Merge some bugfixes for DIANN dataset into Development

R/module-loadpage-server.R

@@ -82,21 +82,22 @@ loadpageServer <- function(id, parent_session, is_web_server = FALSE, app_templa
       }
     })
 
-    # Read first 100 rows for Metamorpheus PTM preview.
-    # TODO: Extend preview reading to other input formats (e.g., Spectronaut, MaxQuant)
-    # for dynamic UI updates. Currently limited to Metamorpheus.
+    # Read first 100 rows for preview-based UI features.
+    # Supported: Metamorpheus PTM (modification ID dropdown), DIANN (version auto-detection).
+    # TODO: Extend to other input formats (Spectronaut, MaxQuant) as needed.
     observe({
-      if (isTRUE(input$filetype == "meta") && isTRUE(input$BIO == "PTM")) {
+      should_preview <- (isTRUE(input$filetype == "meta") && isTRUE(input$BIO == "PTM")) ||
+                       (isTRUE(input$filetype == "diann") && isTRUE(input$BIO != "PTM"))
+      if (should_preview) {
         file_info <- main_data_file()
         if (!is.null(file_info)) {
-          preview <- tryCatch(
-            data.table::fread(file_info$datapath, nrows = 100, header = TRUE),
-            error = function(e) {
-              showNotification(paste("Could not preview file:", conditionMessage(e)),
-                               type = "warning", duration = 5)
-              NULL
-            }
-          )
+          # Reset DIANN detection tracker so a new file re-triggers the notification
+          last_detected_diann_format(NULL)
+          preview <- .read_preview(file_info$datapath, file_info$name)
+          if (is.null(preview)) {
+            showNotification("Could not preview file. Please verify the file format.",
+                             type = "warning", duration = 5)
+          }
           preview_data(preview)
         } else {
           preview_data(NULL)
@@ -106,6 +107,48 @@ loadpageServer <- function(id, parent_session, is_web_server = FALSE, app_templa
       }
     })
 
+    # Track last detected DIANN format to avoid redundant notifications
+    last_detected_diann_format <- reactiveVal(NULL)
+
+    # Auto-toggle DIANN 2.0+ checkbox based on detected file format
+    observe({
+      req(input$filetype == "diann", input$BIO != "PTM")
+      preview <- preview_data()
+      if (is.null(preview)) return()
+
+      is_2plus <- .is_diann_2plus(preview)
+      previous <- last_detected_diann_format()
+      # Only update and notify when the detected state actually changes
+      if (is.null(previous) || previous != is_2plus) {
+        updateCheckboxInput(session, "diann_2plus", value = is_2plus)
+        if (is_2plus) {
+          showNotification("Detected DIANN 2.0+ format (per-fragment columns).",
+                           type = "message", duration = 5)
+        } else {
+          showNotification("Detected DIANN 1.x format (legacy fragment column).",
+                           type = "message", duration = 5)
+        }
+        last_detected_diann_format(is_2plus)
+      }
+    })
+
+    # Warn user if they manually set DIANN 2.0+ checkbox to a value that conflicts with detected format
+    observeEvent(input$diann_2plus, {
+      req(input$filetype == "diann", input$BIO != "PTM")
+      preview <- preview_data()
+      if (is.null(preview)) return()
+      detected_2plus <- .is_diann_2plus(preview)
+      if (isTRUE(input$diann_2plus) != detected_2plus) {
+        showNotification(
+          paste0("Warning: You've ",
+                 if (isTRUE(input$diann_2plus)) "checked" else "unchecked",
+                 " DIANN 2.0+, but the uploaded file appears to be ",
+                 if (detected_2plus) "DIANN 2.0+ format" else "DIANN 1.x format",
+                 ". This mismatch may cause upload to fail."),
+          type = "warning", duration = 10)
+      }
+    }, ignoreInit = TRUE)
+
     # ========= METAMORPHEUS PTM: Dynamic modification ID dropdown =========
     output$mod_id_meta_ui <- renderUI({
       ns <- session$ns
@@ -412,9 +455,17 @@ loadpageServer <- function(id, parent_session, is_web_server = FALSE, app_templa
       getMaxqPtmSites(input)
     })
 
-
     get_data = eventReactive(input$proceed1, {
-      getData(input)
+      tryCatch(
+        getData(input),
+        error = function(e) {
+          tryCatch(remove_modal_spinner(), error = function(e2) NULL)
+          showNotification(
+            paste("Failed to load data:", conditionMessage(e)),
+            type = "error", duration = 12)
+          NULL
+        }
+      )
     })
 
 

R/module-loadpage-ui.R

@@ -602,7 +602,7 @@ create_label_free_options <- function(ns) {
     # DIANN specific options
     conditionalPanel(
       condition = "input['loadpage-filetype'] == 'diann' && input['loadpage-DDA_DIA'] == 'LType'",
-      checkboxInput(ns("diann_2plus"), "DIANN 2.0+", value = TRUE),
+      checkboxInput(ns("diann_2plus"), "DIANN 2.0+", value = FALSE),
       conditionalPanel(
         condition = "!input['loadpage-diann_2plus']",
         textInput(ns("intensity_column"),

R/utils.R

@@ -1,3 +1,47 @@
+#' Read preview columns from a data file (handles CSV, TSV, and Parquet)
+#'
+#' @param filepath Path to the file.
+#' @param filename Original filename (used to detect parquet extension).
+#' @param nrows Number of rows to read. Default 100. Parquet returns columns only.
+#' @return A data frame with up to `nrows` rows, or NULL on error.
+#' @noRd
+.read_preview <- function(filepath, filename = NULL, nrows = 100) {
+  ext <- if (!is.null(filename)) tolower(tools::file_ext(basename(filename))) else ""
+  tryCatch({
+    if (ext %in% c("parquet", "pq")) {
+      # For parquet, read only the schema (column names) to avoid OOM on large files.
+      # Return an empty data frame with the correct column structure for detection.
+      schema <- arrow::open_dataset(filepath, format = "parquet")$schema
+      col_names <- schema$names
+      empty_df <- as.data.frame(
+        setNames(lapply(col_names, function(x) logical(0)), col_names)
+      )
+      empty_df
+    } else {
+      data.table::fread(filepath, nrows = nrows, header = TRUE)
+    }
+  }, error = function(e) NULL)
+}
+
+#' Detect whether a DIANN preview is in 2.0+ format
+#'
+#' DIANN 2.0+ files have per-fragment columns (Fr.0.Quantity, Fr.1.Quantity, etc.)
+#' and no FragmentQuantCorrected column. Older versions use a single
+#' Fragment.Quant.Corrected / FragmentQuantCorrected column.
+#'
+#' @param preview_df Data frame preview of the DIANN file.
+#' @return Logical. TRUE if the file appears to be DIANN 2.0+.
+#' @noRd
+.is_diann_2plus <- function(preview_df) {
+  if (is.null(preview_df) || ncol(preview_df) == 0) return(FALSE)
+  cols <- names(preview_df)
+  # DIANN 2.0+ signature: numbered fragment columns like "Fr.0.Quantity"
+  has_numbered_fragments <- any(grepl("^Fr\\.[0-9]+\\.Quantity$", cols))
+  # DIANN 1.x signature: the legacy fragment column
+  has_legacy_fragments <- any(cols %in% c("Fragment.Quant.Corrected", "FragmentQuantCorrected"))
+  has_numbered_fragments && !has_legacy_fragments
+}
+
 #' Extract unique modification IDs from preview data
 #'
 #' Parses the Full Sequence column to find bracket-enclosed modification IDs.

tests/testthat/test-utils.R

@@ -1650,4 +1650,208 @@ test_that("extract_mod_ids_from_preview handles consecutive modifications", {
   result <- MSstatsShiny:::.extract_mod_ids_from_preview(preview)
   expect_equal(length(result), 3)
   expect_true(all(c("[Mod1]", "[Mod2]", "[Mod3]") %in% result))
+})
+
+# ============================================================================
+# DIANN FORMAT DETECTION TESTS
+# ============================================================================
+
+test_that("is_diann_2plus returns TRUE for DIANN 2.0+ format with numbered fragment columns", {
+  preview <- data.frame(
+    Run = "run1",
+    Protein.Group = "P1",
+    Fr.0.Quantity = 100,
+    Fr.1.Quantity = 200,
+    Fr.2.Quantity = 300,
+    Precursor.Charge = 2
+  )
+  expect_true(MSstatsShiny:::.is_diann_2plus(preview))
+})
+
+test_that("is_diann_2plus returns FALSE for DIANN 1.x format with legacy Fragment.Quant.Corrected", {
+  preview <- data.frame(
+    Run = "run1",
+    Protein.Group = "P1",
+    Fragment.Quant.Corrected = 100,
+    Fragment.Quant.Raw = 95,
+    Precursor.Charge = 2
+  )
+  expect_false(MSstatsShiny:::.is_diann_2plus(preview))
+})
+
+test_that("is_diann_2plus returns FALSE for DIANN 1.x format with FragmentQuantCorrected (no dots)", {
+  preview <- data.frame(
+    Run = "run1",
+    FragmentQuantCorrected = 100,
+    Precursor.Charge = 2
+  )
+  expect_false(MSstatsShiny:::.is_diann_2plus(preview))
+})
+
+test_that("is_diann_2plus returns FALSE when both formats are present (legacy takes precedence)", {
+  preview <- data.frame(
+    Run = "run1",
+    Fragment.Quant.Corrected = 100,
+    Fr.0.Quantity = 200
+  )
+  expect_false(MSstatsShiny:::.is_diann_2plus(preview))
+})
+
+test_that("is_diann_2plus returns FALSE for NULL preview", {
+  expect_false(MSstatsShiny:::.is_diann_2plus(NULL))
+})
+
+test_that("is_diann_2plus returns FALSE for empty data frame", {
+  expect_false(MSstatsShiny:::.is_diann_2plus(data.frame()))
+})
+
+test_that("is_diann_2plus returns FALSE for data with no fragment columns", {
+  preview <- data.frame(
+    Run = "run1",
+    Protein.Group = "P1",
+    Precursor.Charge = 2
+  )
+  expect_false(MSstatsShiny:::.is_diann_2plus(preview))
+})
+
+test_that("is_diann_2plus detects DIANN 2.0+ with many numbered fragment columns", {
+  # Real DIANN 2.0+ files can have Fr.0 through Fr.11
+  cols <- c("Run", "Protein.Group", paste0("Fr.", 0:11, ".Quantity"),
+            paste0("Fr.", 0:11, ".Index"), paste0("Fr.", 0:11, ".Score"))
+  preview <- as.data.frame(setNames(
+    lapply(cols, function(x) if (grepl("Quantity", x)) runif(1) else "x"),
+    cols
+  ))
+  expect_true(MSstatsShiny:::.is_diann_2plus(preview))
+})
+
+# ============================================================================
+# PREVIEW READER TESTS
+# ============================================================================
+
+test_that("read_preview reads CSV files with nrows limit", {
+  tmp <- tempfile(fileext = ".csv")
+  df <- data.frame(a = 1:200, b = letters[1:26][1:200 %% 26 + 1])
+  write.csv(df, tmp, row.names = FALSE)
+
+  preview <- MSstatsShiny:::.read_preview(tmp, "test.csv", nrows = 100)
+  expect_false(is.null(preview))
+  expect_equal(nrow(preview), 100)
+  expect_true(all(c("a", "b") %in% names(preview)))
+
+  unlink(tmp)
+})
+
+test_that("read_preview reads TSV files", {
+  tmp <- tempfile(fileext = ".tsv")
+  df <- data.frame(a = 1:50, b = letters[1:50 %% 26 + 1])
+  write.table(df, tmp, sep = "\t", row.names = FALSE, quote = FALSE)
+
+  preview <- MSstatsShiny:::.read_preview(tmp, "test.tsv", nrows = 100)
+  expect_false(is.null(preview))
+  expect_equal(nrow(preview), 50)
+
+  unlink(tmp)
+})
+
+test_that("read_preview returns NULL for non-existent files", {
+  preview <- MSstatsShiny:::.read_preview("/nonexistent/path.csv", "test.csv")
+  expect_null(preview)
+})
+
+test_that("read_preview returns NULL for malformed files", {
+  tmp <- tempfile(fileext = ".csv")
+  writeBin(as.raw(c(0xFF, 0xFE, 0x00, 0x00)), tmp)  # Garbage bytes
+  preview <- MSstatsShiny:::.read_preview(tmp, "test.csv")
+  # Either NULL or a data frame (fread can sometimes parse garbage) — both acceptable
+  expect_true(is.null(preview) || is.data.frame(preview))
+  unlink(tmp)
+})
+
+test_that("read_preview handles NULL filename gracefully", {
+  tmp <- tempfile(fileext = ".csv")
+  df <- data.frame(a = 1:10)
+  write.csv(df, tmp, row.names = FALSE)
+
+  # Should default to CSV reading path
+  preview <- MSstatsShiny:::.read_preview(tmp, NULL)
+  expect_false(is.null(preview))
+  expect_equal(nrow(preview), 10)
+
+  unlink(tmp)
+})
+
+test_that("read_preview dispatches parquet files to arrow schema reader", {
+  skip_if_not_installed("arrow")
+
+  tmp <- tempfile(fileext = ".parquet")
+  df <- data.frame(a = 1:50, b = runif(50))
+  arrow::write_parquet(df, tmp)
+
+  preview <- MSstatsShiny:::.read_preview(tmp, "test.parquet")
+  expect_false(is.null(preview))
+  # Schema-only read returns 0 rows but correct column names
+  expect_equal(nrow(preview), 0)
+  expect_true(all(c("a", "b") %in% names(preview)))
+
+  unlink(tmp)
+})
+
+test_that("read_preview recognizes .pq extension as parquet", {
+  skip_if_not_installed("arrow")
+
+  tmp <- tempfile(fileext = ".pq")
+  df <- data.frame(a = 1:10)
+  arrow::write_parquet(df, tmp)
+
+  preview <- MSstatsShiny:::.read_preview(tmp, "test.pq")
+  expect_false(is.null(preview))
+
+  unlink(tmp)
+})
+
+# ============================================================================
+# INTEGRATION TESTS: PREVIEW + DIANN DETECTION
+# ============================================================================
+
+test_that("DIANN 1.x CSV file is correctly detected as not 2.0+", {
+  tmp <- tempfile(fileext = ".csv")
+  df <- data.frame(
+    Run = paste0("run", 1:10),
+    Protein.Group = "P1",
+    Fragment.Quant.Corrected = runif(10) * 1000,
+    Fragment.Quant.Raw = runif(10) * 1000,
+    Precursor.Charge = 2,
+    Q.Value = runif(10, 0, 0.01)
+  )
+  write.csv(df, tmp, row.names = FALSE)
+
+  preview <- MSstatsShiny:::.read_preview(tmp, "diann_1x.csv")
+  expect_false(MSstatsShiny:::.is_diann_2plus(preview))
+
+  unlink(tmp)
+})
+
+test_that("DIANN 2.0 parquet file is correctly detected as 2.0+", {
+  skip_if_not_installed("arrow")
+
+  tmp <- tempfile(fileext = ".parquet")
+  df <- data.frame(
+    Run = paste0("run", 1:10),
+    Protein.Group = "P1",
+    Fr.0.Quantity = runif(10) * 1000,
+    Fr.0.Index = 1L,
+    Fr.1.Quantity = runif(10) * 1000,
+    Fr.1.Index = 2L,
+    Fr.2.Quantity = runif(10) * 1000,
+    Fr.2.Index = 3L,
+    Precursor.Charge = 2,
+    Q.Value = runif(10, 0, 0.01)
+  )
+  arrow::write_parquet(df, tmp)
+
+  preview <- MSstatsShiny:::.read_preview(tmp, "diann_2plus.parquet")
+  expect_true(MSstatsShiny:::.is_diann_2plus(preview))
+
+  unlink(tmp)
 })