X-Git-Url: https://git.auder.net/?a=blobdiff_plain;f=epclust%2FR%2Fclustering.R;h=c22678632dcee57df836d4e75978299bb9eaa027;hb=6ad3f3fd0ec4f3cd1fd5de4a287c1893293e5bcc;hp=e27ea353e479fb8185afb124f02fbcab56c9d4e0;hpb=5c6529795907ba1b34d4552cbfd0e0cbb77cac0f;p=epclust.git diff --git a/epclust/R/clustering.R b/epclust/R/clustering.R index e27ea35..c226786 100644 --- a/epclust/R/clustering.R +++ b/epclust/R/clustering.R @@ -1,75 +1,231 @@ -oneIteration = function(..........) +#' @name clustering +#' @rdname clustering +#' @aliases clusteringTask1 computeClusters1 computeClusters2 +#' +#' @title Two-stage clustering, withing one task (see \code{claws()}) +#' +#' @description \code{clusteringTask1()} runs one full stage-1 task, which consists in +#' iterated stage 1 clustering (on nb_curves / ntasks energy contributions, computed +#' through discrete wavelets coefficients). +#' \code{clusteringTask2()} runs a full stage-2 task, which consists in synchrones +#' and then WER distances computations, before applying the clustering algorithm. +#' \code{computeClusters1()} and \code{computeClusters2()} correspond to the atomic +#' clustering procedures respectively for stage 1 and 2. The former applies the +#' clustering algorithm (PAM) on a contributions matrix, while the latter clusters +#' a chunk of series inside one task (~max nb_series_per_chunk) +#' +#' @param indices Range of series indices to cluster in parallel (initial data) +#' @param getContribs Function to retrieve contributions from initial series indices: +#' \code{getContribs(indices)} outpus a contributions matrix +#' @param contribs matrix of contributions (e.g. output of \code{curvesToContribs()}) +#' @param distances matrix of K1 x K1 (WER) distances between synchrones +#' @inheritParams computeSynchrones +#' @inheritParams claws +#' +#' @return For \code{clusteringTask1()} and \code{computeClusters1()}, the indices of the +#' computed (K1) medoids. Indices are irrelevant for stage 2 clustering, thus +#' \code{computeClusters2()} outputs a big.matrix of medoids +#' (of size limited by nb_series_per_chunk) +NULL + +#' @rdname clustering +#' @export +clusteringTask1 = function( + indices, getContribs, K1, nb_series_per_chunk, ncores_clust=1, verbose=FALSE, parll=TRUE) { - cl_clust = parallel::makeCluster(ncores_clust) - parallel::clusterExport(cl_clust, .............., envir=........) - indices_clust = indices_task[[i]] - repeat - { - nb_workers = max( 1, round( length(indices_clust) / nb_series_per_chunk ) ) - indices_workers = list() - #indices[[i]] == (start_index,number_of_elements) - for (i in 1:nb_workers) + if (verbose) + cat(paste("*** Clustering task 1 on ",length(indices)," lines\n", sep="")) + + if (parll) + { + cl = parallel::makeCluster(ncores_clust) + parallel::clusterExport(cl, varlist=c("getContribs","K1","verbose"), envir=environment()) + } + while (length(indices) > K1) + { + indices_workers = .spreadIndices(indices, nb_series_per_chunk) + indices <- + if (parll) { - upper_bound = ifelse( i 0) - { - curves = computeSynchrones(cl) - dists = computeWerDists(curves) - cl = computeClusters(dists, K2) - } - cl + if (verbose) + cat(paste("*** Clustering task 2 on ",nrow(medoids)," lines\n", sep="")) + + if (nrow(medoids) <= K2) + return (medoids) + synchrones = computeSynchrones(medoids, + getRefSeries, nb_ref_curves, nb_series_per_chunk, ncores_clust, verbose, parll) + distances = computeWerDists(synchrones, ncores_clust, verbose, parll) + medoids[ computeClusters2(distances,K2,verbose), ] } -computeClusters = function(data, K) +#' @rdname clustering +#' @export +computeClusters1 = function(contribs, K1, verbose=FALSE) { - library(cluster) - pam_output = cluster::pam(data, K) - return ( list( clusts=pam_output$clustering, medoids=pam_output$medoids, - ranks=pam_output$id.med ) ) + if (verbose) + cat(paste(" computeClusters1() on ",nrow(contribs)," lines\n", sep="")) + cluster::pam(contribs, K1, diss=FALSE)$id.med } -#TODO: appendCoeffs() en C --> serialize et append to file +#' @rdname clustering +#' @export +computeClusters2 = function(distances, K2, verbose=FALSE) +{ + if (verbose) + cat(paste(" computeClusters2() on ",nrow(distances)," lines\n", sep="")) + cluster::pam(distances, K2, diss=TRUE)$id.med +} -computeSynchrones = function(...) +#' computeSynchrones +#' +#' Compute the synchrones curves (sum of clusters elements) from a matrix of medoids, +#' using L2 distances. +#' +#' @param medoids big.matrix of medoids (curves of same length as initial series) +#' @param getRefSeries Function to retrieve initial series (e.g. in stage 2 after series +#' have been replaced by stage-1 medoids) +#' @param nb_ref_curves How many reference series? (This number is known at this stage) +#' @inheritParams claws +#' +#' @return A big.matrix of size K1 x L where L = data_length +#' +#' @export +computeSynchrones = function(medoids, getRefSeries, + nb_ref_curves, nb_series_per_chunk, ncores_clust=1,verbose=FALSE,parll=TRUE) { + if (verbose) + cat(paste("--- Compute synchrones\n", sep="")) + computeSynchronesChunk = function(indices) + { + if (parll) + { + require("bigmemory", quietly=TRUE) + requireNamespace("synchronicity", quietly=TRUE) + require("epclust", quietly=TRUE) + synchrones <- bigmemory::attach.big.matrix(synchrones_desc) + counts <- bigmemory::attach.big.matrix(counts_desc) + medoids <- bigmemory::attach.big.matrix(medoids_desc) + m <- synchronicity::attach.mutex(m_desc) + } + + ref_series = getRefSeries(indices) + nb_series = nrow(ref_series) + + #get medoids indices for this chunk of series + mi = computeMedoidsIndices(medoids@address, ref_series) + + for (i in seq_len(nb_series)) + { + if (parll) + synchronicity::lock(m) + synchrones[ mi[i], ] = synchrones[ mi[i], ] + ref_series[i,] + counts[ mi[i] ] = counts[ mi[i] ] + 1 #TODO: remove counts? + if (parll) + synchronicity::unlock(m) + } + } + + K = nrow(medoids) ; L = ncol(medoids) + # Use bigmemory (shared==TRUE by default) + synchronicity to fill synchrones in // + # TODO: if size > RAM (not our case), use file-backed big.matrix + synchrones = bigmemory::big.matrix(nrow=K, ncol=L, type="double", init=0.) + counts = bigmemory::big.matrix(nrow=K, ncol=1, type="double", init=0) + # synchronicity is only for Linux & MacOS; on Windows: run sequentially + parll = (requireNamespace("synchronicity",quietly=TRUE) + && parll && Sys.info()['sysname'] != "Windows") + if (parll) + { + m <- synchronicity::boost.mutex() + m_desc <- synchronicity::describe(m) + synchrones_desc = bigmemory::describe(synchrones) + counts_desc = bigmemory::describe(counts) + medoids_desc = bigmemory::describe(medoids) + cl = parallel::makeCluster(ncores_clust) + parallel::clusterExport(cl, varlist=c("synchrones_desc","counts_desc","counts", + "verbose","m_desc","medoids_desc","getRefSeries"), envir=environment()) + } + + indices_workers = .spreadIndices(seq_len(nb_ref_curves), nb_series_per_chunk) + ignored <- + if (parll) + parallel::parLapply(cl, indices_workers, computeSynchronesChunk) + else + lapply(indices_workers, computeSynchronesChunk) + + if (parll) + parallel::stopCluster(cl) + + #TODO: can we avoid this loop? ( synchrones = sweep(synchrones, 1, counts, '/') ) + for (i in seq_len(K)) + synchrones[i,] = synchrones[i,] / counts[i,1] + #NOTE: odds for some clusters to be empty? (when series already come from stage 2) + # ...maybe; but let's hope resulting K1' be still quite bigger than K2 + noNA_rows = sapply(seq_len(K), function(i) all(!is.nan(synchrones[i,]))) + if (all(noNA_rows)) + return (synchrones) + # Else: some clusters are empty, need to slice synchrones + synchrones[noNA_rows,] } -#Entrée : courbes synchrones, soit après étape 1 itérée, soit après chaqure étape 1 -computeWerDist = function(conso) +#' computeWerDists +#' +#' Compute the WER distances between the synchrones curves (in rows), which are +#' returned (e.g.) by \code{computeSynchrones()} +#' +#' @param synchrones A big.matrix of synchrones, in rows. The series have same length +#' as the series in the initial dataset +#' @inheritParams claws +#' +#' @return A matrix of size K1 x K1 +#' +#' @export +computeWerDists = function(synchrones, ncores_clust=1,verbose=FALSE,parll=TRUE) { - if (!require("Rwave", quietly=TRUE)) - stop("Unable to load Rwave library") - n <- nrow(conso) - delta <- ncol(conso) + if (verbose) + cat(paste("--- Compute WER dists\n", sep="")) + + + + +#TODO: serializer les CWT, les récupérer via getDataInFile +#--> OK, faut juste stocker comme séries simples de taille delta*ncol (53*17519) + + + + + n <- nrow(synchrones) + delta <- ncol(synchrones) #TODO: automatic tune of all these parameters ? (for other users) nvoice <- 4 - # noctave = 2^13 = 8192 half hours ~ 180 days ; ~log2(ncol(conso)) + # noctave = 2^13 = 8192 half hours ~ 180 days ; ~log2(ncol(synchrones)) noctave = 13 # 4 here represent 2^5 = 32 half-hours ~ 1 day #NOTE: default scalevector == 2^(0:(noctave * nvoice) / nvoice) * s0 (?) - scalevector <- 2^(4:(noctave * nvoice) / nvoice) * 2 + scalevector <- 2^(4:(noctave * nvoice) / nvoice + 1) #condition: ( log2(s0*w0/(2*pi)) - 1 ) * nvoice + 1.5 >= 1 s0=2 w0=2*pi @@ -77,33 +233,113 @@ computeWerDist = function(conso) s0log = as.integer( (log2( s0*w0/(2*pi) ) - 1) * nvoice + 1.5 ) totnoct = noctave + as.integer(s0log/nvoice) + 1 - # (normalized) observations node with CWT - Xcwt4 <- lapply(seq_len(n), function(i) { - ts <- scale(ts(conso[i,]), center=TRUE, scale=scaled) - totts.cwt = Rwave::cwt(ts,totnoct,nvoice,w0,plot=0) - ts.cwt = totts.cwt[,s0log:(s0log+noctave*nvoice)] - #Normalization - sqs <- sqrt(2^(0:(noctave*nvoice)/nvoice)*s0) - sqres <- sweep(ts.cwt,MARGIN=2,sqs,'*') - sqres / max(Mod(sqres)) - }) - - Xwer_dist <- matrix(0., n, n) - fcoefs = rep(1/3, 3) #moving average on 3 values (TODO: very slow! correct?!) - for (i in 1:(n-1)) + Xwer_dist <- bigmemory::big.matrix(nrow=n, ncol=n, type="double") + + # Generate n(n-1)/2 pairs for WER distances computations +# pairs = list() +# V = seq_len(n) +# for (i in 1:n) +# { +# V = V[-1] +# pairs = c(pairs, lapply(V, function(v) c(i,v))) +# } + # Generate "smart" pairs for WER distances computations + pairs = list() + F = floor(2*n/3) + for (i in 1:F) + pairs = c(pairs, lapply((i+1):n, function(v) c(i,v))) + V = (F+1):n + for (i in (F+1):(n-1)) + { + V = V[-1] + pairs = c(pairs, + + # Distance between rows i and j + computeDistancesIJ = function(pair) + { + if (parll) + { + require("bigmemory", quietly=TRUE) + require("epclust", quietly=TRUE) + synchrones <- bigmemory::attach.big.matrix(synchrones_desc) + Xwer_dist <- bigmemory::attach.big.matrix(Xwer_dist_desc) + } + + computeCWT = function(index) + { + ts <- scale(ts(synchrones[index,]), center=TRUE, scale=scaled) + totts.cwt = Rwave::cwt(ts, totnoct, nvoice, w0, plot=FALSE) + ts.cwt = totts.cwt[,s0log:(s0log+noctave*nvoice)] + #Normalization + sqs <- sqrt(2^(0:(noctave*nvoice)/nvoice)*s0) + sqres <- sweep(ts.cwt,2,sqs,'*') + sqres / max(Mod(sqres)) + } + + i = pair[1] ; j = pair[2] + if (verbose && j==i+1) + cat(paste(" Distances (",i,",",j,"), (",i,",",j+1,") ...\n", sep="")) + cwt_i <- computeCWT(i) + cwt_j <- computeCWT(j) + +#print(system.time( { + num <- epclustFilter(Mod(cwt_i * Conj(cwt_j))) + WX <- epclustFilter(Mod(cwt_i * Conj(cwt_i))) + WY <- epclustFilter(Mod(cwt_j * Conj(cwt_j))) + wer2 <- sum(colSums(num)^2) / sum(colSums(WX) * colSums(WY)) + Xwer_dist[i,j] <- sqrt(delta * ncol(cwt_i) * max(1 - wer2, 0.)) #FIXME: wer2 should be < 1 + Xwer_dist[j,i] <- Xwer_dist[i,j] +#} ) ) + Xwer_dist[i,i] = 0. + } + + if (parll) + { + cl = parallel::makeCluster(ncores_clust) + synchrones_desc <- bigmemory::describe(synchrones) + Xwer_dist_desc <- bigmemory::describe(Xwer_dist) + + parallel::clusterExport(cl, varlist=c("synchrones_desc","Xwer_dist_desc","totnoct", + "nvoice","w0","s0log","noctave","s0","verbose"), envir=environment()) + } + + ignored <- + if (parll) + parallel::parLapply(cl, pairs, computeDistancesIJ) + else + lapply(pairs, computeDistancesIJ) + + if (parll) + parallel::stopCluster(cl) + + Xwer_dist[n,n] = 0. + distances <- Xwer_dist[,] + rm(Xwer_dist) ; gc() + distances #~small matrix K1 x K1 +} + +# Helper function to divide indices into balanced sets +.spreadIndices = function(indices, nb_per_chunk) +{ + L = length(indices) + nb_workers = floor( L / nb_per_chunk ) + if (nb_workers == 0) + { + # L < nb_series_per_chunk, simple case + indices_workers = list(indices) + } + else { - for (j in (i+1):n) + indices_workers = lapply( seq_len(nb_workers), function(i) + indices[(nb_per_chunk*(i-1)+1):(nb_per_chunk*i)] ) + # Spread the remaining load among the workers + rem = L %% nb_per_chunk + while (rem > 0) { - #TODO: later, compute CWT here (because not enough storage space for 32M series) - # 'circular=TRUE' is wrong, should just take values on the sides; to rewrite in C - num <- filter(Mod(Xcwt4[[i]] * Conj(Xcwt4[[j]])), fcoefs, circular=TRUE) - WX <- filter(Mod(Xcwt4[[i]] * Conj(Xcwt4[[i]])), fcoefs, circular=TRUE) - WY <- filter(Mod(Xcwt4[[j]] * Conj(Xcwt4[[j]])), fcoefs, circular=TRUE) - wer2 <- sum(colSums(num)^2) / sum( sum(colSums(WX) * colSums(WY)) ) - Xwer_dist[i,j] <- sqrt(delta * ncol(Xcwt4[[1]]) * (1 - wer2)) - Xwer_dist[j,i] <- Xwer_dist[i,j] + index = rem%%nb_workers + 1 + indices_workers[[index]] = c(indices_workers[[index]], indices[L-rem+1]) + rem = rem - 1 } } - diag(Xwer_dist) <- numeric(n) - Xwer_dist + indices_workers }