X-Git-Url: https://git.auder.net/?a=blobdiff_plain;f=epclust%2FR%2Fmain.R;h=bcc650abb8feee40cfc246ef3f95d27b13e478c4;hb=d9bb53c5e1392018bf67f92140edb10137f3423c;hp=86dac64c150c6d1a6c1b4f8bf8756030ca89ae5d;hpb=eef6f6c97277ea3ce760981e5244cbde7fc904a0;p=epclust.git diff --git a/epclust/R/main.R b/epclust/R/main.R index 86dac64..bcc650a 100644 --- a/epclust/R/main.R +++ b/epclust/R/main.R @@ -4,29 +4,35 @@ #' two stage procedure in parallel (see details). #' Input series must be sampled on the same time grid, no missing values. #' -#' @details Summary of the function execution flow: +#' Summary of the function execution flow: +#' \enumerate{ +#' \item Compute and serialize all contributions, obtained through discrete wavelet +#' decomposition (see Antoniadis & al. [2013]) +#' \item Divide series into \code{ntasks} groups to process in parallel. In each task: #' \enumerate{ -#' \item Compute and serialize all contributions, obtained through discrete wavelet -#' decomposition (see Antoniadis & al. [2013]) -#' \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_items_clust} -#' \item optionally, if WER=="mix": -#' a) compute the K1 synchrones curves, -#' b) compute WER distances (K1xK1 matrix) between synchrones and -#' c) apply the second clustering algorithm -#' } -#' \item Launch a final task on the aggregated outputs of all previous tasks: -#' in the case WER=="end" this task takes indices in input, otherwise -#' (medoid) curves +#' \item iterate the first clustering algorithm on its aggregated outputs, +#' on inputs of size \code{nb_items_clust1} +#' \item optionally, if WER=="mix": +#' a) compute the K1 synchrones curves, +#' b) compute WER distances (K1xK1 matrix) between synchrones and +#' c) apply the second clustering algorithm #' } -#' The main argument -- \code{getSeries} -- 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; the name was chosen because all types of arguments are converted to a function. -#' When \code{getSeries} 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. The nature and role of other arguments should be clear +#' \item Launch a final task on the aggregated outputs of all previous tasks: +#' in the case WER=="end" this task takes indices in input, otherwise +#' (medoid) curves +#' } +#' \cr +#' The main argument -- \code{getSeries} -- 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; the name was chosen because all types of arguments are converted to a function. +#' When \code{getSeries} 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. The nature and role of other arguments should be clear +#' \cr +#' Note: Since we don't make assumptions on initial data, there is a possibility that +#' even when serialized, contributions or synchrones 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 getSeries Access to the (time-)series, which can be of one of the three #' following types: @@ -39,10 +45,18 @@ #' } #' @param K1 Number of clusters to be found after stage 1 (K1 << N [number of series]) #' @param K2 Number of clusters to be found after stage 2 (K2 << K1) -#' @param nb_per_chunk (Maximum) number of items to retrieve in one batch, for both types of -#' retrieval: resp. series and contribution; in a vector of size 2 -#' @param nb_items_clust1 (Maximum) number of items in input of the clustering algorithm -#' for stage 1 +#' @param nb_series_per_chunk (Maximum) number of series to retrieve in one batch +#' @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 +#' @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 synchrones +#' @param nb_items_clust1 (~Maximum) number of items in input of the clustering algorithm +#' for stage 1. At worst, a clustering algorithm might be called with ~2*nb_items_clust1 +#' items; but this could only happen at the last few iterations. #' @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 @@ -78,12 +92,12 @@ #' 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) -#' medoids_ascii = claws(series, K1=60, K2=6, nb_per_chunk=c(200,500), verbose=TRUE) +#' medoids_ascii = claws(series, K1=60, K2=6, 200, verbose=TRUE) #' #' # Same example, from CSV file #' csv_file = "/tmp/epclust_series.csv" #' write.table(series, csv_file, sep=",", row.names=FALSE, col.names=FALSE) -#' medoids_csv = claws(csv_file, K1=60, K2=6, nb_per_chunk=c(200,500)) +#' medoids_csv = claws(csv_file, K1=60, K2=6, 200) #' #' # Same example, from binary file #' bin_file <- "/tmp/epclust_series.bin" @@ -91,7 +105,7 @@ #' endian <- "little" #' binarize(csv_file, bin_file, 500, nbytes, endian) #' getSeries <- function(indices) getDataInFile(indices, bin_file, nbytes, endian) -#' medoids_bin <- claws(getSeries, K1=60, K2=6, nb_per_chunk=c(200,500)) +#' medoids_bin <- claws(getSeries, K1=60, K2=6, 200) #' unlink(csv_file) #' unlink(bin_file) #' @@ -118,7 +132,7 @@ #' df_series <- dbGetQuery(series_db, request) #' as.matrix(df_series[,"value"], nrow=serie_length) #' } -#' medoids_db = claws(getSeries, K1=60, K2=6, nb_per_chunk=c(200,500)) +#' medoids_db = claws(getSeries, K1=60, K2=6, 200)) #' dbDisconnect(series_db) #' #' # All computed medoids should be the same: @@ -128,15 +142,12 @@ #' digest::sha1(medoids_db) #' } #' @export -claws <- function(getSeries, K1, K2, - nb_per_chunk,nb_items_clust1=7*K1 #volumes of data - wav_filt="d8",contrib_type="absolute", #stage 1 - WER="end", #stage 2 - random=TRUE, #randomize series order? - ntasks=1, ncores_tasks=1, ncores_clust=4, #parallelism - sep=",", #ASCII input separator - nbytes=4, endian=.Platform$endian, #serialization (write,read) - verbose=FALSE, parll=TRUE) +claws <- function(getSeries, K1, K2, nb_series_per_chunk, nb_items_clust1=7*K1, + algoClust1=function(data,K) cluster::pam(t(data),K,diss=FALSE)$id.med, + algoClust2=function(dists,K) cluster::pam(dists,K,diss=TRUE)$id.med, + wav_filt="d8", contrib_type="absolute", WER="end", random=TRUE, + ntasks=1, ncores_tasks=1, ncores_clust=4, sep=",", nbytes=4, + endian=.Platform$endian, verbose=FALSE, parll=TRUE) { # Check/transform arguments if (!is.matrix(getSeries) && !bigmemory::is.big.matrix(getSeries) @@ -147,19 +158,15 @@ claws <- function(getSeries, K1, K2, } K1 <- .toInteger(K1, function(x) x>=2) K2 <- .toInteger(K2, function(x) x>=2) - if (!is.numeric(nb_per_chunk) || length(nb_per_chunk)!=2) - stop("'nb_per_chunk': numeric, size 2") - nb_per_chunk[1] <- .toInteger(nb_per_chunk[1], function(x) x>=1) - # A batch of contributions should have at least as many elements as a batch of series, - # because it always contains much less values - nb_per_chunk[2] <- max(.toInteger(nb_per_chunk[2],function(x) x>=1), nb_per_chunk[1]) + nb_series_per_chunk <- .toInteger(nb_series_per_chunk, function(x) x>=1) + # K1 (number of clusters at step 1) cannot exceed nb_series_per_chunk, because we will need + # to load K1 series in memory for clustering stage 2. + if (K1 > nb_series_per_chunk) + stop("'K1' cannot exceed 'nb_series_per_chunk'") nb_items_clust1 <- .toInteger(nb_items_clust1, function(x) x>K1) random <- .toLogical(random) - tryCatch - ( - {ignored <- wavelets::wt.filter(wav_filt)}, - error = function(e) stop("Invalid wavelet filter; see ?wavelets::wt.filter") - ) + tryCatch( {ignored <- wavelets::wt.filter(wav_filt)}, + error = function(e) stop("Invalid wavelet filter; see ?wavelets::wt.filter") ) ctypes = c("relative","absolute","logit") contrib_type = ctypes[ pmatch(contrib_type,ctypes) ] if (is.na(contrib_type)) @@ -176,60 +183,89 @@ claws <- function(getSeries, K1, K2, verbose <- .toLogical(verbose) parll <- .toLogical(parll) - # Serialize series if required, to always use a function - bin_dir <- ".epclust_bin/" - dir.create(bin_dir, showWarnings=FALSE, mode="0755") + # Binarize series if getSeries 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 + # bigmemory::big.matrix, but it would break the "all-is-function" pattern. if (!is.function(getSeries)) { if (verbose) cat("...Serialize time-series\n") - series_file = paste(bin_dir,"data",sep="") ; unlink(series_file) + series_file = ".series.bin" ; unlink(series_file) binarize(getSeries, series_file, nb_series_per_chunk, sep, nbytes, endian) getSeries = function(inds) getDataInFile(inds, series_file, nbytes, endian) } # Serialize all computed wavelets contributions into a file - contribs_file = paste(bin_dir,"contribs",sep="") ; unlink(contribs_file) + contribs_file = ".contribs.bin" ; unlink(contribs_file) index = 1 nb_curves = 0 if (verbose) cat("...Compute contributions and serialize them\n") nb_curves = binarizeTransform(getSeries, - function(series) curvesToContribs(series, wf, ctype), + function(series) curvesToContribs(series, wav_filt, contrib_type), contribs_file, nb_series_per_chunk, nbytes, endian) getContribs = function(indices) getDataInFile(indices, contribs_file, nbytes, endian) + # A few sanity checks: do not continue if too few data available. if (nb_curves < K2) stop("Not enough data: less series than final number of clusters") nb_series_per_task = round(nb_curves / ntasks) if (nb_series_per_task < K2) stop("Too many tasks: less series in one task than final number of clusters") + # Generate a random permutation of 1:N (if random==TRUE); + # otherwise just use arrival (storage) order. + indices_all = if (random) sample(nb_curves) else seq_len(nb_curves) + # Split (all) indices into ntasks groups of ~same size + indices_tasks = lapply(seq_len(ntasks), function(i) { + upper_bound = ifelse( i1) + { + # 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("getSeries","getContribs","K1","K2","algoClust1","algoClust2", + "nb_series_per_chunk","nb_items_clust1","ncores_clust", + "sep","nbytes","endian","verbose","parll") + if (WER=="mix" && ntasks>1) + varlist = c(varlist, "medoids_file") + parallel::clusterExport(cl, varlist, envir = environment()) + } + + # This function achieves one complete clustering task, divided in stage 1 + stage 2. + # stage 1: n indices --> clusteringTask1(...) --> K1 medoids + # stage 2: K1 medoids --> clusteringTask2(...) --> K2 medoids, + # where n = N / ntasks, N being the total number of curves. runTwoStepClustering = function(inds) { + # When running in parallel, the environment is blank: we need to load the required + # packages, and pass useful variables. if (parll && ntasks>1) require("epclust", quietly=TRUE) indices_medoids = clusteringTask1( - inds, getContribs, K1, nb_series_per_chunk, ncores_clust, verbose, parll) - if (WER=="mix") + inds, getContribs, K1, algoClust1, nb_series_per_chunk, ncores_clust, verbose, parll) + if (WER=="mix" && ntasks>1) { - if (parll && ntasks>1) + if (parll) require("bigmemory", quietly=TRUE) medoids1 = bigmemory::as.big.matrix( getSeries(indices_medoids) ) - medoids2 = clusteringTask2(medoids1, K2, getSeries, nb_curves, nb_series_per_chunk, - nbytes, endian, ncores_clust, verbose, parll) - binarize(medoids2, synchrones_file, nb_series_per_chunk, sep, nbytes, endian) + medoids2 = clusteringTask2(medoids1, K2, algoClust2, getSeries, nb_curves, + nb_series_per_chunk, nbytes, endian, ncores_clust, verbose, parll) + binarize(medoids2, medoids_file, nb_series_per_chunk, sep, nbytes, endian) return (vector("integer",0)) } indices_medoids } - # Cluster contributions in parallel (by nb_series_per_chunk) - indices_all = if (random) sample(nb_curves) else seq_len(nb_curves) - indices_tasks = lapply(seq_len(ntasks), function(i) { - upper_bound = ifelse( i1) + {medoids_file = ".medoids.bin" ; unlink(medoids_file)} + if (verbose) { message = paste("...Run ",ntasks," x stage 1", sep="") @@ -237,19 +273,9 @@ claws <- function(getSeries, K1, K2, message = paste(message," + stage 2", sep="") cat(paste(message,"\n", sep="")) } - if (WER=="mix") - {synchrones_file = paste(bin_dir,"synchrones",sep="") ; unlink(synchrones_file)} - if (parll && ntasks>1) - { - cl = parallel::makeCluster(ncores_tasks, outfile="") - varlist = c("getSeries","getContribs","K1","K2","verbose","parll", - "nb_series_per_chunk","ntasks","ncores_clust","sep","nbytes","endian") - if (WER=="mix") - varlist = c(varlist, "synchrones_file") - parallel::clusterExport(cl, varlist=varlist, envir = environment()) - } - # 1000*K1 indices [if WER=="end"], or empty vector [if WER=="mix"] --> series on file + # As explained above, indices will be assigned to ntasks*K1 medoids indices [if WER=="end"], + # or nothing (empty vector) if WER=="mix"; in this case, synchrones are stored in a file. indices <- if (parll && ntasks>1) unlist( parallel::parLapply(cl, indices_tasks, runTwoStepClustering) ) @@ -258,34 +284,45 @@ claws <- function(getSeries, K1, K2, if (parll && ntasks>1) parallel::stopCluster(cl) + # Right before the final stage, two situations are possible: + # a. data to be processed now sit in a binary format in medoids_file (if WER=="mix") + # b. data still is the initial set of curves, referenced by the ntasks*K1 indices + # So, the function getSeries() will potentially change. However, computeSynchrones() + # requires a function retrieving the initial series. Thus, the next line saves future + # conditional instructions. getRefSeries = getSeries - if (WER=="mix") + + if (WER=="mix" && ntasks>1) { indices = seq_len(ntasks*K2) - #Now series must be retrieved from synchrones_file - getSeries = function(inds) getDataInFile(inds, synchrones_file, nbytes, endian) - #Contributions must be re-computed + # Now series (synchrones) must be retrieved from medoids_file + getSeries = function(inds) getDataInFile(inds, medoids_file, nbytes, endian) + # Contributions must be re-computed unlink(contribs_file) index = 1 if (verbose) cat("...Serialize contributions computed on synchrones\n") ignored = binarizeTransform(getSeries, - function(series) curvesToContribs(series, wf, ctype), + function(series) curvesToContribs(series, wav_filt, contrib_type), contribs_file, nb_series_per_chunk, nbytes, endian) } # Run step2 on resulting indices or series (from file) if (verbose) cat("...Run final // stage 1 + stage 2\n") - indices_medoids = clusteringTask1( - indices, getContribs, K1, nb_series_per_chunk, ncores_tasks*ncores_clust, verbose, parll) + indices_medoids = clusteringTask1(indices, getContribs, K1, algoClust1, + nb_series_per_chunk, ncores_tasks*ncores_clust, verbose, parll) medoids1 = bigmemory::as.big.matrix( getSeries(indices_medoids) ) - medoids2 = clusteringTask2(medoids1, K2, getRefSeries, nb_curves, nb_series_per_chunk, - nbytes, endian, ncores_tasks*ncores_clust, verbose, parll) + medoids2 = clusteringTask2(medoids1, K2, algoClust2, getRefSeries, nb_curves, + nb_series_per_chunk, nbytes, endian, ncores_tasks*ncores_clust, verbose, parll) - # Cleanup - unlink(bin_dir, recursive=TRUE) + # Cleanup: remove temporary binary files + tryCatch( + {unlink(series_file); unlink(contribs_file); unlink(medoids_file)}, + error = function(e) {}) + # Return medoids as a standard matrix, since K2 series have to fit in RAM + # (clustering algorithm 1 takes K1 > K2 of them as input) medoids2[,] } @@ -300,14 +337,17 @@ claws <- function(getSeries, K1, K2, #' @return A [big.]matrix of size log(L) x n containing contributions in columns #' #' @export -curvesToContribs = function(series, wav_filt, contrib_type) +curvesToContribs = function(series, wav_filt, contrib_type, coin=FALSE) { + series = as.matrix(series) #1D serie could occur L = nrow(series) D = ceiling( log2(L) ) + # Series are interpolated to all have length 2^D nb_sample_points = 2^D apply(series, 2, function(x) { interpolated_curve = spline(1:L, x, n=nb_sample_points)$y - W = wavelets::dwt(interpolated_curve, filter=wf, D)@W + W = wavelets::dwt(interpolated_curve, filter=wav_filt, D)@W + # Compute the sum of squared discrete wavelet coefficients, for each scale nrj = rev( sapply( W, function(v) ( sqrt( sum(v^2) ) ) ) ) if (contrib_type!="absolute") nrj = nrj / sum(nrj)