[CW-7210] Add a sample sheet reader

bin/workflow_glue_r/R/bambu.R

@@ -168,14 +168,22 @@ bambu_resolve_chunk_dirs <- function(args) {
     chunk_dirs
 }
 
+bambu_read_sample_sheet <- function(path) {
+    header <- names(utils::read.csv(path, nrows = 0, check.names = FALSE))
+    char_cols <- intersect(c("alias", "sample_id"), header)
+    col_classes <- stats::setNames(rep("character", length(char_cols)), char_cols)
+    utils::read.csv(
+        path,
+        check.names = FALSE,
+        stringsAsFactors = FALSE,
+        colClasses = col_classes
+    )
+}
+
 bambu_resolve_inputs <- function(args, bamfile_list_ctor = Rsamtools::BamFileList) {
     sample_df <- NULL
     if (!bambu_missing(args$sample_sheet)) {
-        sample_df <- utils::read.csv(
+        sample_df <- bambu_read_sample_sheet(args$sample_sheet)
-            args$sample_sheet,
-            check.names = FALSE,
-            stringsAsFactors = FALSE
-        )
         if (!"alias" %in% names(sample_df)) {
             stop("Sample sheet must contain an 'alias' column.", call. = FALSE)
         }
@@ -1028,11 +1036,7 @@ bambu_collate_chunk_outputs <- function(
     })
     raw_se <- bambu_combine_transcript_chunks(tx_ses)
 
-    sample_df <- utils::read.csv(
+    sample_df <- bambu_read_sample_sheet(file.path(chunk_dirs[[1]], "samples.csv"))
-        file.path(chunk_dirs[[1]], "samples.csv"),
-        check.names = FALSE,
-        stringsAsFactors = FALSE
-    )
 
     gene_se <- gene_expression_fn(raw_se)
     filtered <- bambu_filter_transcripts(raw_se)

bin/workflow_glue_r/tests/testthat/test_bambu.R

@@ -223,6 +223,35 @@ testthat::test_that("sample sheet reordered to match BAMs", {
     )
 })
 
+testthat::test_that("numeric alias and sample_id values are preserved as strings", {
+    sample_sheet <- tempfile(fileext = ".csv")
+    writeLines(
+        paste(
+            "barcode,sample_id,alias,condition",
+            "barcode01,01,01,control",
+            "barcode02,02,02,treated",
+            sep = "\n"
+        ),
+        sample_sheet
+    )
+
+    args <- list(
+        bams = "sample1.bam,sample2.bam",
+        aliases = "01,02",
+        sample_sheet = sample_sheet
+    )
+    resolved <- bambu_resolve_inputs(
+        args,
+        bamfile_list_ctor = function(paths, yieldSize) paths
+    )
+
+    testthat::expect_equal(resolved$aliases, c("01", "02"))
+    testthat::expect_equal(resolved$sample_df$alias, c("01", "02"))
+    testthat::expect_equal(resolved$sample_df$sample_id, c("01", "02"))
+    testthat::expect_type(resolved$sample_df$alias, "character")
+    testthat::expect_type(resolved$sample_df$sample_id, "character")
+})
+
 # Explicit discovery/quant flags are passed through to bambu consistently.
 # NDR is only passed during discovery and omitted when automatic selection is wanted.
 testthat::test_that("bambu args include requested discovery and quant flags", {