Merge pull request #195 from stephenslab/revamp-dscquery-2

Revise dscquery to address Issue #182, implemented new function dscread

dscrutils/DESCRIPTION

@@ -1,8 +1,8 @@
 Package: dscrutils
 Encoding: UTF-8
 Type: Package
-Version: 0.3.8
-Date: 2019-04-16
+Version: 0.3.8.7
+Date: 2019-05-15
 Title: Dynamic Statistical Comparisons R Interface
 Authors@R: c(person("Gao","Wang",role=c("aut","cre"),
                     email="[email protected]"),

dscrutils/NAMESPACE

@@ -1,9 +1,9 @@
 # Generated by roxygen2: do not edit by hand
 
 export(dscquery)
+export(dscread)
 export(dscreval)
 export(load_inputs)
-export(read_dsc)
 export(run_cmd)
 export(save_session)
 export(shiny_plot)

dscrutils/R/dscread.R

@@ -0,0 +1,90 @@
+#' @title Read DSC Module Outputs
+#'
+#' @description Reads in DSC module outputs generated from a single
+#' run of a module instance.
+#'
+#' @details DSC module outputs are either stored in RDS files (see
+#' \code{\link{readRDS}}) or a Python "pickle" file. For DSC module
+#' outputs stored as Python pickle files, the reticulate package is
+#' used to import the data into R.
+#' 
+#' @param outdir Directory where the DSC output is stored.
+#'
+#' @param outfile File specifying the file path relative to the DSC
+#' directory. You can use \code{\link{dscquery}} with the
+#' \code{module.output.file} to obtain a correct file path. Note that
+#' the file path should not contain the file extension (".rds" or
+#' ".pkl").
+#'
+#' @return The return file is a list containing the DSC module
+#' outputs. This list always includes a "DSC_DEBUG" list element
+#' containing additional information recorded by DSC, such as the
+#' replicate id.
+#'
+#' @seealso \code{\link{dscquery}}
+#' 
+#' @examples
+#'
+#' dsc.dir <- system.file("datafiles","one_sample_location",
+#'                        "dsc_result",package = "dscrutils")
+#' dat <- dscquery(dsc.dir,targets = "simulate",
+#'                 module.output.file = "simulate")
+#' out <- dscread(dsc.dir,dat$simulate.output.file[1])
+#' 
+#' @importFrom tools file_ext
+#' @importFrom yaml yaml.load_file
+#' 
+#' @export
+#'
+dscread <- function (outdir, outfile) {
+
+  # Check the input arguments.
+  if (!(is.character(outdir) & length(outdir) == 1))
+    stop("Argument \"outdir\" should be a character vector of length 1")
+  if (!(is.character(outfile) & length(outfile) == 1))
+    stop("Argument \"outfile\" should be a character vector of length 1")
+
+  # Look for files with extensions "rds" and "pkl".
+  outfile <- path.expand(file.path(outdir,outfile))
+  rds     <- paste0(outfile,".rds")
+  pkl     <- paste0(outfile,".pkl")
+  if (file.exists(rds) & file.exists(pkl))
+    stop(sprintf(paste("Both %s and %s DSC output files exist; files should",
+                       "be cleaned up by running \"dsc --clean\""),rds,pkl))
+  else if (file.exists(rds))
+
+    # Read from the .rds file.
+    out <- tryCatch(readRDS(rds),
+      error = function (e) {
+        warning(sprintf("Unable to read from %s; file may be corrupted",rds))
+        return(NULL)
+      })
+  else if (file.exists(pkl)) {
+
+    # Read from the .pkl file.
+    if (!requireNamespace("reticulate",quietly = TRUE))
+      stop("Cannot read from .pkl file due to missing reticulate package")
+    out <- tryCatch(reticulate::py_load_object(pkl),
+      error = function (e) {
+        warning(sprintf("Unable to read from %s; file may be corrupted",pkl))
+        return(NULL)
+      })
+
+    # This additional processing step is needed to convert more
+    # complex Python data structures such as a pandas data frames.
+    out <- rapply(out,reticulate::py_to_r,classes = "python.builtin.object",
+                  how = "replace")
+  } else {
+    warning(sprintf(paste("Unable to read from DSC output file %s as one or",
+                          "more files may be missing; returning NULL"),
+                    outfile))
+    out <- NULL
+  }
+
+  # We may use this code in the future to read from YAML files:
+  #
+  #   yaml.load_file(outfile)
+  #
+  return(out)
+}
+

dscrutils/R/utils.R

@@ -9,38 +9,6 @@ merge_lists <- function(x, y, ...)
   return(x)
 }
 
-#' @importFrom tools file_ext
-#' @importFrom yaml yaml.load_file
-#' @export
-#'
-read_dsc <- function(infile) {
-  inext = file_ext(infile)
-  if (inext == "") {
-    for (item in c("rds", "pkl", "yml")) {
-      if (file.exists(paste0(infile, ".", item))) {
-        inext = item
-        infile = paste0(infile, ".", item)
-        break
-      }
-    }
-  }
-  if (inext == "") stop(paste("Cannot determine extension for input file", infile))
-  if (inext == 'pkl') {
-    ## Do not use `importFrom` for reticulate because
-    ## this is not always required
-    ## and when it is required the DSC parser code
-    ## will check and install the package.
-    if (!requireNamespace("reticulate",quietly = TRUE))
-      stop("Cannot read Python's `pkl` files due to missing `reticulate` package.")
-    result = reticulate::py_load_object(infile)
-    return(rapply(result, reticulate::py_to_r, classes = "python.builtin.object", how = "replace"))
-  } else if (inext == 'yml') {
-    return(yaml.load_file(infile))
-  } else {
-    return(readRDS(infile))
-  }
-}
-
 #' @export
 load_inputs <- function(files, loader) {
   if (length(files) == 1) {
@@ -107,4 +75,4 @@ source_dir <- function(folder, recursive = TRUE, ...)
 
 source_dirs <- function(folders, ...) {
   for (folder in folders) source_dir(folder, ...)
-}
+}

dscrutils/README.md

@@ -49,7 +49,7 @@ command-line shell:
 
 ```bash
 R CMD build dscrutils
-R CMD check --as-cran dscrutils_0.3.6.4.tar.gz
+R CMD check --as-cran dscrutils_0.3.8.7.tar.gz
 ```
 
 To update the package documentation from the