-#' @include defaults.R
-
-#' @title Cluster power curves with PAM in parallel
+#' 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 start index
-#' (start) and number of curves (n); see example in package vignette.
-#' }
+#' @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
+#' }
+#' @inheritParams clustering
#' @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 wf Wavelet transform filter; see ?wavelets::wt.filter
+#' @param ctype Type of contribution: "relative" or "absolute" (or 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 random TRUE (default) for random chunks repartition
#' @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 ncores_tasks "MPI" number of parallel tasks (1 to disable: sequential tasks)
+#' @param ncores_clust "OpenMP" number of parallel clusterings in one task
+#' @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 writeTmp Function to write temporary wavelets coefficients (+ identifiers);
-#' see defaults in defaults.R
-#' @param readTmp Function to read temporary wavelets coefficients (see defaults.R)
-#' @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 number of parallel tasks (1 to disable: 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 to use for (de)serialization. Use "little" or "big" for portability
+#' @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)
#'
-#' @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,
-#' "LIMIT ", n, " ORDER BY date", sep=""))
-#' return (df)
+#' \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)),
+#' byrow=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, K1=60, K2=6, "d8", "rel", 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, "d8", "rel", nb_series_per_chunk=500)
+#'
+#' # Same example, from binary file
+#' bin_file = "/tmp/epclust_series.bin"
+#' nbytes = 8
+#' endian = "little"
+#' epclust::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, "d8", "rel", 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)
+#' 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)) )
+#' 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"] )
+#' getSeries = function(indices) {
+#' request = "SELECT id,value FROM times_values WHERE id in ("
+#' for (i in indices)
+#' request = paste(request, i, ",", sep="")
+#' request = paste(request, ")", sep="")
+#' df_series = dbGetQuery(series_db, request)
+#' # Assume that all series share same length at this stage
+#' ts_length = sum(df_series[,"id"] == df_series[1,"id"])
+#' t( as.matrix(df_series[,"value"], nrow=ts_length) )
+#' }
+#' medoids_db = claws(getSeries, K1=60, K2=6, "d8", "rel", nb_series_per_chunk=500)
+#' dbDisconnect(series_db)
+#'
+#' # All computed medoids should be the same:
+#' digest::sha1(medoids_ascii)
+#' digest::sha1(medoids_csv)
+#' digest::sha1(medoids_bin)
+#' digest::sha1(medoids_db)
#' }
-#' cl = epclust(getData, K1=200, K2=15, ntasks=1000, nb_series_per_chunk=5000, WER="mix")
#' @export
-epclust = function(data, K1, K2,
- ntasks=1, nb_series_per_chunk=50*K1, min_series_per_chunk=5*K1,
- writeTmp=defaultWriteTmp, readTmp=defaultReadTmp, wf="haar", WER="end",
- ncores_tasks=1, ncores_clust=4)
+claws = function(getSeries, K1, K2,
+ wf,ctype, #stage 1
+ WER="end", #stage 2
+ random=TRUE, #randomize series order?
+ 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)
+ verbose=FALSE, parll=TRUE)
{
- #TODO: setRefClass(...) to avoid copy data:
- #http://stackoverflow.com/questions/2603184/r-pass-by-reference
-
- #0) check arguments
- if (!is.data.frame(data) && !is.function(data))
- tryCatch(
- {
- if (is.character(data))
- {
- data_con = file(data, open="r")
- } else if (!isOpen(data))
- {
- open(data)
- data_con = data
- }
- },
- error="data should be a data.frame, a function or a valid connection")
- if (!is.integer(K) || K < 2)
- stop("K should be an integer greater or equal to 2")
- if (!is.integer(nb_series_per_chunk) || nb_series_per_chunk < K)
- stop("nb_series_per_chunk should be an integer greater or equal to K")
- if (!is.function(writeTmp) || !is.function(readTmp))
- stop("read/writeTmp should be functional (see defaults.R)")
+ # Check/transform arguments
+ if (!is.matrix(getSeries) && !bigmemory::is.big.matrix(getSeries)
+ && !is.function(getSeries)
+ && !methods::is(getSeries,"connection") && !is.character(getSeries))
+ {
+ stop("'getSeries': [big]matrix, function, file or valid connection (no NA)")
+ }
+ K1 = .toInteger(K1, function(x) x>=2)
+ K2 = .toInteger(K2, function(x) x>=2)
+ if (!is.logical(random))
+ stop("'random': logical")
+ tryCatch(
+ {ignored <- wavelets::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'}")
- #concerning ncores, any non-integer type will be treated as "use parallel:detectCores()/4"
+ 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)
- #1) acquire data (process curves, get as coeffs)
- #TODO: for data.frame and custom function, run in parallel (connections are sequential[?!])
- index = 1
- nb_curves = 0
- repeat
+ # 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))
{
- coeffs_chunk = NULL
- if (is.data.frame(data))
- {
- #full data matrix
- if (index < nrow(data))
- {
- coeffs_chunk = curvesToCoeffs(
- data[index:(min(index+nb_series_per_chunk-1,nrow(data))),], wf)
- }
- } else if (is.function(data))
- {
- #custom user function to retrieve next n curves, probably to read from DB
- coeffs_chunk = curvesToCoeffs( data(index, nb_series_per_chunk), wf )
- } else
- {
- #incremental connection
- #TODO: find a better way to parse than using a temp file
- ascii_lines = readLines(data_con, nb_series_per_chunk)
- if (length(ascii_lines > 0))
- {
- series_chunk_file = ".tmp/series_chunk"
- writeLines(ascii_lines, series_chunk_file)
- coeffs_chunk = curvesToCoeffs( read.csv(series_chunk_file), wf )
- }
- }
- if (is.null(coeffs_chunk))
- break
- writeTmp(coeffs_chunk)
- nb_curves = nb_curves + nrow(coeffs_chunk)
- index = index + nb_series_per_chunk
+ if (verbose)
+ cat("...Serialize time-series\n")
+ series_file = paste(bin_dir,"data",sep="") ; unlink(series_file)
+ binarize(getSeries, series_file, nb_series_per_chunk, sep, nbytes, endian)
+ getSeries = function(inds) getDataInFile(inds, series_file, nbytes, endian)
}
- if (exists(data_con))
- close(data_con)
+
+ # Serialize all computed wavelets contributions onto a file
+ contribs_file = paste(bin_dir,"contribs",sep="") ; 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),
+ contribs_file, nb_series_per_chunk, nbytes, endian)
+ getContribs = function(indices) getDataInFile(indices, contribs_file, nbytes, endian)
+
if (nb_curves < min_series_per_chunk)
stop("Not enough data: less rows than min_series_per_chunk!")
+ nb_series_per_task = round(nb_curves / ntasks)
+ if (nb_series_per_task < min_series_per_chunk)
+ stop("Too many tasks: less series in one task than min_series_per_chunk!")
- #2) process coeffs (by nb_series_per_chunk) and cluster them in parallel
- library(parallel)
- cl_tasks = parallel::makeCluster(ncores_tasks)
- #Nothing to export because each worker retrieve and put data from/on files (or DB)
- #parallel::clusterExport(cl=cl, varlist=c("nothing","to","export"), envir=environment())
- #TODO: be careful of writing to a new temp file, then flush initial one, then re-use it...
- res_tasks = parallel::parSapply(cl_tasks, 1:ntasks, function() {
- cl_clust = parallel::makeCluster(ncores_clust)
- repeat
- {
- #while there are jobs to do
- #(i.e. size of tmp "file" is greater than ntasks * nb_series_per_chunk)
- nb_workers = nb_curves %/% nb_series_per_chunk
- indices = list()
- #indices[[i]] == (start_index,number_of_elements)
- for (i in 1:nb_workers)
- indices[[i]] = c(nb_series_per_chunk*(i-1)+1, nb_series_per_chunk)
- remainder = nb_curves %% nb_series_per_chunk
- if (remainder >= min_series_per_chunk)
- {
- nb_workers = nb_workers + 1
- indices[[nb_workers]] = c(nb_curves-remainder+1, nb_curves)
- } else if (remainder > 0)
- {
- #spread the load among other workers
- #...
- }
- res_clust = parallel::parSapply(cl, indices, processChunk, K, WER=="mix")
- #C) flush tmp file (current parallel processes will write in it)
- }
- parallel:stopCluster(cl_clust)
+ # 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( i<ntasks, min(nb_series_per_task*i,nb_curves), nb_curves )
+ indices_all[((i-1)*nb_series_per_task+1):upper_bound]
})
- parallel::stopCluster(cl_tasks)
+ if (verbose)
+ cat(paste("...Run ",ntasks," x stage 1 in parallel\n",sep=""))
+ if (parll)
+ {
+ cl = parallel::makeCluster(ncores_tasks)
+ parallel::clusterExport(cl, varlist=c("getSeries","getContribs","K1","K2","verbose","parll",
+ "nb_series_per_chunk","ncores_clust","synchrones_file","sep","nbytes","endian"),
+ envir = environment())
+ }
- #3) readTmp last results, apply PAM on it, and return medoids + identifiers
- final_coeffs = readTmp(1, nb_series_per_chunk)
- if (nrow(final_coeffs) == K)
+ runTwoStepClustering = function(inds)
{
- return ( list( medoids=coeffsToCurves(final_coeffs[,2:ncol(final_coeffs)]),
- ids=final_coeffs[,1] ) )
+ if (parll)
+ require("epclust", quietly=TRUE)
+ indices_medoids = clusteringTask1(
+ inds, getContribs, K1, nb_series_per_chunk, ncores_clust, verbose, parll)
+ if (WER=="mix")
+ {
+ medoids2 = computeClusters2(getSeries(indices_medoids),
+ K2, getSeries, nb_curves, nb_series_per_chunk, ncores_clust, verbose, parll)
+ binarize(medoids2, synchrones_file, nb_series_per_chunk, sep, nbytes, endian)
+ return (vector("integer",0))
+ }
+ indices_medoids
}
- pam_output = getClusters(as.matrix(final_coeffs[,2:ncol(final_coeffs)]), K)
- medoids = coeffsToCurves(pam_output$medoids, wf)
- ids = final_coeffs[,1] [pam_output$ranks]
- #4) apply stage 2 (in parallel ? inside task 2) ?)
- if (WER == "end")
+ # 1000*K1 indices [if WER=="end"], or empty vector [if WER=="mix"] --> series on file
+ if (parll)
+ indices = unlist( parallel::parLapply(cl, indices_tasks, runTwoStepClustering) )
+ else
+ indices = unlist( lapply(indices_tasks, runTwoStepClustering) )
+ if (parll)
+ parallel::stopCluster(cl)
+
+ getRefSeries = getSeries
+ synchrones_file = paste(bin_dir,"synchrones",sep="") ; unlink(synchrones_file)
+ if (WER=="mix")
{
- #from center curves, apply stage 2...
- #TODO:
+ 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
+ unlink(contribs_file)
+ index = 1
+ if (verbose)
+ cat("...Serialize contributions computed on synchrones\n")
+ ignored = binarizeTransform(getSeries,
+ function(series) curvesToContribs(series, wf, ctype),
+ contribs_file, nb_series_per_chunk, nbytes, endian)
}
- return (list(medoids=medoids, ids=ids))
+ # 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)
+ medoids = computeClusters2(getSeries(indices_medoids),
+ K2, getRefSeries, nb_curves, nb_series_per_chunk, ncores_tasks*ncores_clust, verbose)
+
+ # Cleanup
+ unlink(bin_dir, recursive=TRUE)
+
+ medoids
}
-processChunk = function(indice, K, WER)
+#' curvesToContribs
+#'
+#' Compute the discrete wavelet coefficients for each series, and aggregate them in
+#' energy contribution across scales as described in https://arxiv.org/abs/1101.4744v2
+#'
+#' @param series Matrix of series (in rows), of size n x L
+#' @inheritParams claws
+#'
+#' @return A matrix of size n x log(L) containing contributions in rows
+#'
+#' @export
+curvesToContribs = function(series, wf, ctype)
{
- #1) retrieve data
- coeffs = readTmp(indice[1], indice[2])
- #2) cluster
- cl = getClusters(as.matrix(coeffs[,2:ncol(coeffs)]), K)
- #3) WER (optional)
- #TODO:
+ L = length(series[1,])
+ D = ceiling( log2(L) )
+ nb_sample_points = 2^D
+ cont_types = c("relative","absolute")
+ ctype = cont_types[ pmatch(ctype,cont_types) ]
+ 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
+ nrj = rev( sapply( W, function(v) ( sqrt( sum(v^2) ) ) ) )
+ if (ctype=="relative") nrj / sum(nrj) else nrj
+ }) )
}
-#TODO: difficulté : retrouver courbe à partir de l'identifiant (DB ok mais le reste ?)
-#aussi : que passe-t-on aux noeuds ? curvesToCoeffs en // ?
-#enfin : WER ?!
+# Check integer arguments with functional conditions
+.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
+}
projects / epclust.git / blobdiff