-
Notifications
You must be signed in to change notification settings - Fork 46
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
added the files required for R bindings. (#182)
* added the files required for R bindings. * added seed parameter. Seed isn't working, but parameter can be set. * removed seed check until seeding is fixed. * update install instructions and travis. * Add proper roxygen comment to export forestPackingRConversion to NAMESPACE file. Update documentation and tests. * update Makevars flags. * update docs * update docs * Update install instructions update fpRerF example and forest-types * add check for correct forestType * update test and yaml * fix travis command error. * update tests for R-Dev. * should fix travis??? * add flag to stop test on failure. * update test-predict * 🚧 omp flags for mac os with clang. * 🚧 this possibly breaks -ffast-math on systems other than JLP's * update travis to run `R CMD build` and `R CMD check` * update README install for Mac OS. * update to README * set warnings_are_errors <- FALSE this will allow merging of the Rbindings with openMP support. * update travis. * Revert "update test-predict" This reverts commit 6ff5e88. * trying to revert to a state where TravisCI can install. * remove from Makevars.
- Loading branch information
Showing
23 changed files
with
687 additions
and
67 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,33 @@ | ||
| #' Packs a forest and saves modified forest to disk for use by PackPredict function | ||
| #' | ||
| #' Efficiently packs a forest trained with the RF option. Two intermediate data structures are written to disk, forestPackTempFile.csv and traversalPackTempFile.csv. The size of these data structures is proportional to a trained forest and training data respectively. Both data structures are removed at the end of the operation. The resulting forest is saved as forest.out. The size of this file is similar to the size of the trained forest. | ||
| #' | ||
| #' @param forest forest data structure returned from fpRerF function. | ||
| #' @param X an n by d numeric matrix (preferable) or data frame used to train the forest. | ||
| #' | ||
| #' | ||
| #' @export | ||
| #' | ||
| #' @examples | ||
| #' | ||
| #' library(rerf) | ||
| #' forest <- fpRerF(as.matrix(iris[, 1:4]), as.numeric(iris[[5L]])-1) | ||
| #' predictions <- fpPredict(forest, as.matrix(iris[, 1:4])) | ||
| # | ||
| fpPredict <- | ||
| function(forest=NULL,X=NULL){ | ||
| if(class(forest)!="Rcpp_forestPackingRConversion"){ | ||
| stop("forest is not of correct type.") | ||
| } | ||
|
|
||
| if(is.null(X)){ | ||
| stop("no observations to predict.") | ||
| } | ||
|
|
||
| X <- as.matrix(X) | ||
| if(class(X[1,1])=="integer"){ | ||
| storage.mode(X) <- "numeric" | ||
| } | ||
|
|
||
| forest$predict(X) | ||
| } |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,120 @@ | ||
| #' Packs a forest and saves modified forest to disk for use by PackPredict function | ||
| #' | ||
| #' Efficiently packs a forest trained with the RF option. Two intermediate data structures are written to disk, forestPackTempFile.csv and traversalPackTempFile.csv. The size of these data structures is proportional to a trained forest and training data respectively. Both data structures are removed at the end of the operation. The resulting forest is saved as forest.out. The size of this file is similar to the size of the trained forest. | ||
| #' | ||
| #' @param X an n by d numeric matrix (preferable) or data frame used to train the forest. | ||
| #' @param Y a numeric vector of size n. If the Y vector used to train the forest was not of type numeric then a simple call to as.numeric(Y) will suffice as input. | ||
| #' @param csvFileName the name of a headerless csv file containing combined data and labels. | ||
| #' @param columnWithY is the column in the headerless csv file containing class lables. | ||
| #' @param maxDepth int the maximum allowed tree height/depth (path distance between root and leaves). (maxDepth = Inf, i.e. largest system int) | ||
| #' @param minParent is the size of nodes that will not be split (minParent=1) | ||
| #' @param numTreesInForest the number of trees to grow in the forest (numTreesInForest=500) | ||
| #' @param numCores is the number of cores to use when training and predicting with the forest (numCores=1) | ||
| #' @param numTreeBins the number of bins to store the forest. Each bin will contain numTreesInForest/numTreeBins trees. Only used when forestType=="binned*" (numTreeBins= numCores) | ||
| #' @param forestType the type of forest to grow: binnedBase, binnedBaseRerF (forestType="binnedBaseRerF") | ||
| #' @param NodeSizeToBin the minimum node size to use stratified subsampling (NodeSizeToBin=NULL) | ||
| #' @param NodeSizeBin the size of the stratified subsample chosen when NodeSizeToBin criteria is met (NodeSizeBin=NULL) | ||
| #' @param forestType the type of forest to grow: binnedBase, binnedBaseRerF, rfBase, rerf (forestType="binnedBaseRerF") | ||
| #' @param mtry the number of features to consider when splitting a node (mtry=ncol(X)^.5) | ||
| #' @param mtryMult the average number of features combined to form a new feature when using RerF (mtryMult=1) | ||
| #' | ||
| #' | ||
| #' @export | ||
| #' | ||
| #' @examples | ||
| #' library(rerf) | ||
| #' ## setup data | ||
| #' X <- as.matrix(iris[, 1:4]) | ||
| #' Y <- as.numeric(iris[[5]]) - 1 | ||
| #' forest <- fpRerF(X, Y, numCores = 2L) | ||
| #' (training.error <- mean(fpPredict(forest, X) != Y)) | ||
| #' | ||
|
|
||
|
|
||
| fpRerF <- | ||
| function(X=NULL, Y=NULL,csvFileName=NULL, columnWithY=NULL, maxDepth = Inf, minParent=1, numTreesInForest=500, numCores=1,numTreeBins=NULL, forestType="binnedBaseRerF", nodeSizeToBin=NULL, nodeSizeBin=NULL,mtry=NULL, mtryMult=NULL,seed=sample(1:1000000,1)){ | ||
|
|
||
| ##### Basic Checks | ||
| ################################################ | ||
| if(numCores < 1){ | ||
| stop("at least one core must be used.") | ||
| } | ||
| if(minParent < 1){ | ||
| stop("at least one observation must be used in each node.") | ||
| } | ||
| if(numTreesInForest < 1){ | ||
| stop("at least one tree must be used.") | ||
| } | ||
| if(!(forestType %in% c('rfBase', 'rerf', 'binnedBase', 'binnedBaseRerF'))){ | ||
| stop("must pick a forest type from the following:\n rfBase, rerf, inPlace, inPlaceRerF, binnedBase, binnedBaseRerF") | ||
| } | ||
| if(is.null(numTreeBins)){ | ||
| numTreeBins <- numCores | ||
| } | ||
|
|
||
| forest_module <- methods::new(forestPackingRConversion) | ||
| forest_module$setParameterString("forestType", forestType) | ||
| forest_module$setParameterInt("numTreesInForest", numTreesInForest) | ||
| if(is.finite(maxDepth) && (maxDepth > 0)){ | ||
| forest_module$setParameterInt("maxDepth", maxDepth) | ||
| } | ||
| forest_module$setParameterInt("minParent", minParent) | ||
| forest_module$setParameterInt("numCores", numCores) | ||
| forest_module$setParameterInt("useRowMajor",0) | ||
| forest_module$setParameterInt("seed",seed) | ||
|
|
||
| if(!is.null(nodeSizeToBin) & !is.null(nodeSizeBin)){ | ||
| if(nodeSizeBin > nodeSizeToBin){ | ||
| stop("nodeSizeBin must be less than or greater than nodeSizeToBin.") | ||
| } | ||
| forest_module$setParameterInt("binSize",nodeSizeBin) | ||
| forest_module$setParameterInt("binMin",nodeSizeToBin) | ||
| } | ||
|
|
||
| ### Set MTRY if not NULL | ||
| if(!is.null(mtry)){ | ||
| forest_module$setParameterInt("mtry",mtry) | ||
| } | ||
| if(!is.null(mtryMult)){ | ||
| forest_module$setParameterDouble("mtryMult",mtryMult) | ||
| } | ||
|
|
||
| ##### Check X and Y inputs | ||
| ################################################ | ||
| if(xor(is.null(X), is.null(Y))){ | ||
| stop("X and Y must be set or both must be blank.") | ||
| } | ||
|
|
||
| if(!is.null(X)){ | ||
| X <- as.matrix(X) | ||
| if(class(X[1,1])=="integer"){ | ||
| storage.mode(X) <- "numeric" | ||
| } | ||
|
|
||
| Y <- as.integer(Y) | ||
|
|
||
| if(nrow(X) != length(Y)){ | ||
| stop("number of observations in X is different from Y length.") | ||
| } | ||
| forest_module$growForestGivenX(X,Y) | ||
|
|
||
|
|
||
| ##### Check CSV info | ||
| ################################################ | ||
| }else if(!is.null(csvFileName)){ | ||
| if(!file.exists(csvFileName)){ | ||
| stop("file does not exist.") | ||
| } | ||
| if(is.null(columnWithY)){ | ||
| stop("columnWithY cannot be NULL when using CSV.") | ||
| } | ||
| forest_module$setParameterString("CSVFileName", csvFileName) | ||
| forest_module$setParameterInt("columnWithY", columnWithY); | ||
| forest_module$growForestCSV() | ||
| }else{ | ||
| stop("no input provided.") | ||
| } | ||
| #forest_module$printParameters() | ||
| return(forest_module) | ||
|
|
||
| } |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,4 @@ | ||
| #' Rcpp module: forestPackingRConversion | ||
| #' @name forestPackingRConversion | ||
| #' @export | ||
| loadModule("forestPackingRConversion_mod", TRUE) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.