update

taxonomicallyinformedannotation · Jul 31, 2023 · 6bc879e · 6bc879e
1 parent 2baf503
commit 6bc879e
Show file tree

Hide file tree

Showing 24 changed files with 139 additions and 278 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -19,13 +19,11 @@ Imports:
     docopt (>= 0.7.1),
     dplyr (>= 1.1.2),
     httr (>= 1.4.6),
-    future (>= 1.33.0),
-    future.apply (>= 1.11.0),
     igraph (>= 1.5.0.1),
     jsonlite (>= 1.8.7),
     MsBackendMgf (>= 1.8.0),
     msentropy (>= 0.1.3),
-    progressr (>= 0.13.0),
+    pbapply (>= 1.7.2),
     R.utils (>= 2.12.2),
     rotl (>= 3.1.0),
     Spectra (>= 1.10.1),
@@ -81,7 +79,7 @@ Collate:
     'create_adducts_neg.R'
     'create_adducts_pos.R'
     'create_components.R'
-    'create_edges_parallel.R'
+    'create_edges.R'
     'keep_peaks.R'
     'sanitize_spectra.R'
     'remove_above_precursor.R'

diff --git a/NAMESPACE b/NAMESPACE
@@ -11,7 +11,7 @@ export(create_adducts_neg)
 export(create_adducts_pos)
 export(create_components)
 export(create_dir)
-export(create_edges_parallel)
+export(create_edges)
 export(create_edges_spectra)
 export(decorate_bio)
 export(decorate_chemo)

diff --git a/NEWS.md b/NEWS.md
@@ -2,7 +2,8 @@
 
 # timaR 2.8.2
 
-* Change from pbmclapply to future.lapply
+* Change from pbmclapply to pblapply
+* Added spectral entropy
 * Fix empty chemical classes
 * [renv](https://rstudio.github.io/renv/index.html) removal
 * Performance improvement by replacing the [tidyverse](https://www.tidyverse.org) by the [fastverse](https://fastverse.github.io/fastverse) (in progress)

diff --git a/R/annotate_spectra.R b/R/annotate_spectra.R
@@ -27,7 +27,6 @@ utils::globalVariables(
 #'
 #' @details It takes two files as input.
 #'    A query file that will be matched against a library file.
-#'    Parallel processing is also made available.
 #'
 #' @include export_output.R
 #' @include export_params.R
@@ -44,7 +43,6 @@ utils::globalVariables(
 #' @param condition Condition to be fulfilled.
 #'    Either 'OR' or 'AND' (mass and peaks minima).
 #' @param qutoff Intensity under which ms2 fragments will be removed.
-#' @param parallel Boolean. Process in parallel
 #' @param approx Perform matching without precursor match
 #' @param parameters Params
 #'
@@ -62,7 +60,6 @@ annotate_spectra <- function(input = params$files$spectral$raw,
                              ppm = params$ms$tolerances$mass$ppm$ms2,
                              dalton = params$ms$tolerances$mass$dalton$ms2,
                              qutoff = params$ms$intensity$thresholds$ms2,
-                             parallel = params$options$parallel,
                              approx = params$annotations$ms2$approx,
                              parameters = params) {
   stopifnot("Your input file does not exist." = file.exists(input))
@@ -235,65 +232,34 @@ annotate_spectra <- function(input = params$files$spectral$raw,
                     )
                   }
                 }
-              )
+              ) |>
+              Filter(f = Negate(is.null)) |>
+              dplyr::bind_rows()
 
-            inner_list <- Filter(Negate(is.null), inner_list)
-            inner_list <- dplyr::bind_rows(inner_list)
             return(inner_list)
           }
 
         log_debug("Performing spectral comparison")
-        ## TODO investigate issue with Windows
-        if (parallel && Sys.info()[["sysname"]] != "Windows") {
-          options(future.globals.onReference = "error")
-          future::plan(future::multisession)
-          progressr::handlers(list(
-            progressr::handler_progress(
-              format = ":current/:total [:bar] :percent in :elapsed ETA: :eta"
-            )
-          ))
-          outer_list <-
-            future.apply::future_lapply(
-              X = seq_along(spectra),
-              p = progressr::progressor(along = seq_along(spectra)),
-              future.seed = TRUE,
-              FUN = function(spectrum, qp = query_precursors, p) {
-                p(sprintf("spectra=%g", length(qp)))
-                precursor <- qp[spectrum]
-                calculate_score_and_create_inner_list(
-                  spectrum = spectrum,
-                  precursor = precursor,
-                  spectral_lib = lib_spectra,
-                  query_spectra = query_spectra,
-                  query_rts = query_rts,
-                  lib_id = lib_id,
-                  minimal = minimal,
-                  maximal = maximal
-                )
-              }
-            ) |>
-            progressr::with_progress()
-        } else {
-          outer_list <-
-            lapply(
-              X = seq_along(spectra),
-              FUN = function(spectrum, qp = query_precursors) {
-                precursor <- qp[spectrum]
-                calculate_score_and_create_inner_list(
-                  spectrum = spectrum,
-                  precursor = precursor,
-                  spectral_lib = lib_spectra,
-                  query_spectra = query_spectra,
-                  query_rts = query_rts,
-                  lib_id = lib_id,
-                  minimal = minimal,
-                  maximal = maximal
-                )
-              }
-            )
-        }
+        outer_list <-
+          pbapply::pblapply(
+            X = seq_along(spectra),
+            FUN = function(spectrum, qp = query_precursors) {
+              precursor <- qp[spectrum]
+              calculate_score_and_create_inner_list(
+                spectrum = spectrum,
+                precursor = precursor,
+                spectral_lib = lib_spectra,
+                query_spectra = query_spectra,
+                query_rts = query_rts,
+                lib_id = lib_id,
+                minimal = minimal,
+                maximal = maximal
+              )
+            }
+          ) |>
+          dplyr::bind_rows()
 
-        return(dplyr::bind_rows(outer_list))
+        return(outer_list)
       }
 
     df_final <-

diff --git a/R/calculate_entropy.R b/R/calculate_entropy.R
@@ -7,6 +7,7 @@
 #' @param frags Fragments
 #' @param ms2_tolerance MS2 tolerance
 #' @param ppm_tolerance ppm tolerance
+#' @param threshold Threshold
 #'
 #' @return NULL
 #'
@@ -17,7 +18,8 @@ calculate_entropy <- function(index,
                               target,
                               frags,
                               ms2_tolerance,
-                              ppm_tolerance) {
+                              ppm_tolerance,
+                              threshold = 0.1) {
   score <-
     msentropy::calculate_entropy_similarity(
       frags[[index]],
@@ -31,7 +33,7 @@ calculate_entropy <- function(index,
       clean_spectra = TRUE
     )
 
-  if (score >= 0.1) {
+  if (score >= threshold) {
     return(
       list(
         "feature_id" = index,

diff --git a/R/create_edges.R b/R/create_edges.R
@@ -0,0 +1,48 @@
+utils::globalVariables(
+  c(
+    "params"
+  )
+)
+
+#' @title Create edges
+#'
+#' @description This function applies similarity calculation to a list of
+#'        spectra to create edges
+#'
+#' @include calculate_entropy.R
+#'
+#' @param index Indices
+#' @param frags Fragments
+#' @param precs Precursors
+#' @param nspecs Number of spectra
+#' @param ms2_tolerance MS2 tolerance
+#' @param ppm_tolerance ppm tolerance
+#' @param threshold Threshold
+#'
+#' @return NULL
+#'
+#' @export
+#'
+#' @examples NULL
+create_edges <- function(index,
+                         frags,
+                         precs,
+                         nspecs,
+                         ms2_tolerance,
+                         ppm_tolerance,
+                         threshold) {
+  # Calculate the similarity using lapply
+  inner_list <- lapply(X = (index + 1):nspecs, FUN = function(target) {
+    calculate_entropy(
+      index,
+      target,
+      frags,
+      ms2_tolerance,
+      ppm_tolerance,
+      threshold
+    )
+  }) |>
+    dplyr::bind_rows()
+
+  return(inner_list)
+}
diff --git a/R/create_edges_parallel.R b/R/create_edges_parallel.R
diff --git a/R/create_edges_spectra.R b/R/create_edges_spectra.R
@@ -10,7 +10,7 @@ utils::globalVariables(c(
 #' @description This function create edges
 #'    based on fragmentation spectra similarity
 #'
-#' @include create_edges_parallel.R
+#' @include create_edges.R
 #' @include import_spectra.R
 #' @include normalize_peaks.R
 #' @include remove_above_precursor.R
@@ -25,7 +25,6 @@ utils::globalVariables(c(
 #' @param ppm Relative ppm tolerance to be used
 #' @param dalton Absolute Dalton tolerance to be used
 #' @param qutoff Intensity under which ms2 fragments will be removed.
-#' @param parallel Boolean. Process in parallel
 #' @param parameters Params
 #'
 #' @return NULL
@@ -42,7 +41,6 @@ create_edges_spectra <- function(
     ppm = params$ms$tolerances$mass$ppm$ms2,
     dalton = params$ms$tolerances$mass$dalton$ms2,
     qutoff = params$ms$intensity$thresholds$ms2,
-    parallel = params$options$parallel,
     parameters = params) {
   stopifnot("Your input file does not exist." = file.exists(input))
   ## Not checking for ppm and Da limits, everyone is free.
@@ -62,62 +60,26 @@ create_edges_spectra <- function(
 
   log_debug("Performing spectral comparison")
   log_debug(
-    "As we do not bin the spectra,
-      nor limit the precursors delta,
-      expect a long processing time."
+    "As we do not limit the precursors delta,
+      expect a (relatively) long processing time."
   )
   log_debug("Take yourself a break, you deserve it.")
   nspecz <- length(spectra)
   precz <- spectra$precursorMz
   fragz <- spectra@backend@peaksData
 
-  if (parallel) {
-    options(future.globals.onReference = "error")
-    future::plan(future::multisession)
-    progressr::handlers(
-      list(
-        progressr::handler_progress(
-          format = ":current/:total [:bar] :percent in :elapsed ETA: :eta"
-        )
-      )
-    )
-    result_list <-
-      future.apply::future_lapply(
-        X = 1:(nspecz - 1),
-        FUN = create_edges_parallel,
-        p = progressr::progressor(along = 1:(nspecz - 1)),
-        future.seed = TRUE,
-        future.chunk.size = structure(TRUE, ordering = "random"),
-        frags = fragz,
-        precs = precz,
-        nspecs = nspecz,
-        ms2_tolerance = dalton,
-        ppm_tolerance = ppm,
-        parallel = parallel
-      ) |>
-      progressr::with_progress()
-  } else {
-    result_list <-
-      lapply(
-        X = 1:(nspecz - 1),
-        FUN = create_edges_parallel,
-        frags = fragz,
-        precs = precz,
-        nspecs = nspecz,
-        ms2_tolerance = dalton,
-        ppm_tolerance = ppm,
-        p = NA,
-        parallel = parallel
-      )
-  }
-  edges <- do.call(
-    what = rbind,
-    args = unlist(result_list, recursive = FALSE)
-  ) |>
-    data.frame()
-
-  log_debug("Collecting garbage ...")
-  gc()
+  edges <-
+    pbapply::pblapply(
+      X = 1:(nspecz - 1),
+      FUN = create_edges,
+      frags = fragz,
+      precs = precz,
+      nspecs = nspecz,
+      ms2_tolerance = dalton,
+      ppm_tolerance = ppm,
+      threshold = threshold
+    ) |>
+    dplyr::bind_rows()
 
   edges <- edges |>
     tidytable::select(

diff --git a/R/parse_cli_params.R b/R/parse_cli_params.R
@@ -345,9 +345,6 @@ parse_cli_params <- function() {
     if (!is.null(arguments$force)) {
       params$options$force <- as.logical(arguments$force)
     }
-    if (!is.null(arguments$parallel)) {
-      params$options$parallel <- as.logical(arguments$parallel)
-    }
     if (!is.null(arguments$summarise)) {
       params$options$summarise <- as.logical(arguments$summarise)
     }