version 1.8.2

cran · Nov 22, 2023 · 194efc2 · 194efc2
1 parent 111fef2
commit 194efc2
Show file tree

Hide file tree

Showing 20 changed files with 559 additions and 602 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,7  1,7 @@
 Package: outliertree
 Type: Package
 Title: Explainable Outlier Detection Through Decision Tree Conditioning
-Version: 1.8.1-1
 Version: 1.8.2
 Author: David Cortes
 Maintainer: David Cortes <[email protected]>
 URL: https://github.com/david-cortes/outliertree
@@ -12,14  12,14 @@ Description: Outlier detection method that flags suspicious values within observ
   Full procedure is described in Cortes (2020) <arXiv:2001.00636>.
   Loosely based on the 'GritBot' <https://www.rulequest.com/gritbot-info.html> software.
 License: GPL (>= 3)
-Imports: Rcpp (>= 1.0.1)
-Depends: R (>= 3.5.0)
 Imports: Rcpp (>= 1.0.1), methods
 Depends: R (>= 4.3.0)
 Suggests: knitr, rmarkdown, kableExtra, data.table
 LinkingTo: Rcpp, Rcereal
 LazyData: true
 VignetteBuilder: knitr
-RoxygenNote: 7.1.2
 RoxygenNote: 7.2.3
 NeedsCompilation: yes
-Packaged: 2022-07-27 18:40:52 UTC; david
 Packaged: 2023-11-21 21:14:45 UTC; david
 Repository: CRAN
-Date/Publication: 2022-08-06 20:00:02 UTC
 Date/Publication: 2023-11-21 21:50:03 UTC
diff --git a/MD5 b/MD5
@@ -1,44  1,43 @@
-0bb6d82dfe028de1d1c9a8955257361e *DESCRIPTION
-f2f84a6270f5aef385a2a886f1246758 *NAMESPACE
-a3549c32ad34b6c7bd1a2024982a62ba *R/RcppExports.R
-a8b53087fcf6e29fdf5e8519ba1442fd *R/helpers.R
 0191b4ce8f38c3e76d42ee8c94c531cc *DESCRIPTION
 f7ddfbf401ff4029ee143ac87b34bf18 *NAMESPACE
 146ce017ca8e585317f46365f599d82c *R/RcppExports.R
 8f9d710b5dc39d73a8055c336f9f6d95 *R/helpers.R
 3479eaf06504f010287a711cd5aeed33 *R/hypothyroid.R
-f5fb84a0c34408e213347e6d0e3b2fcf *R/obj_methods.R
-e37dfd4db46c0a06daa5c7bd9184c2cd *R/outliertree.R
 b000c22a8e5f588636e309fac6d2b800 *R/obj_methods.R
 867d3986ac0d780a4f2a31ff53899075 *R/outliertree.R
 6e9aeeba689a25e3a5714c376f892c9d *R/titanic.R
-710acbc770c9ab441ea689536bd6e539 *build/vignette.rds
 2e174a8feb7b3f62911d9b56c4e70fbe *build/vignette.rds
 c18cb6d69093cb91329c328ff4a117a0 *configure
 46f6ca6bf4d93b32c26870fe7d3009e1 *configure.ac
 d41d8cd98f00b204e9800998ecf8427e *configure.win
 68bd4e6455985c40493fc54c36e99ec0 *data/hypothyroid.rda
 88bb214afac793d369d89a07007f007b *data/titanic.rda
 707cb95bdf71178625753b2ec698c97e *inst/doc/Explainable_Outlier_Detection_in_Titanic_dataset.R
 cf5ae8500528392a73f9c329806776ea *inst/doc/Explainable_Outlier_Detection_in_Titanic_dataset.Rmd
-dc21f304cce8ab5ac20213243137bd9a *inst/doc/Explainable_Outlier_Detection_in_Titanic_dataset.html
 6f1202eec1dd71943944b016461bfb84 *inst/doc/Explainable_Outlier_Detection_in_Titanic_dataset.html
 c7d31c586d00d153e28c3ffc0e42e9ca *inst/doc/Introducing_OutlierTree.R
 520ac4aa8c838850aaf36d4e11426308 *inst/doc/Introducing_OutlierTree.Rmd
-e7ffd977382d9b679d30c5f4b7278926 *inst/doc/Introducing_OutlierTree.html
 486241bd107ceced52b2db1d3c4d6fab *inst/doc/Introducing_OutlierTree.html
 9e980d45ebf1314f6a7ab4d0735c34f9 *man/as.list.outlieroutputs.Rd
 c1db44adae56d90f7e038ce359f3a1d2 *man/check.outlierness.bounds.Rd
 289001788632a6cb1e1fa4faaa5f6502 *man/extract.training.outliers.Rd
 894e7d20bd0bee99b5154b6fcdc3c25c *man/hypothyroid.Rd
-569135efd7b6147a88c2ff25ec617c7a *man/outlier.tree.Rd
-a3af014eecf636332bb89a3dc30b4431 *man/predict.outliertree.Rd
-007649f14261ce69628d323c54c11831 *man/print.outlieroutputs.Rd
 4a7c7e409ee44b5f970203634f0b4604 *man/outlier.tree.Rd
 50de24911cc616b2ae4e1ddce5e5f798 *man/predict.outliertree.Rd
 1f927ee92106a94883ca322aecc8b048 *man/print.outlieroutputs.Rd
 b7678d0ade3f8be54a2e3952e9b4a0b3 *man/print.outliertree.Rd
 78650c26bdca4bcc0233fec124b45605 *man/sub-.outlieroutputs.Rd
 506e9f197f434716d44c4d944858b87d *man/summary.outlieroutputs.Rd
 e4f9598822d1683a4810f8b70b3f719d *man/summary.outliertree.Rd
 7e6cc5ce3242ff2eda4d9487d18babc1 *man/titanic.Rd
-9031913794b4e414be12a8bc60cc571d *man/unpack.outlier.tree.Rd
-bfcd1b4e4980ae851a22408af8aaeb79 *src/Makevars.in
-6d09028baad6a6136c1c02345ea0cd17 *src/Makevars.win
-9feb25e069cc93179718532dad3ba95c *src/RcppExports.cpp
-ee55e310e831eaf1d5548c9e679bfa38 *src/Rwrapper.cpp
 c751a87f5827956018d5ce3aa35194c0 *src/Makevars.in
 72f4c11f78af661d8cbaf40c179a8318 *src/Makevars.win
 0b663d0a726e67eac4918a72c1edd747 *src/RcppExports.cpp
 a52f73c46b570268b0af8aa1981a4459 *src/Rwrapper.cpp
 c7a837ecba6b9af232973a449a1e3a93 *src/cat_outlier.cpp
 ca45be6248ced1c119cbe38b5d5da6e3 *src/clusters.cpp
-2dce205f669042b7b1b466e7afd88a45 *src/fit_model.cpp
-4987b3276c61ebbf65a920d66c84e781 *src/misc.cpp
 f3456306c44f650494f9f33e9a3ab6da *src/fit_model.cpp
 06347e747355f1f9f95b8e2a0f748224 *src/misc.cpp
 f67bff9e697a7cee33db78443f870edf *src/outlier_tree.h
 9773402db2b299cbaf88feb87a83deac *src/outliertree-win.def
 08abb4798dd6572955e7ec90c1d660fe *src/predict.cpp

diff --git a/NAMESPACE b/NAMESPACE
@@ -10,8  10,8 @@ S3method(summary,outliertree)
 export(check.outlierness.bounds)
 export(extract.training.outliers)
 export(outlier.tree)
-export(unpack.outlier.tree)
 importFrom(Rcpp,evalCpp)
 importFrom(methods,new)
 importFrom(parallel,detectCores)
 importFrom(stats,predict)
 importFrom(utils,head)

diff --git a/R/RcppExports.R b/R/RcppExports.R
@@ -1,14  1,6 @@
 # Generated by using Rcpp::compileAttributes() -> do not edit by hand
 # Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
 
-deserialize_OutlierTree <- function(src, ptr_obj) {
-    .Call(`_outliertree_deserialize_OutlierTree`, src, ptr_obj)
-}
-
-check_null_ptr_model <- function(ptr_model) {
-    .Call(`_outliertree_check_null_ptr_model`, ptr_model)
-}
-
 fit_OutlierTree <- function(arr_num, ncols_numeric, arr_cat, ncols_categ, ncat, arr_ord, ncols_ord, ncat_ord, nrows, cols_ignore_r, nthreads, categ_as_bin, ord_as_bin, cat_bruteforce_subset, categ_from_maj, take_mid, max_depth, max_perc_outliers, min_size_numeric, min_size_categ, min_gain, follow_all, gain_as_pct, z_norm, z_outlier, return_outliers, cat_levels, ord_levels, colnames_num, colnames_cat, colnames_ord, min_date, min_ts) {
     .Call(`_outliertree_fit_OutlierTree`, arr_num, ncols_numeric, arr_cat, ncols_categ, ncat, arr_ord, ncols_ord, ncat_ord, nrows, cols_ignore_r, nthreads, categ_as_bin, ord_as_bin, cat_bruteforce_subset, categ_from_maj, take_mid, max_depth, max_perc_outliers, min_size_numeric, min_size_categ, min_gain, follow_all, gain_as_pct, z_norm, z_outlier, return_outliers, cat_levels, ord_levels, colnames_num, colnames_cat, colnames_ord, min_date, min_ts)
 }

diff --git a/R/helpers.R b/R/helpers.R
@@ -261,17  261,15 @@ check.nthreads <- function(nthreads) {
         nthreads <- 1L
     } else if (is.na(nthreads)) {
         nthreads <- 1L
-    } else if (nthreads == "auto") {
-        nthreads <- parallel::detectCores()
     } else if (nthreads < 1L) {
-        nthreads <- parallel::detectCores()
         nthreads <- 1L
     }
     nthreads <- as.integer(nthreads)
     if (nthreads > 1L && !R_has_openmp()) {
         msg <- paste0("Attempting to use more than 1 thread, but ",
                       "package was compiled without OpenMP support.")
         if (tolower(Sys.info()[["sysname"]]) == "darwin")
-            msg <- paste0(msg, " See https://mac.r-project.org/openmp/")
             msg <- paste0(msg, " See https://github.com/david-cortes/installing-optimized-libraries#4-macos-install-and-enable-openmp")
         warning(msg)
     }
     return(nthreads)
@@ -292,7  290,9 @@ check.is.model.obj <- function(model_obj) {
     if (!("outliertree" %in% class(model_obj))) {
         stop("Must pass an Outlier Tree model object as generated by function 'outlier.tree'.")
     }
-    if (is.null(model_obj$obj_from_cpp)) {
     if (is.null(model_obj$obj_from_cpp$ptr_model$ptr) ||
         identical(model_obj$obj_from_cpp$ptr_model$ptr, methods::new("externalptr"))
     ) {
         stop("Outlier Tree model object has been corrupted.")
     }
 }

diff --git a/R/obj_methods.R b/R/obj_methods.R
@@ -10,7  10,6 @@
 #' @return The same input `x` that was passed (as `invisible`).
 #' @export
 print.outliertree <- function(x, ...) {
-    unpack.outlier.tree(x)
     cat("Outlier Tree model\n")
     if (NROW(x$cols_num))  cat(sprintf("\tNumeric variables: %d\n",     NROW(x$cols_num)))
     if (NROW(x$cols_date)) cat(sprintf("\tDate variables: %d\n",        NROW(x$cols_date)))

diff --git a/R/outliertree.R b/R/outliertree.R
@@ -1,6  1,7 @@
 #' @importFrom parallel detectCores
 #' @importFrom stats predict
 #' @importFrom utils head
 #' @importFrom methods new
 #' @importFrom Rcpp evalCpp
 #' @useDynLib outliertree, .registration=TRUE
 NULL
@@ -113,7  114,7 @@ NULL
 #'   \item GritBot software: \url{https://www.rulequest.com/gritbot-info.html}
 #'   \item Cortes, David. "Explainable outlier detection through decision tree conditioning." arXiv preprint arXiv:2001.00636 (2020).
 #' }
-#' @seealso \link{predict.outliertree} \link{extract.training.outliers} \link{hypothyroid} \link{unpack.outlier.tree}
 #' @seealso \link{predict.outliertree} \link{extract.training.outliers} \link{hypothyroid}
 #' @examples 
 #' library(outliertree)
 #' 
@@ -139,7  140,7 @@ NULL
 #' ### use custom row names
 #' df.w.names <- hypothyroid
 #' row.names(df.w.names) <- paste0("rownum", 1:nrow(hypothyroid))
-#' outliers.w.names <- predict(model, df.w.names, return_outliers=TRUE)
 #' outliers.w.names <- predict(model, df.w.names, return_outliers=TRUE, nthreads=1)
 #' outliers.w.names[["rownum745"]]
 #' @export
 outlier.tree <- function(df, max_depth = 4L, min_gain = 1e-2, z_norm = 2.67, z_outlier = 8.0,
@@ -212,8  213,6 @@ outlier.tree <- function(df, max_depth = 4L, min_gain = 1e-2, z_norm = 2.67, z_o
                                                as.character(model_data$cols_ord),
                                                model_data$date_min,
                                                model_data$ts_min)
-    if (!NROW(model_data$obj_from_cpp$serialized_obj))
-        stop("Model object is too big. Try smaller inputs and/or changing hyperparameters.")
     names(model_data$obj_from_cpp$bounds) <- get.cols.ordered(model_data)
     model_data$obj_from_cpp$bounds        <- model_data$obj_from_cpp$bounds[names(df)]
 
@@ -252,6  251,7 @@ outlier.tree <- function(df, max_depth = 4L, min_gain = 1e-2, z_norm = 2.67, z_o
 #' outliers. The number of decimals will be dynamically increased according to the relative magnitudes of the
 #' values being reported. Ignored when passing `outliers_print=0` or `outliers_print=FALSE`.
 #' @param return_outliers Whether to return the outliers in an R object (otherwise will just print them).
 #' @param nthreads Number of parallel threads to use. Parallelization is done by rows.
 #' @param ... Not used.
 #' @return If passing `return_outliers` = `TRUE`, will return a list of lists with the outliers and their
 #' information (each row is an entry in the first list, with the same names as the rows in the input data
@@ -264,7  264,7 @@ outlier.tree <- function(df, max_depth = 4L, min_gain = 1e-2, z_norm = 2.67, z_o
 #' @details Note that after loading a serialized object from `outlier.tree` through `readRDS` or `load`,
 #' it will only de-serialize the underlying C   object upon running `predict` or `print`, so the first run will
 #' be slower, while subsequent runs will be faster as the C   object will already be in-memory.
-#' @seealso \link{outlier.tree} \link{print.outlieroutputs} \link{unpack.outlier.tree}
 #' @seealso \link{outlier.tree} \link{print.outlieroutputs}
 #' @examples 
 #' library(outliertree)
 #' ### random data frame with an obvious outlier
@@ -291,7  291,8 @@ outlier.tree <- function(df, max_depth = 4L, min_gain = 1e-2, z_norm = 2.67, z_o
 #' 
 #' ### find the test outlier
 #' test_outliers = predict(outliers_model, df_test,
-#'     outliers_print = 1, return_outliers = TRUE)
 #'     outliers_print = 1, return_outliers = TRUE,
 #'     nthreads = 1)
 #' 
 #' ### retrieve the outlier info (for row 1) as an R list
 #' test_outliers[[1]]
@@ -300,8  301,7 @@ outlier.tree <- function(df, max_depth = 4L, min_gain = 1e-2, z_norm = 2.67, z_o
 #' # dt = t(data.table::as.data.table(test_outliers))
 #' @export 
 predict.outliertree <- function(object, newdata, outliers_print = 15L, min_decimals = 2L,
-                                return_outliers = TRUE, ...) {
-    unpack.outlier.tree(object)
                                 return_outliers = TRUE, nthreads = object$nthreads, ...) {
     outliers_print  <- check.outliers.print(outliers_print)
     return_outliers <- as.logical(return_outliers)
     if (NROW(newdata) == 0) {
@@ -314,7  314,7 @@ predict.outliertree <- function(object, newdata, outliers_print = 15L, min_decim
     }
 
     c_arr_data    <- split.types.new(newdata, object)
-    outliers_info <- predict_OutlierTree(object$obj_from_cpp$ptr_model, NROW(newdata), object$nthreads,
     outliers_info <- predict_OutlierTree(object$obj_from_cpp$ptr_model$ptr, NROW(newdata), check.nthreads(nthreads),
                                          c_arr_data$arr_num, c_arr_data$arr_cat, c_arr_data$arr_ord,
                                          object$cat_levels,
                                          object$ord_levels,
@@ -365,7  365,8 @@ predict.outliertree <- function(object, newdata, outliers_print = 15L, min_decim
 #' pred <- predict(otree,
 #'   hypothyroid,
 #'   outliers_print=0,
-#'   return_outliers=TRUE)
 #'   return_outliers=TRUE,
 #'   nthreads=1)
 #'   
 #' ### Print stored predictions
 #' ### Row 531 is an outlier, but 532 is not
@@ -442,52  443,3 @@ check.outlierness.bounds <- function(outlier_tree_model) {
     check.is.model.obj(outlier_tree_model)
     return(outlier_tree_model$obj_from_cpp$bounds)
 }
-
-#' @title Unpack Outlier Tree model after de-serializing
-#' @description  After persisting an outlier tree model object through `saveRDS`, `save`,
-#' or restarting a session, the underlying C   objects that constitute the outlier tree
-#' model and which live only on the C   heap memory are not saved along, thus not
-#' restored after loading a saved model through `readRDS` or `load`.
-#' 
-#' The model object however keeps serialized versions of the C   objects as raw bytes,
-#' from which the C   objects can be reconstructed, and are done so automatically after
-#' calling `predict`, `print`, or `summary` on the freshly-loaded object from
-#' `readRDS` or `load`.
-#' 
-#' This function allows de-serializing the object bytes without invoking any extra
-#' side effects or computations, akin to XGBoost's `xgb.Booster.complete` or
-#' CatBoost's `catboost.restore_handle`.
-#' @details If the model is going to be used in a production system, it's possible
-#' after de-serialization to delete the raw bytes in order to save memory (e.g.
-#' `otree$obj_from_cpp$serialized_obj <- NULL`). The memory will however not be
-#' freed automatically, as it's managed by R's garbage collector.
-#' @param model An Outlier Tree object as returned by `outlier.tree`, which has
-#' been just loaded from a disk file through `readRDS`, `load`, or a session restart.
-#' @return No return value. Object is modified in-place.
-#' @examples 
-#' ### Warning: this example will generate a temporary .Rds
-#' ### file in your temp folder, and will then delete it
-#' library(outliertree)
-#' set.seed(1)
-#' df <- as.data.frame(matrix(rnorm(1000), nrow = 250))
-#' otree <- outlier.tree(df, outliers_print=0, nthreads=1)
-#' temp_file <- file.path(tempdir(), "otree.Rds")
-#' saveRDS(otree, temp_file)
-#' otree2 <- readRDS(temp_file)
-#' file.remove(temp_file)
-#' 
-#' cat("Pointer after loading model is this: \n")
-#' print(otree2$obj_from_cpp$ptr_model)
-#' 
-#' ### now unpack the raw bytes
-#' unpack.outlier.tree(otree2)
-#' cat("Pointer after unpacking is this: \n")
-#' print(otree2$obj_from_cpp$ptr_model)
-#' @export
-unpack.outlier.tree <- function(model)  {
-    check.is.model.obj(model)
-    if (check_null_ptr_model(model$obj_from_cpp$ptr_model)) {
-        deserialize_OutlierTree(model$obj_from_cpp$serialized_obj, model$obj_from_cpp$ptr_model)
-    }
-    return(invisible(NULL))
-}
diff --git a/build/vignette.rds b/build/vignette.rds
diff --git a/inst/doc/Explainable_Outlier_Detection_in_Titanic_dataset.html b/inst/doc/Explainable_Outlier_Detection_in_Titanic_dataset.html