X-Git-Url: https://git.auder.net/?a=blobdiff_plain;f=epclust%2FR%2Fmain.R;h=09e1ed7feca546b476b3e69d8b9e379e535a967d;hb=14c10f2d252f45349e0b4fbf87e17dfbfae39f92;hp=00d2a882e8745950599d839e953db56ef4d746fd;hpb=282342bafdc9ff65c5df98c6e2304d63b33b9fb2;p=epclust.git diff --git a/epclust/R/main.R b/epclust/R/main.R index 00d2a88..09e1ed7 100644 --- a/epclust/R/main.R +++ b/epclust/R/main.R @@ -11,30 +11,31 @@ #' \item Divide series into \code{ntasks} groups to process in parallel. In each task: #' \enumerate{ #' \item iterate the first clustering algorithm on its aggregated outputs, -#' on inputs of size \code{nb_series_per_chunk} -#' \item optionally, if WER=="mix": -#' a) compute the K1 synchrones curves, -#' a) compute WER distances (K1xK1 matrix) between medoids and -#' b) apply the second clustering algorithm (output: K2 indices) +#' on inputs of size \code{nb_items_clust}\cr +#' -> K1 medoids indices +#' \item optionally, if WER=="mix":\cr +#' a. compute WER distances (K1xK1) between medoids\cr +#' b. apply the 2nd clustering algorithm\cr +#' -> K2 medoids indices #' } #' \item Launch a final task on the aggregated outputs of all previous tasks: #' ntasks*K1 if WER=="end", ntasks*K2 otherwise #' \item Compute synchrones (sum of series within each final group) #' } -#' \cr +#' #' The main argument -- \code{series} -- has a quite misleading name, since it can be #' either a [big.]matrix, a CSV file, a connection or a user function to retrieve series. -#' When \code{series} is given as a function, it must take a single argument, -#' 'indices', integer vector equal to the indices of the curves to retrieve; +#' When \code{series} is given as a function it must take a single argument, +#' 'indices': integer vector equal to the indices of the curves to retrieve; #' see SQLite example. #' WARNING: the return value must be a matrix (in columns), or NULL if no matches. -#' \cr +#' #' Note: Since we don't make assumptions on initial data, there is a possibility that #' even when serialized, contributions do not fit in RAM. For example, #' 30e6 series of length 100,000 would lead to a +4Go contribution matrix. Therefore, #' it's safer to place these in (binary) files; that's what we do. #' -#' @param series Access to the (time-)series, which can be of one of the three +#' @param series Access to the N (time-)series, which can be of one of the four #' following types: #' \itemize{ #' \item [big.]matrix: each column contains the (time-ordered) values of one time-serie @@ -43,41 +44,38 @@ #' \item function: a custom way to retrieve the curves; it has only one argument: #' the indices of the series to be retrieved. See SQLite example #' } -#' @param K1 Number of clusters to be found after stage 1 (K1 << N [number of series]) +#' @param K1 Number of clusters to be found after stage 1 (K1 << N) #' @param K2 Number of clusters to be found after stage 2 (K2 << K1) -#' @param nb_series_per_chunk (Maximum) number of series to retrieve in one batch -#' @param nb_items_clust (~Maximum) number of items in clustering algorithm 1 input +#' @param nb_series_per_chunk Number of series to retrieve in one batch +#' @param nb_items_clust Number of items in 1st clustering algorithm input #' @param algoClust1 Clustering algorithm for stage 1. A function which takes (data, K) #' as argument where data is a matrix in columns and K the desired number of clusters, -#' and outputs K medoids ranks. Default: PAM. In our method, this function is called -#' on iterated medoids during stage 1 +#' and outputs K medoids ranks. Default: PAM. #' @param algoClust2 Clustering algorithm for stage 2. A function which takes (dists, K) #' as argument where dists is a matrix of distances and K the desired number of clusters, -#' and outputs K medoids ranks. Default: PAM. In our method, this function is called -#' on a matrix of K1 x K1 (WER) distances computed between medoids after algorithm 1 +#' and outputs K medoids ranks. Default: PAM. #' @param wav_filt Wavelet transform filter; see ?wavelets::wt.filter #' @param contrib_type Type of contribution: "relative", "logit" or "absolute" (any prefix) #' @param WER "end" to apply stage 2 after stage 1 has fully iterated, or "mix" to apply #' stage 2 at the end of each task -#' @param smooth_lvl Smoothing level: odd integer, 1 == no smoothing. 3 seems good +#' @param smooth_lvl Smoothing level: odd integer, 1 == no smoothing. #' @param nvoice Number of voices within each octave for CWT computations #' @param random TRUE (default) for random chunks repartition #' @param ntasks Number of tasks (parallel iterations to obtain K1 [if WER=="end"] -#' or K2 [if WER=="mix"] medoids); default: 1. +#' or K2 [if WER=="mix"] medoids); default: 1.\cr #' Note: ntasks << N (number of series), so that N is "roughly divisible" by ntasks -#' @param ncores_tasks Number of parallel tasks (1 to disable: sequential tasks) -#' @param ncores_clust Number of parallel clusterings in one task (3 should be a minimum) +#' @param ncores_tasks Number of parallel tasks ('1' == sequential tasks) +#' @param ncores_clust Number of parallel clusterings in one task #' @param sep Separator in CSV input file (if any provided) -#' @param nbytes Number of bytes to serialize a floating-point number; 4 or 8 -#' @param endian Endianness for (de)serialization ("little" or "big") -#' @param verbose Level of verbosity (0/FALSE for nothing or 1/TRUE for all; devel stage) -#' @param parll TRUE to fully parallelize; otherwise run sequentially (debug, comparison) +#' @param nbytes 4 or 8 bytes to (de)serialize a floating-point number +#' @param endian Endianness for (de)serialization: "little" or "big" +#' @param verbose FALSE: nothing printed; TRUE: some execution traces #' -#' @return A list with +#' @return A list: #' \itemize{ -#' medoids: a matrix of the final K2 medoids curves, in columns -#' ranks: corresponding indices in the dataset -#' synchrones: a matrix of the K2 sum of series within each final group +#' \item medoids: matrix of the final K2 medoids curves +#' \item ranks: corresponding indices in the dataset +#' \item synchrones: sum of series within each final group #' } #' #' @references Clustering functional data using Wavelets [2013]; @@ -88,29 +86,34 @@ #' @examples #' \dontrun{ #' # WER distances computations are too long for CRAN (for now) +#' # Note: on this small example, sequential run is faster #' #' # Random series around cos(x,2x,3x)/sin(x,2x,3x) -#' x <- seq(0,500,0.05) -#' L <- length(x) #10001 +#' x <- seq(0,50,0.05) +#' L <- length(x) #1001 #' ref_series <- matrix( c(cos(x),cos(2*x),cos(3*x),sin(x),sin(2*x),sin(3*x)), ncol=6 ) #' library(wmtsa) #' series <- do.call( cbind, lapply( 1:6, function(i) -#' do.call(cbind, wmtsa::wavBootstrap(ref_series[,i], n.realization=400)) ) ) -#' #dim(series) #c(2400,10001) -#' res_ascii <- claws(series, K1=60, K2=6, 200, verbose=TRUE) +#' do.call(cbind, wmtsa::wavBootstrap(ref_series[,i], n.realization=40)) ) ) +#' # Mix series so that all groups are evenly spread +#' permut <- (0:239)%%6 * 40 + (0:239)%/%6 + 1 +#' series = series[,permut] +#' #dim(series) #c(240,1001) +#' res_ascii <- claws(series, K1=30, K2=6, nb_series_per_chunk=500, +#' nb_items_clust=100, random=FALSE, verbose=TRUE, ncores_clust=1) #' #' # Same example, from CSV file -#' csv_file <- "/tmp/epclust_series.csv" -#' write.table(series, csv_file, sep=",", row.names=FALSE, col.names=FALSE) -#' res_csv <- claws(csv_file, K1=60, K2=6, 200) +#' csv_file <- tempfile(pattern="epclust_series.csv_") +#' write.table(t(series), csv_file, sep=",", row.names=FALSE, col.names=FALSE) +#' res_csv <- claws(csv_file, 30, 6, 500, 100, random=FALSE, ncores_clust=1) #' #' # Same example, from binary file -#' bin_file <- "/tmp/epclust_series.bin" +#' bin_file <- tempfile(pattern="epclust_series.bin_") #' nbytes <- 8 #' endian <- "little" -#' binarize(csv_file, bin_file, 500, nbytes, endian) +#' binarize(csv_file, bin_file, 500, ",", nbytes, endian) #' getSeries <- function(indices) getDataInFile(indices, bin_file, nbytes, endian) -#' res_bin <- claws(getSeries, K1=60, K2=6, 200) +#' res_bin <- claws(getSeries, 30, 6, 500, 100, random=FALSE, ncores_clust=1) #' unlink(csv_file) #' unlink(bin_file) #' @@ -118,45 +121,46 @@ #' library(DBI) #' series_db <- dbConnect(RSQLite::SQLite(), "file::memory:") #' # Prepare data.frame in DB-format -#' n <- nrow(series) -#' time_values <- data.frame( -#' id <- rep(1:n,each=L), -#' time <- rep( as.POSIXct(1800*(0:n),"GMT",origin="2001-01-01"), L ), -#' value <- as.double(t(series)) ) +#' n <- ncol(series) +#' times_values <- data.frame( +#' id = rep(1:n,each=L), +#' time = rep( as.POSIXct(1800*(1:L),"GMT",origin="2001-01-01"), n ), +#' value = as.double(series) ) #' dbWriteTable(series_db, "times_values", times_values) #' # Fill associative array, map index to identifier #' indexToID_inDB <- as.character( -#' dbGetQuery(series_db, 'SELECT DISTINCT id FROM time_values')[,"id"] ) +#' dbGetQuery(series_db, 'SELECT DISTINCT id FROM times_values')[,"id"] ) #' serie_length <- as.integer( dbGetQuery(series_db, -#' paste("SELECT COUNT * FROM time_values WHERE id == ",indexToID_inDB[1],sep="")) ) +#' paste("SELECT COUNT(*) FROM times_values WHERE id == ",indexToID_inDB[1],sep="")) ) #' getSeries <- function(indices) { +#' indices = indices[ indices <= length(indexToID_inDB) ] +#' if (length(indices) == 0) +#' return (NULL) #' request <- "SELECT id,value FROM times_values WHERE id in (" -#' for (i in indices) -#' request <- paste(request, indexToID_inDB[i], ",", sep="") +#' for (i in seq_along(indices)) { +#' request <- paste(request, indexToID_inDB[ indices[i] ], sep="") +#' if (i < length(indices)) +#' request <- paste(request, ",", sep="") +#' } #' request <- paste(request, ")", sep="") #' df_series <- dbGetQuery(series_db, request) -#' if (length(df_series) >= 1) -#' as.matrix(df_series[,"value"], nrow=serie_length) -#' else -#' NULL +#' matrix(df_series[,"value"], nrow=serie_length) #' } -#' res_db <- claws(getSeries, K1=60, K2=6, 200)) +#' res_db <- claws(getSeries, 30, 6, 500, 100, random=FALSE, ncores_clust=1) #' dbDisconnect(series_db) #' -#' # All results should be the same: -#' library(digest) -#' digest::sha1(res_ascii) -#' digest::sha1(res_csv) -#' digest::sha1(res_bin) -#' digest::sha1(res_db) +#' # All results should be equal: +#' all(res_ascii$ranks == res_csv$ranks +#' & res_ascii$ranks == res_bin$ranks +#' & res_ascii$ranks == res_db$ranks) #' } #' @export -claws <- function(series, K1, K2, nb_series_per_chunk, nb_items_clust=7*K1, +claws <- function(series, K1, K2, nb_series_per_chunk, nb_items_clust=5*K1, algoClust1=function(data,K) cluster::pam(t(data),K,diss=FALSE,pamonce=1)$id.med, algoClust2=function(dists,K) cluster::pam(dists,K,diss=TRUE,pamonce=1)$id.med, wav_filt="d8", contrib_type="absolute", WER="end", smooth_lvl=3, nvoice=4, random=TRUE, ntasks=1, ncores_tasks=1, ncores_clust=3, sep=",", nbytes=4, - endian=.Platform$endian, verbose=FALSE, parll=TRUE) + endian=.Platform$endian, verbose=FALSE) { # Check/transform arguments if (!is.matrix(series) && !bigmemory::is.big.matrix(series) @@ -186,7 +190,6 @@ claws <- function(series, K1, K2, nb_series_per_chunk, nb_items_clust=7*K1, stop("'sep': character") nbytes <- .toInteger(nbytes, function(x) x==4 || x==8) verbose <- .toLogical(verbose) - parll <- .toLogical(parll) # Binarize series if it is not a function; the aim is to always use a function, # to uniformize treatments. An equally good alternative would be to use a file-backed @@ -205,8 +208,6 @@ claws <- function(series, K1, K2, nb_series_per_chunk, nb_items_clust=7*K1, # Serialize all computed wavelets contributions into a file contribs_file <- ".contribs.epclust.bin" - index <- 1 - nb_curves <- 0 if (verbose) cat("...Compute contributions and serialize them (or retrieve past binary file)\n") if (!file.exists(contribs_file)) @@ -240,12 +241,17 @@ claws <- function(series, K1, K2, nb_series_per_chunk, nb_items_clust=7*K1, indices_all[((i-1)*nb_series_per_task+1):upper_bound] }) + parll <- (ncores_tasks > 1) if (parll && ntasks>1) { # Initialize parallel runs: outfile="" allow to output verbose traces in the console # under Linux. All necessary variables are passed to the workers. - cl <- parallel::makeCluster(ncores_tasks, outfile="") - varlist <- c("ncores_clust","verbose","parll", #task 1 & 2 + cl <- + if (verbose) + parallel::makeCluster(ncores_tasks, outfile="") + else + parallel::makeCluster(ncores_tasks) + varlist <- c("ncores_clust","verbose", #task 1 & 2 "K1","getContribs","algoClust1","nb_items_clust") #task 1 if (WER=="mix") { @@ -267,11 +273,11 @@ claws <- function(series, K1, K2, nb_series_per_chunk, nb_items_clust=7*K1, if (parll && ntasks>1) require("epclust", quietly=TRUE) indices_medoids <- clusteringTask1(inds, getContribs, K1, algoClust1, - nb_items_clust, ncores_clust, verbose, parll) + nb_items_clust, ncores_clust, verbose) if (WER=="mix") { indices_medoids <- clusteringTask2(indices_medoids, getSeries, K2, algoClust2, - nb_series_per_chunk,smooth_lvl,nvoice,nbytes,endian,ncores_clust,verbose,parll) + nb_series_per_chunk,smooth_lvl,nvoice,nbytes,endian,ncores_clust,verbose) } indices_medoids } @@ -285,7 +291,7 @@ claws <- function(series, K1, K2, nb_series_per_chunk, nb_items_clust=7*K1, } # As explained above, we obtain after all runs ntasks*[K1 or K2] medoids indices, - # depending wether WER=="end" or "mix", respectively. + # depending whether WER=="end" or "mix", respectively. indices_medoids_all <- if (parll && ntasks>1) unlist( parallel::parLapply(cl, indices_tasks, runTwoStepClustering) ) @@ -306,15 +312,16 @@ claws <- function(series, K1, K2, nb_series_per_chunk, nb_items_clust=7*K1, if (verbose) cat("...Run final // stage 1 + stage 2\n") indices_medoids <- clusteringTask1(indices_medoids_all, getContribs, K1, algoClust1, - nb_items_clust, ncores_tasks*ncores_clust, verbose, parll) - indices_medoids <- clusteringTask2(indices_medoids, getContribs, K2, algoClust2, - nb_series_per_chunk,smooth_lvl,nvoice,nbytes,endian,ncores_last_stage,verbose,parll) + nb_items_clust, ncores_tasks*ncores_clust, verbose) + + indices_medoids <- clusteringTask2(indices_medoids, getSeries, K2, algoClust2, + nb_series_per_chunk,smooth_lvl,nvoice,nbytes,endian,ncores_last_stage,verbose) # Compute synchrones, that is to say the cumulated power consumptions for each of the K2 # final groups. medoids <- getSeries(indices_medoids) synchrones <- computeSynchrones(medoids, getSeries, nb_curves, nb_series_per_chunk, - ncores_last_stage, verbose, parll) + ncores_last_stage, verbose) # NOTE: no need to use big.matrix here, since there are only K2 << K1 << N remaining curves list("medoids"=medoids, "ranks"=indices_medoids, "synchrones"=synchrones)