised_meta[[dataset]])
-unharmonised_tbl
+harmonised <- get_metadata() |> dplyr::filter(tissue == "kidney blood vessel")
+unharmonised <- get_unharmonised_metadata(harmonised)
+unharmonised
 ```
 
-If we have metadata from the normal metadata table that is from a single dataset,
-we can even join this additional metadata into one big data frame:
+Notice that the columns differ between each dataset's data frame:
+
 ```{r}
-harmonised_meta = get_metadata() |> dplyr::filter(file_id == dataset) |> dplyr::collect()
-dplyr::left_join(harmonised_meta, unharmonised_tbl, by=c("file_id", "cell_"))
+dplyr::pull(unharmonised, unharmonised) |> head(2)
 ```
 
 # Cell metadata