X-Git-Url: https://git.auder.net/rpsls.js?a=blobdiff_plain;f=epclust%2FR%2Fmain.R;h=1347fae27a623a43cf9db53a74580c85690f5313;hb=4efef8ccd1522278f53aa5ce265f3a6cfb6fbd9f;hp=27fbb7488394bcc39009c86b74251b3c458cce4b;hpb=3eef8d3df59ded9a281cff51f79fe824198a7427;p=epclust.git diff --git a/epclust/R/main.R b/epclust/R/main.R index 27fbb74..1347fae 100644 --- a/epclust/R/main.R +++ b/epclust/R/main.R @@ -1,91 +1,151 @@ -#' @title Cluster power curves with PAM in parallel +#' @include de_serialize.R +#' @include clustering.R +NULL + +#' CLAWS: CLustering with wAvelets and Wer distanceS #' -#' @description Groups electricity power curves (or any series of similar nature) by applying PAM -#' algorithm in parallel to chunks of size \code{nb_series_per_chunk} +#' Groups electricity power curves (or any series of similar nature) by applying PAM +#' algorithm in parallel to chunks of size \code{nb_series_per_chunk}. Input series +#' must be sampled on the same time grid, no missing values. #' -#' @param data Access to the data, which can be of one of the three following types: -#' \itemize{ -#' \item data.frame: each line contains its ID in the first cell, and all values after -#' \item connection: any R connection object (e.g. a file) providing lines as described above -#' \item function: a custom way to retrieve the curves; it has two arguments: the ranks to be -#' retrieved, and the IDs - at least one of them must be present (priority: ranks). -#' } +#' @param getSeries Access to the (time-)series, which can be of one of the three +#' following types: +#' \itemize{ +#' \item matrix: each line contains all the values for one time-serie, ordered by time +#' \item connection: any R connection object (e.g. a file) providing lines as described above +#' \item function: a custom way to retrieve the curves; it has only one argument: +#' the indices of the series to be retrieved. See examples +#' } #' @param K1 Number of super-consumers to be found after stage 1 (K1 << N) #' @param K2 Number of clusters to be found after stage 2 (K2 << K1) +#' @param random TRUE (default) for random chunks repartition +#' @param wf Wavelet transform filter; see ?wavelets::wt.filter. Default: haar +#' @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 ntasks Number of tasks (parallel iterations to obtain K1 medoids); default: 1. #' Note: ntasks << N, so that N is "roughly divisible" by N (number of series) -#' @param nb_series_per_chunk (Maximum) number of series in each group, inside a task -#' @param min_series_per_chunk Minimum number of series in each group -#' @param wf Wavelet transform filter; see ?wt.filter. Default: haar -#' @param WER "end" to apply stage 2 after stage 1 has iterated and finished, or "mix" -#' to apply it after every stage 1 #' @param ncores_tasks "MPI" number of parallel tasks (1 to disable: sequential tasks) #' @param ncores_clust "OpenMP" number of parallel clusterings in one task -#' @param random Randomize chunks repartition -#' @param ... Other arguments to be passed to \code{data} function +#' @param nb_series_per_chunk (~Maximum) number of series in each group, inside a task +#' @param min_series_per_chunk Minimum number of series in each group +#' @param sep Separator in CSV input file (relevant only if getSeries is a file name) +#' @param nbytes Number of bytes to serialize a floating-point number; 4 or 8 +#' @param endian Endianness to use for (de)serialization. Use "little" or "big" for portability #' -#' @return A data.frame of the final medoids curves (identifiers + values) +#' @return A matrix of the final medoids curves (K2) in rows #' #' @examples -#' getData = function(start, n) { -#' con = dbConnect(drv = RSQLite::SQLite(), dbname = "mydata.sqlite") -#' df = dbGetQuery(con, paste( -#' "SELECT * FROM times_values GROUP BY id OFFSET ",start, +#' \dontrun{ +#' # WER distances computations are a bit too long for CRAN (for now) +#' +#' # Random series around cos(x,2x,3x)/sin(x,2x,3x) +#' x = seq(0,500,0.05) +#' L = length(x) #10001 +#' ref_series = matrix( c(cos(x), cos(2*x), cos(3*x), sin(x), sin(2*x), sin(3*x)), +#' byrows=TRUE, ncol=L ) +#' library(wmtsa) +#' series = do.call( rbind, lapply( 1:6, function(i) +#' do.call(rbind, wmtsa::wavBootstrap(ref_series[i,], n.realization=400)) ) ) +#' #dim(series) #c(2400,10001) +#' medoids_ascii = claws(series_RData, K1=60, K2=6, wf="d8", nb_series_per_chunk=500) +#' +#' # 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, wf="d8", nb_series_per_chunk=500) +#' +#' # Same example, from binary file +#' bin_file = "/tmp/epclust_series.bin" +#' nbytes = 8 +#' endian = "little" +#' epclust::serialize(csv_file, bin_file, 500, nbytes, endian) +#' getSeries = function(indices) getDataInFile(indices, bin_file, nbytes, endian) +#' medoids_bin = claws(getSeries, K1=60, K2=6, wf="d8", nb_series_per_chunk=500) +#' unlink(csv_file) +#' unlink(bin_file) +#' +#' # Same example, from SQLite database +#' library(DBI) +#' series_db <- dbConnect(RSQLite::SQLite(), "file::memory:") +#' # Prepare data.frame in DB-format +#' n = nrow(series) +#' formatted_series = data.frame( +#' ID = rep(1:n,each=L), +#' time = as.POSIXct(1800*(0:n),"GMT",origin="2001-01-01"), +#' value + + + + +#' TODO + + +#' times_values = as.data.frame(series) +#' dbWriteTable(series_db, "times_values", times_values) +#' # NOTE: assume that DB internal data is not reorganized when computing coefficients +#' indexToID_inDB <<- list() +#' getSeries = function(indices) { +#' con = dbConnect(drv = RSQLite::SQLite(), dbname = db_file) +#' if (indices %in% indexToID_inDB) +#' { +#' df = dbGetQuery(con, paste( +#' "SELECT value FROM times_values GROUP BY id OFFSET ",start, #' "LIMIT ", n, " ORDER BY date", sep="")) -#' return (df) +#' return (df) +#' } +#' else +#' { +#' ... +#' } +#' } +#' dbDisconnect(mydb) #' } -#' #####TODO: if DB, array rank --> ID at first retrieval, when computing coeffs; so:: NO use of IDs ! -#' #TODO: 3 examples, data.frame / binary file / DB sqLite -#' + sampleCurves : wavBootstrap de package wmtsa -#' cl = epclust(getData, K1=200, K2=15, ntasks=1000, nb_series_per_chunk=5000, WER="mix") #' @export -epclust = function(series,K1,K2,ntasks=1,nb_series_per_chunk=50*K1,min_series_per_chunk=5*K1, - wf="haar",WER="end",ncores_tasks=1,ncores_clust=4,random=TRUE,ftype="float",...) +claws = function(getSeries, K1, K2, + random=TRUE, #randomize series order? + wf="haar", #stage 1 + WER="end", #stage 2 + ntasks=1, ncores_tasks=1, ncores_clust=4, #control parallelism + nb_series_per_chunk=50*K1, min_series_per_chunk=5*K1, #chunk size + sep=",", #ASCII input separator + nbytes=4, endian=.Platform$endian) #serialization (write,read) { # Check/transform arguments - bin_dir = "epclust.bin/" - dir.create(bin_dir, showWarnings=FALSE, mode="0755") - if (!is.function(series)) - { - series_file = paste(bin_dir,"data",sep="") - unlink(series_file) - } - if (is.matrix(series)) - serialize(series, series_file, ftype, nb_series_per_chunk) - else if (!is.function(series)) + if (!is.matrix(getSeries) && !is.function(getSeries) && + !is(getSeries, "connection" && !is.character(getSeries))) { - tryCatch( - { - if (is.character(series)) - series_con = file(series, open="r") - else if (!isOpen(series)) - { - open(series) - series_con = series - } - serialize(series_con, series_file, ftype, nb_series_per_chunk) - close(series_con) - }, - error=function(e) "series should be a data.frame, a function or a valid connection" - ) + stop("'getSeries': matrix, function, file or valid connection (no NA)") } - if (!is.function(series)) - series = function(indices) getDataInFile(indices, series_file) - getSeries = series - - K1 = toInteger(K1, function(x) x>=2) - K2 = toInteger(K2, function(x) x>=2) - ntasks = toInteger(ntasks, function(x) x>=1) - nb_series_per_chunk = toInteger(nb_series_per_chunk, function(x) x>=K1) - min_series_per_chunk = toInteger(K1, function(x) x>=K1 && x<=nb_series_per_chunk) - ncores_tasks = toInteger(ncores_tasks, function(x) x>=1) - ncores_clust = toInteger(ncores_clust, function(x) x>=1) + K1 = .toInteger(K1, function(x) x>=2) + K2 = .toInteger(K2, function(x) x>=2) + if (!is.logical(random)) + stop("'random': logical") + tryCatch( + {ignored <- wt.filter(wf)}, + error = function(e) stop("Invalid wavelet filter; see ?wavelets::wt.filter")) if (WER!="end" && WER!="mix") stop("WER takes values in {'end','mix'}") + ntasks = .toInteger(ntasks, function(x) x>=1) + ncores_tasks = .toInteger(ncores_tasks, function(x) x>=1) + ncores_clust = .toInteger(ncores_clust, function(x) x>=1) + nb_series_per_chunk = .toInteger(nb_series_per_chunk, function(x) x>=K1) + min_series_per_chunk = .toInteger(K1, function(x) x>=K1 && x<=nb_series_per_chunk) + if (!is.character(sep)) + stop("'sep': character") + nbytes = .toInteger(nbytes, function(x) x==4 || x==8) + + # Serialize series if required, to always use a function + bin_dir = "epclust.bin/" + dir.create(bin_dir, showWarnings=FALSE, mode="0755") + if (!is.function(getSeries)) + { + series_file = paste(bin_dir,"data",sep="") ; unlink(series_file) + serialize(getSeries, series_file, nb_series_per_chunk, sep, nbytes, endian) + getSeries = function(indices) getDataInFile(indices, series_file, nbytes, endian) + } # Serialize all wavelets coefficients (+ IDs) onto a file - coefs_file = paste(bin_dir,"coefs",sep="") - unlink(coefs_file) + coefs_file = paste(bin_dir,"coefs",sep="") ; unlink(coefs_file) index = 1 nb_curves = 0 repeat @@ -93,12 +153,12 @@ epclust = function(series,K1,K2,ntasks=1,nb_series_per_chunk=50*K1,min_series_pe series = getSeries((index-1)+seq_len(nb_series_per_chunk)) if (is.null(series)) break - coeffs_chunk = curvesToCoeffs(series, wf) - serialize(coeffs_chunk, coefs_file, ftype, nb_series_per_chunk) + coefs_chunk = curvesToCoefs(series, wf) + serialize(coefs_chunk, coefs_file, nb_series_per_chunk, sep, nbytes, endian) index = index + nb_series_per_chunk - nb_curves = nb_curves + nrow(coeffs_chunk) + nb_curves = nb_curves + nrow(coefs_chunk) } - getCoefs = function(indices) getDataInFile(indices, coefs_file) + getCoefs = function(indices) getDataInFile(indices, coefs_file, nbytes, endian) if (nb_curves < min_series_per_chunk) stop("Not enough data: less rows than min_series_per_chunk!") @@ -107,26 +167,34 @@ epclust = function(series,K1,K2,ntasks=1,nb_series_per_chunk=50*K1,min_series_pe stop("Too many tasks: less series in one task than min_series_per_chunk!") # Cluster coefficients in parallel (by nb_series_per_chunk) - indices = if (random) sample(nb_curves) else seq_len(nb_curves) + indices_all = if (random) sample(nb_curves) else seq_len(nb_curves) indices_tasks = lapply(seq_len(ntasks), function(i) { upper_bound = ifelse( i series on file) + # 1000*K1 indices [if WER=="end"], or empty vector [if WER=="mix"] --> series on file indices = unlist( parallel::parLapply(cl, indices_tasks, function(inds) { - clusteringTask(inds, getSeries, getSeries, getCoefs, K1, K2*(WER=="mix"), - nb_series_per_chunk,ncores_clust,to_file=TRUE, ftype) + require("epclust", quietly=TRUE) + indices_medoids = clusteringTask(inds,getCoefs,K1,nb_series_per_chunk,ncores_clust) + if (WER=="mix") + { + medoids2 = computeClusters2( + getSeries(indices_medoids), K2, getSeries, nb_series_per_chunk) + serialize(medoids2, synchrones_file, nb_series_per_chunk, sep, nbytes, endian) + return (vector("integer",0)) + } + indices_medoids }) ) parallel::stopCluster(cl) - getSeriesForSynchrones = getSeries - synchrones_file = paste(bin_dir,"synchrones",sep="") + getRefSeries = getSeries + synchrones_file = paste(bin_dir,"synchrones",sep="") ; unlink(synchrones_file) if (WER=="mix") { indices = seq_len(ntasks*K2) #Now series must be retrieved from synchrones_file - getSeries = function(inds) getDataInFile(inds, synchrones_file) + getSeries = function(inds) getDataInFile(inds, synchrones_file, nbytes, endian) #Coefs must be re-computed unlink(coefs_file) index = 1 @@ -135,13 +203,40 @@ epclust = function(series,K1,K2,ntasks=1,nb_series_per_chunk=50*K1,min_series_pe series = getSeries((index-1)+seq_len(nb_series_per_chunk)) if (is.null(series)) break - coeffs_chunk = curvesToCoeffs(series, wf) - serialize(coeffs_chunk, coefs_file, ftype, nb_series_per_chunk) + coefs_chunk = curvesToCoefs(series, wf) + serialize(coefs_chunk, coefs_file, nb_series_per_chunk, sep, nbytes, endian) index = index + nb_series_per_chunk } } # Run step2 on resulting indices or series (from file) - clusteringTask(indices, getSeries, getSeriesForSynchrones, getCoefs, K1, K2, - nb_series_per_chunk, ncores_tasks*ncores_clust, to_file=FALSE, ftype) + indices_medoids = clusteringTask( + indices, getCoefs, K1, nb_series_per_chunk, ncores_tasks*ncores_clust) + computeClusters2(getSeries(indices_medoids),K2,getRefSeries,nb_series_per_chunk) +} + +# helper +curvesToCoefs = function(series, wf) +{ + L = length(series[1,]) + D = ceiling( log2(L) ) + nb_sample_points = 2^D + t( apply(series, 1, function(x) { + interpolated_curve = spline(1:L, x, n=nb_sample_points)$y + W = wavelets::dwt(interpolated_curve, filter=wf, D)@W + rev( sapply( W, function(v) ( sqrt( sum(v^2) ) ) ) ) + }) ) +} + +# helper +.toInteger <- function(x, condition) +{ + if (!is.integer(x)) + tryCatch( + {x = as.integer(x)[1]}, + error = function(e) paste("Cannot convert argument",substitute(x),"to integer") + ) + if (!condition(x)) + stop(paste("Argument",substitute(x),"does not verify condition",body(condition))) + x }