Merge pull request #343 from rformassspectrometry/phili

Addition of cbind2() to append multiple spectra variables to the spectra data

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: Spectra
 Title: Spectra Infrastructure for Mass Spectrometry Data
-Version: 1.17.1
+Version: 1.17.2
 Description: The Spectra package defines an efficient infrastructure
    for storing and handling mass spectrometry spectra and functionality to
    subset, process, visualize and compare spectra data. It provides different

NAMESPACE

@@ -75,6 +75,7 @@ exportMethods(backendParallelFactor)
 exportMethods(backendRequiredSpectraVariables)
 exportMethods(bin)
 exportMethods(c)
+exportMethods(cbind2)
 exportMethods(centroided)
 exportMethods(collisionEnergy)
 exportMethods(combinePeaks)
@@ -309,4 +310,5 @@ importMethodsFrom(S4Vectors,extractROWS)
 importMethodsFrom(S4Vectors,isEmpty)
 importMethodsFrom(S4Vectors,lapply)
 importMethodsFrom(S4Vectors,split)
+importMethodsFrom(methods,cbind2)
 importMethodsFrom(methods,show)

NEWS.md

@@ -1,5 +1,11 @@
 # Spectra 1.17
 
+## Change in 1.17.2
+
+- Add `cbind2()` method to easily add multiple `spectraVariables` and their 
+  content to the `spectraData` of a `Spectra` object. 
+  See also [issue #342](https://github.com/rformassspectrometry/Spectra/issues/342)
+
 ## Changes in 1.17.1
 
 - Refactor `containsMz()` to support chunk-wise processing.

R/MsBackend.R

@@ -182,10 +182,15 @@
 #' @param value replacement value for `<-` methods. See individual
 #'     method description or expected data type.
 #'
-#' @param values for `filterValues()`: A `numeric` vector that define the
+#' @param values For `filterValues()`: A `numeric` vector that define the
 #'     values to filter the `object`. `values` needs to be of same length than
 #'     parameter `spectraVariables` and in the same order.
 #'
+#' @param y For `cbind2()`: A `data.frame` or `DataFrame` with the
+#'     spectra variables to be added to the backend. Need to be of the same
+#'     length as the number of spectra in the backend. The number of rows and
+#'     their order has to match the number of spectra and their order in x.
+#'
 #' @param x Object extending `MsBackend`.
 #'
 #' @param ... Additional arguments.
@@ -313,6 +318,12 @@
 #'   `dropNaSpectraVariables()` might still show columns containing `NA` values
 #'   for *core* spectra variables.
 #'
+#' - `cbind2()`: allows to appends multiple spectra variables to the backend at
+#'   once. The `Spectra` and the values for the new spectra variables have to
+#'   be in a matching order. Replacing existing spectra variables is not
+#'   supported through this function. For a more controlled way of adding
+#'   spectra variables, the `joinSpectraData()` should be used.
+#'
 #' - `centroided()`, `centroided<-`: gets or sets the centroiding
 #'   information of the spectra. `centroided()` returns a `logical`
 #'   vector of length equal to the number of spectra with `TRUE` if a
@@ -1022,6 +1033,29 @@ setMethod("peaksVariables", "MsBackend", function(object) {
     c("mz", "intensity")
 })
 
+
+setClassUnion("dataframeOrDataFrameOrmatrix", c("data.frame", "DataFrame", "matrix"))
+#' @exportMethod cbind2
+#'
+#' @importMethodsFrom methods cbind2
+#'
+#' @rdname MsBackend
+setMethod("cbind2", signature = c("MsBackend", "dataframeOrDataFrameOrmatrix"),
+          function(x, y = data.frame(), ...) {
+    if (is(y, "matrix"))
+        y <- as.data.frame(y)
+    if (any(colnames(spectraData(x)) %in% colnames(y)))
+        stop("spectra variables in 'y' are already present in 'x' ",
+             "replacing them is not allowed")
+    if (nrow(y) != length(x))
+        stop("Number of row in 'y' does not match the number of spectra in 'x'")
+    for (i in colnames(y)) {
+        x[[i]] <- y[, i]
+    }
+    x
+})
+
+
 #' @exportMethod centroided
 #'
 #' @aliases centroided<-,MsBackend-method
@@ -1344,7 +1378,7 @@ setMethod("filterRanges", "MsBackend",
                   return(object)
               if (!is.numeric(ranges))
                   stop("filterRanges only support filtering for numerical ",
-                  "'spectraVariables'")
+                       "'spectraVariables'")
               match <- match.arg(match)
               if (is.character(spectraVariables)){
                   if(!all(spectraVariables %in% spectraVariables(object)))
@@ -1354,7 +1388,7 @@ setMethod("filterRanges", "MsBackend",
                            "function to list possible values.")
               } else
                   stop("The 'spectraVariables' parameter needs to be of type ",
-                  "'character'.")
+                       "'character'.")
               if (length(spectraVariables) != length(ranges) / 2)
                   stop("Length of 'ranges' needs to be twice the length of ",
                        "the parameter 'spectraVariables' and define the lower ",

R/MsBackendDataFrame.R

@@ -567,6 +567,25 @@ setMethod("[", "MsBackendDataFrame", function(x, i, j, ..., drop = FALSE) {
     .subset_backend_data_frame(x, i)
 })
 
+#' @importMethodsFrom methods cbind2
+#'
+#' @rdname hidden_aliases
+setMethod("cbind2", signature = c("MsBackendDataFrame",
+                                  "dataframeOrDataFrameOrmatrix"),
+          function(x, y = data.frame(), ...) {
+              if (is(y, "matrix"))
+                  y <- as.data.frame(y)
+              if (any(colnames(spectraData(x)) %in% colnames(y)))
+                  stop("spectra variables in 'y' are already present in 'x' ",
+                       "replacing them is not allowed")
+              if (nrow(y) != length(x))
+                  stop("Number of row in 'y' does not match the number of ",
+                       "spectra in 'x'")
+              x@spectraData <- cbind(x@spectraData, y)
+              validObject(x)
+              x
+          })
+
 #' @rdname hidden_aliases
 setMethod("split", "MsBackendDataFrame", function(x, f, drop = FALSE, ...) {
     if (!is.factor(f))

R/MsBackendMemory.R

@@ -670,6 +670,26 @@ setMethod("[", "MsBackendMemory", function(x, i, j, ..., drop = FALSE) {
     .df_subset(x, i)
 })
 
+#' @importMethodsFrom methods cbind2
+#'
+#' @rdname hidden_aliases
+setMethod("cbind2", signature = c("MsBackendMemory",
+                                  "dataframeOrDataFrameOrmatrix"),
+          function(x, y = data.frame(), ...) {
+              if (is(y, "matrix"))
+                  y <- as.data.frame(y)
+              if (any(colnames(spectraData(x)) %in% colnames(y)))
+                  stop("spectra variables in 'y' are already present in 'x' ",
+                       "replacing them is not allowed")
+
+              if (nrow(y) != length(x))
+                  stop("Number of row in'y' does not match the number of ",
+                       "spectra in 'x'")
+              x@spectraData <- cbind(x@spectraData, y)
+              validObject(x)
+              x
+          })
+
 #' @rdname hidden_aliases
 setMethod("split", "MsBackendMemory", function(x, f, drop = FALSE, ...) {
     if (!is.factor(f))

R/Spectra.R

@@ -1447,6 +1447,7 @@ setReplaceMethod("[[", "Spectra", function(x, i, j, ..., value) {
 #' @aliases combineSpectra
 #' @aliases split
 #' @aliases joinSpectraData
+#' @aliases cbind2
 #'
 #' @description
 #'
@@ -1463,6 +1464,16 @@ setReplaceMethod("[[", "Spectra", function(x, i, j, ..., value) {
 #'   function and to eventually (if needed) apply the processing queue using
 #'   the [applyProcessing()] function.
 #'
+#' - `cbind2()`: Appends multiple spectra variables from a `data.frame`,
+#'   `DataFrame` or `matrix` to the `Spectra`  object at once. It does so
+#'   *blindly* (e.g. do not check rownames compatibility) and is therefore at
+#'   the risk of the user. The function also does not allow to replace existing
+#'   spectra variables. For a more  controlled way of adding spectra
+#'   variables, the `joinSpectraData()`  should be used. It will return a
+#'   `Spectra` object with the appended spectra variables. `cbind2()` does
+#'   check however that the number of rows of the `data.frame` or `DataFrame`
+#'   matches the number of spectra in the `Spectra` object.
+#'
 #' - `combineSpectra()`: combines sets of spectra (defined with parameter `f`)
 #'   into a single spectrum per set aggregating their MS data (i.e. their
 #'   *peaks data* matrices with the *m/z* and intensity values of their
@@ -1507,6 +1518,8 @@ setReplaceMethod("[[", "Spectra", function(x, i, j, ..., value) {
 #'      should be explored and ideally be removed using for
 #'      `QFeatures::reduceDataFrame()`, `PMS::reducePSMs()` or similar
 #'      functions.
+#'    For a more general function that allows to append `data.frame`,
+#'    `DataFrame` and `matrix` see `cbind2()`.
 #'
 #' - `split()`: splits the `Spectra` object based on parameter `f` into a `list`
 #'   of `Spectra` objects.
@@ -1543,7 +1556,10 @@ setReplaceMethod("[[", "Spectra", function(x, i, j, ..., value) {
 #'
 #' @param x A `Spectra` object.
 #'
-#' @param y A `DataFrame` with the spectra variables to join/add.
+#' @param y For `joinSpectraData()`: `DataFrame` with the spectra variables
+#'     to join/add. For `cbind2()`: a `data.frame`, `DataFrame` or
+#'     `matrix`. The number of rows and their order has to match the
+#'     number of spectra in `x`, respectively their order.
 #'
 #' @param ... Additional arguments.
 #'
@@ -1660,6 +1676,10 @@ setReplaceMethod("[[", "Spectra", function(x, i, j, ..., value) {
 #'
 #' spectraVariables(sciex2)
 #' spectraData(sciex2)[1:13, c("spectrumId", "var1", "var2")]
+#'
+#' ## Append new spectra variables with cbind2()
+#' df <- data.frame(cola = seq_len(length(sciex1)), colb = "b")
+#' data_append <- cbind2(sciex1, df)
 NULL
 
 #' @rdname combineSpectra
@@ -1669,6 +1689,15 @@ setMethod("c", "Spectra", function(x, ...) {
     .concatenate_spectra(unname(list(unname(x), ...)))
 })
 
+#' @rdname combineSpectra
+#'
+#' @export
+setMethod("cbind2", signature(x = "Spectra",
+                              y = "dataframeOrDataFrameOrmatrix"),
+          function(x, y, ...) {
+              x@backend <- cbind2(x@backend, y, ...)
+              })
+
 #' @rdname combineSpectra
 setMethod("split", "Spectra", function(x, f, drop = FALSE, ...) {
     bcknds <- split(x@backend, f, ...)

inst/test_backends/test_MsBackend/test_spectra_subsetting.R

@@ -59,6 +59,28 @@ test_that("[", {
     expect_equal(res, be[which(l)])
 })
 
+test_that("cbind2 works", {
+    seql <- length(be)
+    df <- data.frame(cola = seq_len(seql), colb = "b", colz = "z")
+    res <- cbind2(be, df)
+    expect_true(validObject(res))
+    expect_equal(ncol(spectraData(res)), length(spectraVariables(be)) + 3)
+    expect_equal(res$cola, seq_len(seql))
+    expect_equal(res$colb, rep("b", seql))
+    expect_equal(res$colz, rep("z", seql))
+    df2  <- data.frame(cola = 3:6, colb = "b", colz = "z")
+    expect_error(cbind2(be, df2), "does not match")
+    ## with matrix
+    m <- matrix(1:seql, ncol = 1, dimnames = list(NULL, "m"))
+    res <- cbind2(be, m)
+    expect_true(validObject(res))
+    expect_equal(ncol(spectraData(res)), length(spectraVariables(be)) + 1)
+    expect_equal(res$m, 1:seql)
+    ## no replacing
+    expect_error(cbind2(be, data.frame(scanIndex = 1:seql)),
+                 "are already present")
+})
+
 #' extractByIndex. Uses [ if not implemented
 test_that("extractByIndex", {
     i <- sample(seq_along(be), floor(length(be) / 2))