epclust/R/main.R

   1 #' CLAWS: CLustering with wAvelets and Wer distanceS
   2 #'
   3 #' Cluster electricity power curves (or any series of similar nature) by applying a
   4 #' two stage procedure in parallel (see details).
   5 #' Input series must be sampled on the same time grid, no missing values.
   6 #'
   7 #' Summary of the function execution flow:
   8 #' \enumerate{
   9 #'   \item Compute and serialize all contributions, obtained through discrete wavelet
  10 #'     decomposition (see Antoniadis & al. [2013])
  11 #'   \item Divide series into \code{ntasks} groups to process in parallel. In each task:
  12 #'   \enumerate{
  13 #'     \item iterate the first clustering algorithm on its aggregated outputs,
  14 #'       on inputs of size \code{nb_items_clust}\cr
  15 #'         -> K1 medoids indices
  16 #'     \item optionally, if WER=="mix":\cr
  17 #'       a. compute WER distances (K1xK1) between medoids\cr
  18 #'       b. apply the 2nd clustering algorithm\cr
  19 #'          -> K2 medoids indices
  20 #'   }
  21 #'   \item Launch a final task on the aggregated outputs of all previous tasks:
  22 #'     ntasks*K1 if WER=="end", ntasks*K2 otherwise
  23 #'   \item Compute synchrones (sum of series within each final group)
  24 #' }
  25 #'
  26 #' The main argument -- \code{series} -- has a quite misleading name, since it can be
  27 #' either a [big.]matrix, a CSV file, a connection or a user function to retrieve series.
  28 #' When \code{series} is given as a function it must take a single argument,
  29 #' 'indices': integer vector equal to the indices of the curves to retrieve;
  30 #' see SQLite example.
  31 #' WARNING: the return value must be a matrix (in columns), or NULL if no matches.
  32 #'
  33 #' Note: Since we don't make assumptions on initial data, there is a possibility that
  34 #' even when serialized, contributions do not fit in RAM. For example,
  35 #' 30e6 series of length 100,000 would lead to a +4Go contribution matrix. Therefore,
  36 #' it's safer to place these in (binary) files; that's what we do.
  37 #'
  38 #' @param series Access to the N (time-)series, which can be of one of the four
  39 #'   following types:
  40 #'   \itemize{
  41 #'     \item [big.]matrix: each column contains the (time-ordered) values of one time-serie
  42 #'     \item connection: any R connection object providing lines as described above
  43 #'     \item character: name of a CSV file containing series in rows (no header)
  44 #'     \item function: a custom way to retrieve the curves; it has only one argument:
  45 #'       the indices of the series to be retrieved. See SQLite example
  46 #'   }
  47 #' @param K1 Number of clusters to be found after stage 1 (K1 << N)
  48 #' @param K2 Number of clusters to be found after stage 2 (K2 << K1)
  49 #' @param nb_series_per_chunk Number of series to retrieve in one batch
  50 #' @param nb_items_clust Number of items in 1st clustering algorithm input
  51 #' @param algoClust1 Clustering algorithm for stage 1. A function which takes (data, K)
  52 #'   as argument where data is a matrix in columns and K the desired number of clusters,
  53 #'   and outputs K medoids ranks. Default: PAM.
  54 #' @param algoClust2 Clustering algorithm for stage 2. A function which takes (dists, K)
  55 #'   as argument where dists is a matrix of distances and K the desired number of clusters,
  56 #'   and outputs K medoids ranks. Default: PAM.
  57 #' @param wav_filt Wavelet transform filter; see ?wavelets::wt.filter
  58 #' @param contrib_type Type of contribution: "relative", "logit" or "absolute" (any prefix)
  59 #' @param WER "end" to apply stage 2 after stage 1 has fully iterated, or "mix" to apply
  60 #'   stage 2 at the end of each task
  61 #' @param smooth_lvl Smoothing level: odd integer, 1 == no smoothing.
  62 #' @param nvoice Number of voices within each octave for CWT computations
  63 #' @param random TRUE (default) for random chunks repartition
  64 #' @param ntasks Number of tasks (parallel iterations to obtain K1 [if WER=="end"]
  65 #'   or K2 [if WER=="mix"] medoids); default: 1.
  66 #'   Note: ntasks << N (number of series), so that N is "roughly divisible" by ntasks
  67 #' @param ncores_tasks Number of parallel tasks ('1' == sequential tasks)
  68 #' @param ncores_clust Number of parallel clusterings in one task
  69 #' @param sep Separator in CSV input file (if any provided)
  70 #' @param nbytes Number of bytes to serialize a floating-point number: 4 or 8
  71 #' @param endian Endianness for (de)serialization: "little" or "big"
  72 #' @param verbose FALSE: nothing printed; TRUE: some execution traces
  73 #' @param parll TRUE: run in parallel. FALSE: run sequentially
  74 #' @param reuse_bin Re-use previously stored binary series and contributions
  75 #'
  76 #' @return A list:
  77 #' \itemize{
  78 #'   \item medoids: matrix of the final K2 medoids curves
  79 #'   \item ranks: corresponding indices in the dataset
  80 #'   \item synchrones: sum of series within each final group
  81 #' }
  82 #'
  83 #' @references Clustering functional data using Wavelets [2013];
  84 #'   A. Antoniadis, X. Brossat, J. Cugliari & J.-M. Poggi.
  85 #'   Inter. J. of Wavelets, Multiresolution and Information Procesing,
  86 #'   vol. 11, No 1, pp.1-30. doi:10.1142/S0219691313500033
  87 #'
  88 #' @examples
  89 #' \dontrun{
  90 #' # WER distances computations are too long for CRAN (for now)
  91 #'
  92 #' # Random series around cos(x,2x,3x)/sin(x,2x,3x)
  93 #' x <- seq(0,50,0.05)
  94 #' L <- length(x) #1001
  95 #' ref_series <- matrix( c(cos(x),cos(2*x),cos(3*x),sin(x),sin(2*x),sin(3*x)), ncol=6 )
  96 #' library(wmtsa)
  97 #' series <- do.call( cbind, lapply( 1:6, function(i)
  98 #'   do.call(cbind, wmtsa::wavBootstrap(ref_series[,i], n.realization=40)) ) )
  99 #' # Mix series so that all groups are evenly spread
 100 #' permut <- (0:239)%%6 * 40 + (0:239)%/%6 + 1
 101 #' series = series[,permut]
 102 #' #dim(series) #c(240,1001)
 103 #' res_ascii <- claws(series, K1=30, K2=6, 100, random=FALSE, verbose=TRUE)
 104 #'
 105 #' # Same example, from CSV file
 106 #' csv_file <- tempfile(pattern="epclust_series.csv_")
 107 #' write.table(t(series), csv_file, sep=",", row.names=FALSE, col.names=FALSE)
 108 #' res_csv <- claws(csv_file, K1=30, K2=6, 100, random=FALSE)
 109 #'
 110 #' # Same example, from binary file
 111 #' bin_file <- tempfile(pattern="epclust_series.bin_")
 112 #' nbytes <- 8
 113 #' endian <- "little"
 114 #' binarize(csv_file, bin_file, 500, ",", nbytes, endian)
 115 #' getSeries <- function(indices) getDataInFile(indices, bin_file, nbytes, endian)
 116 #' res_bin <- claws(getSeries, K1=30, K2=6, 100, random=FALSE)
 117 #' unlink(csv_file)
 118 #' unlink(bin_file)
 119 #'
 120 #' # Same example, from SQLite database
 121 #' library(DBI)
 122 #' series_db <- dbConnect(RSQLite::SQLite(), "file::memory:")
 123 #' # Prepare data.frame in DB-format
 124 #' n <- ncol(series)
 125 #' times_values <- data.frame(
 126 #'   id = rep(1:n,each=L),
 127 #'   time = rep( as.POSIXct(1800*(1:L),"GMT",origin="2001-01-01"), n ),
 128 #'   value = as.double(series) )
 129 #' dbWriteTable(series_db, "times_values", times_values)
 130 #' # Fill associative array, map index to identifier
 131 #' indexToID_inDB <- as.character(
 132 #'   dbGetQuery(series_db, 'SELECT DISTINCT id FROM times_values')[,"id"] )
 133 #' serie_length <- as.integer( dbGetQuery(series_db,
 134 #'   paste("SELECT COUNT(*) FROM times_values WHERE id == ",indexToID_inDB[1],sep="")) )
 135 #' getSeries <- function(indices) {
 136 #'   request <- "SELECT id,value FROM times_values WHERE id in ("
 137 #'   for (i in seq_along(indices)) {
 138 #'     request <- paste(request, indexToID_inDB[ indices[i] ],  sep="")
 139 #'     if (i < length(indices))
 140 #'       request <- paste(request, ",", sep="")
 141 #'   }
 142 #'   request <- paste(request, ")", sep="")
 143 #'   df_series <- dbGetQuery(series_db, request)
 144 #'   if (nrow(df_series) >= 1)
 145 #'     matrix(df_series[,"value"], nrow=serie_length)
 146 #'   else
 147 #'     NULL
 148 #' }
 149 #' # reuse_bin==FALSE: DB do not garantee ordering
 150 #' res_db <- claws(getSeries, K1=30, K2=6, 100, random=FALSE, reuse_bin=FALSE)
 151 #' dbDisconnect(series_db)
 152 #'
 153 #' # All results should be equal:
 154 #' all(res_ascii$ranks == res_csv$ranks
 155 #'   & res_ascii$ranks == res_bin$ranks
 156 #'   & res_ascii$ranks == res_db$ranks)
 157 #' }
 158 #' @export
 159 claws <- function(series, K1, K2, nb_series_per_chunk, nb_items_clust=7*K1,
 160     algoClust1=function(data,K) cluster::pam(t(data),K,diss=FALSE,pamonce=1)$id.med,
 161     algoClust2=function(dists,K) cluster::pam(dists,K,diss=TRUE,pamonce=1)$id.med,
 162     wav_filt="d8", contrib_type="absolute", WER="end", smooth_lvl=3, nvoice=4,
 163     random=TRUE, ntasks=1, ncores_tasks=1, ncores_clust=3, sep=",", nbytes=4,
 164     endian=.Platform$endian, verbose=FALSE, parll=TRUE, reuse_bin=TRUE)
 165 {
 166
 167
 168 #TODO: comprendre differences.......... debuguer getSeries for DB
 169
 170
 171     # Check/transform arguments
 172     if (!is.matrix(series) && !bigmemory::is.big.matrix(series)
 173         && !is.function(series)
 174         && !methods::is(series,"connection") && !is.character(series))
 175     {
 176         stop("'series': [big]matrix, function, file or valid connection (no NA)")
 177     }
 178     K1 <- .toInteger(K1, function(x) x>=2)
 179     K2 <- .toInteger(K2, function(x) x>=2)
 180     nb_series_per_chunk <- .toInteger(nb_series_per_chunk, function(x) x>=1)
 181     nb_items_clust <- .toInteger(nb_items_clust, function(x) x>K1)
 182     random <- .toLogical(random)
 183     tryCatch({ignored <- wavelets::wt.filter(wav_filt)},
 184         error=function(e) stop("Invalid wavelet filter; see ?wavelets::wt.filter") )
 185     ctypes <- c("relative","absolute","logit")
 186     contrib_type <- ctypes[ pmatch(contrib_type,ctypes) ]
 187     if (is.na(contrib_type))
 188         stop("'contrib_type' in {'relative','absolute','logit'}")
 189     if (WER!="end" && WER!="mix")
 190         stop("'WER': in {'end','mix'}")
 191     random <- .toLogical(random)
 192     ntasks <- .toInteger(ntasks, function(x) x>=1)
 193     ncores_tasks <- .toInteger(ncores_tasks, function(x) x>=1)
 194     ncores_clust <- .toInteger(ncores_clust, function(x) x>=1)
 195     if (!is.character(sep))
 196         stop("'sep': character")
 197     nbytes <- .toInteger(nbytes, function(x) x==4 || x==8)
 198     verbose <- .toLogical(verbose)
 199     parll <- .toLogical(parll)
 200
 201     # Binarize series if it is not a function; the aim is to always use a function,
 202     # to uniformize treatments. An equally good alternative would be to use a file-backed
 203     # bigmemory::big.matrix, but it would break the "all-is-function" pattern.
 204     if (!is.function(series))
 205     {
 206         if (verbose)
 207             cat("...Serialize time-series (or retrieve past binary file)\n")
 208         series_file <- ".series.epclust.bin"
 209         if (!file.exists(series_file) || !reuse_bin)
 210         {
 211             unlink(series_file,)
 212             binarize(series, series_file, nb_series_per_chunk, sep, nbytes, endian)
 213         }
 214         getSeries <- function(inds) getDataInFile(inds, series_file, nbytes, endian)
 215     }
 216     else
 217         getSeries <- series
 218
 219     # Serialize all computed wavelets contributions into a file
 220     contribs_file <- ".contribs.epclust.bin"
 221     if (verbose)
 222         cat("...Compute contributions and serialize them (or retrieve past binary file)\n")
 223     if (!file.exists(contribs_file) || !reuse_bin)
 224     {
 225         unlink(contribs_file,)
 226         nb_curves <- binarizeTransform(getSeries,
 227             function(curves) curvesToContribs(curves, wav_filt, contrib_type),
 228             contribs_file, nb_series_per_chunk, nbytes, endian)
 229     }
 230     else
 231     {
 232         # TODO: duplicate from getDataInFile() in de_serialize.R
 233         contribs_size <- file.info(contribs_file)$size #number of bytes in the file
 234         contrib_length <- readBin(contribs_file, "integer", n=1, size=8, endian=endian)
 235         nb_curves <- (contribs_size-8) / (nbytes*contrib_length)
 236     }
 237     getContribs <- function(indices) getDataInFile(indices, contribs_file, nbytes, endian)
 238
 239     # A few sanity checks: do not continue if too few data available.
 240     if (nb_curves < K2)
 241         stop("Not enough data: less series than final number of clusters")
 242     nb_series_per_task <- round(nb_curves / ntasks)
 243     if (nb_series_per_task < K2)
 244         stop("Too many tasks: less series in one task than final number of clusters")
 245
 246     # Generate a random permutation of 1:N (if random==TRUE);
 247     # otherwise just use arrival (storage) order.
 248     indices_all <- if (random) sample(nb_curves) else seq_len(nb_curves)
 249     # Split (all) indices into ntasks groups of ~same size
 250     indices_tasks <- lapply(seq_len(ntasks), function(i) {
 251         upper_bound <- ifelse( i<ntasks, min(nb_series_per_task*i,nb_curves), nb_curves )
 252         indices_all[((i-1)*nb_series_per_task+1):upper_bound]
 253     })
 254
 255     if (parll && ntasks>1)
 256     {
 257         # Initialize parallel runs: outfile="" allow to output verbose traces in the console
 258         # under Linux. All necessary variables are passed to the workers.
 259         cl <-
 260             if (verbose)
 261                 parallel::makeCluster(ncores_tasks, outfile="")
 262             else
 263                 parallel::makeCluster(ncores_tasks)
 264         varlist <- c("ncores_clust","verbose","parll", #task 1 & 2
 265             "K1","getContribs","algoClust1","nb_items_clust") #task 1
 266         if (WER=="mix")
 267         {
 268             # Add variables for task 2
 269             varlist <- c(varlist, "K2","getSeries","algoClust2","nb_series_per_chunk",
 270                 "smooth_lvl","nvoice","nbytes","endian")
 271         }
 272         parallel::clusterExport(cl, varlist, envir <- environment())
 273     }
 274
 275     # This function achieves one complete clustering task, divided in stage 1 + stage 2.
 276     # stage 1: n indices  --> clusteringTask1(...) --> K1 medoids (indices)
 277     # stage 2: K1 indices --> K1xK1 WER distances --> clusteringTask2(...) --> K2 medoids,
 278     # where n == N / ntasks, N being the total number of curves.
 279     runTwoStepClustering <- function(inds)
 280     {
 281         # When running in parallel, the environment is blank: we need to load the required
 282         # packages, and pass useful variables.
 283         if (parll && ntasks>1)
 284             require("epclust", quietly=TRUE)
 285         indices_medoids <- clusteringTask1(inds, getContribs, K1, algoClust1,
 286             nb_items_clust, ncores_clust, verbose, parll)
 287         if (WER=="mix")
 288         {
 289             indices_medoids <- clusteringTask2(indices_medoids, getSeries, K2, algoClust2,
 290                 nb_series_per_chunk,smooth_lvl,nvoice,nbytes,endian,ncores_clust,verbose,parll)
 291         }
 292         indices_medoids
 293     }
 294
 295     if (verbose)
 296     {
 297         message <- paste("...Run ",ntasks," x stage 1", sep="")
 298         if (WER=="mix")
 299             message <- paste(message," + stage 2", sep="")
 300         cat(paste(message,"\n", sep=""))
 301     }
 302
 303     # As explained above, we obtain after all runs ntasks*[K1 or K2] medoids indices,
 304     # depending wether WER=="end" or "mix", respectively.
 305     indices_medoids_all <-
 306         if (parll && ntasks>1)
 307             unlist( parallel::parLapply(cl, indices_tasks, runTwoStepClustering) )
 308         else
 309             unlist( lapply(indices_tasks, runTwoStepClustering) )
 310
 311     if (parll && ntasks>1)
 312         parallel::stopCluster(cl)
 313
 314     # For the last stage, ncores_tasks*(ncores_clusts+1) cores should be available:
 315     #  - ntasks for level 1 parallelism
 316     #  - ntasks*ncores_clust for level 2 parallelism,
 317     # but since an extension MPI <--> tasks / OpenMP <--> sub-tasks is on the way,
 318     # it's better to just re-use ncores_clust
 319     ncores_last_stage <- ncores_clust
 320
 321     # Run last clustering tasks to obtain only K2 medoids indices
 322     if (verbose)
 323         cat("...Run final // stage 1 + stage 2\n")
 324     indices_medoids <- clusteringTask1(indices_medoids_all, getContribs, K1, algoClust1,
 325         nb_items_clust, ncores_tasks*ncores_clust, verbose, parll)
 326
 327     indices_medoids <- clusteringTask2(indices_medoids, getSeries, K2, algoClust2,
 328         nb_series_per_chunk,smooth_lvl,nvoice,nbytes,endian,ncores_last_stage,verbose,parll)
 329
 330     # Compute synchrones, that is to say the cumulated power consumptions for each of the K2
 331     # final groups.
 332     medoids <- getSeries(indices_medoids)
 333     synchrones <- computeSynchrones(medoids, getSeries, nb_curves, nb_series_per_chunk,
 334         ncores_last_stage, verbose, parll)
 335
 336     # NOTE: no need to use big.matrix here, since there are only K2 << K1 << N remaining curves
 337     list("medoids"=medoids, "ranks"=indices_medoids, "synchrones"=synchrones)
 338 }