Commit | Line | Data |
---|---|---|
8702eb86 | 1 | #' CLAWS: CLustering with wAvelets and Wer distanceS |
7f0781b7 | 2 | #' |
eef6f6c9 BA |
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 | #' | |
d9bb53c5 BA |
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: | |
eef6f6c9 | 12 | #' \enumerate{ |
d9bb53c5 | 13 | #' \item iterate the first clustering algorithm on its aggregated outputs, |
40f12a2f | 14 | #' on inputs of size \code{nb_series_per_chunk} |
d9bb53c5 BA |
15 | #' \item optionally, if WER=="mix": |
16 | #' a) compute the K1 synchrones curves, | |
40f12a2f BA |
17 | #' a) compute WER distances (K1xK1 matrix) between medoids and |
18 | #' b) apply the second clustering algorithm (output: K2 indices) | |
eef6f6c9 | 19 | #' } |
d9bb53c5 | 20 | #' \item Launch a final task on the aggregated outputs of all previous tasks: |
40f12a2f BA |
21 | #' ntasks*K1 if WER=="end", ntasks*K2 otherwise |
22 | #' \item Compute synchrones (sum of series within each final group) | |
d9bb53c5 BA |
23 | #' } |
24 | #' \cr | |
40f12a2f BA |
25 | #' The main argument -- \code{series} -- has a quite misleading name, since it can be |
26 | #' either a [big.]matrix, a CSV file, a connection or a user function to retrieve series. | |
27 | #' When \code{series} is given as a function, it must take a single argument, | |
d9bb53c5 | 28 | #' 'indices', integer vector equal to the indices of the curves to retrieve; |
40f12a2f | 29 | #' see SQLite example. |
a52836b2 | 30 | #' WARNING: the return value must be a matrix (in columns), or NULL if no matches. |
d9bb53c5 BA |
31 | #' \cr |
32 | #' Note: Since we don't make assumptions on initial data, there is a possibility that | |
40f12a2f | 33 | #' even when serialized, contributions do not fit in RAM. For example, |
d9bb53c5 BA |
34 | #' 30e6 series of length 100,000 would lead to a +4Go contribution matrix. Therefore, |
35 | #' it's safer to place these in (binary) files; that's what we do. | |
7f0781b7 | 36 | #' |
40f12a2f | 37 | #' @param series Access to the (time-)series, which can be of one of the three |
8702eb86 BA |
38 | #' following types: |
39 | #' \itemize{ | |
eef6f6c9 | 40 | #' \item [big.]matrix: each column contains the (time-ordered) values of one time-serie |
bf5c0844 BA |
41 | #' \item connection: any R connection object providing lines as described above |
42 | #' \item character: name of a CSV file containing series in rows (no header) | |
8702eb86 | 43 | #' \item function: a custom way to retrieve the curves; it has only one argument: |
eef6f6c9 | 44 | #' the indices of the series to be retrieved. See SQLite example |
8702eb86 | 45 | #' } |
eef6f6c9 | 46 | #' @param K1 Number of clusters to be found after stage 1 (K1 << N [number of series]) |
1c6f223e | 47 | #' @param K2 Number of clusters to be found after stage 2 (K2 << K1) |
40f12a2f BA |
48 | #' @param nb_series_per_chunk (Maximum) number of series to retrieve in one batch; |
49 | #' this value is also used for the (maximum) number of series to cluster at a time | |
0486fbad | 50 | #' @param algoClust1 Clustering algorithm for stage 1. A function which takes (data, K) |
2b9f5356 | 51 | #' as argument where data is a matrix in columns and K the desired number of clusters, |
9f05a4a0 BA |
52 | #' and outputs K medoids ranks. Default: PAM. In our method, this function is called |
53 | #' on iterated medoids during stage 1 | |
0486fbad | 54 | #' @param algoClust2 Clustering algorithm for stage 2. A function which takes (dists, K) |
2b9f5356 | 55 | #' as argument where dists is a matrix of distances and K the desired number of clusters, |
9f05a4a0 | 56 | #' and outputs K medoids ranks. Default: PAM. In our method, this function is called |
40f12a2f | 57 | #' on a matrix of K1 x K1 (WER) distances computed between medoids after algorithm 1 |
eef6f6c9 BA |
58 | #' @param wav_filt Wavelet transform filter; see ?wavelets::wt.filter |
59 | #' @param contrib_type Type of contribution: "relative", "logit" or "absolute" (any prefix) | |
60 | #' @param WER "end" to apply stage 2 after stage 1 has fully iterated, or "mix" to apply | |
61 | #' stage 2 at the end of each task | |
a52836b2 | 62 | #' @param nvoice Number of voices within each octave for CWT computations |
4bcfdbee | 63 | #' @param random TRUE (default) for random chunks repartition |
eef6f6c9 BA |
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 to disable: sequential tasks) | |
40f12a2f | 68 | #' @param ncores_clust Number of parallel clusterings in one task (3 should be a minimum) |
4bcfdbee | 69 | #' @param sep Separator in CSV input file (if any provided) |
8702eb86 | 70 | #' @param nbytes Number of bytes to serialize a floating-point number; 4 or 8 |
eef6f6c9 | 71 | #' @param endian Endianness for (de)serialization ("little" or "big") |
4bcfdbee | 72 | #' @param verbose Level of verbosity (0/FALSE for nothing or 1/TRUE for all; devel stage) |
492cd9e7 | 73 | #' @param parll TRUE to fully parallelize; otherwise run sequentially (debug, comparison) |
7f0781b7 | 74 | #' |
40f12a2f BA |
75 | #' @return A list with |
76 | #' \itemize{ | |
77 | #' medoids: a matrix of the final K2 medoids curves, in columns | |
78 | #' ranks: corresponding indices in the dataset | |
79 | #' synchrones: a matrix of the K2 sum of series within each final group | |
80 | #' } | |
eef6f6c9 BA |
81 | #' |
82 | #' @references Clustering functional data using Wavelets [2013]; | |
83 | #' A. Antoniadis, X. Brossat, J. Cugliari & J.-M. Poggi. | |
84 | #' Inter. J. of Wavelets, Multiresolution and Information Procesing, | |
85 | #' vol. 11, No 1, pp.1-30. doi:10.1142/S0219691313500033 | |
1c6f223e BA |
86 | #' |
87 | #' @examples | |
4efef8cc | 88 | #' \dontrun{ |
eef6f6c9 | 89 | #' # WER distances computations are too long for CRAN (for now) |
4efef8cc BA |
90 | #' |
91 | #' # Random series around cos(x,2x,3x)/sin(x,2x,3x) | |
92 | #' x = seq(0,500,0.05) | |
93 | #' L = length(x) #10001 | |
eef6f6c9 | 94 | #' ref_series = matrix( c(cos(x),cos(2*x),cos(3*x),sin(x),sin(2*x),sin(3*x)), ncol=6 ) |
4efef8cc | 95 | #' library(wmtsa) |
eef6f6c9 | 96 | #' series = do.call( cbind, lapply( 1:6, function(i) |
a52836b2 | 97 | #' do.call(cbind, wmtsa::wavBootstrap(ref_series[,i], n.realization=400)) ) ) |
4efef8cc | 98 | #' #dim(series) #c(2400,10001) |
40f12a2f | 99 | #' res_ascii = claws(series, K1=60, K2=6, 200, verbose=TRUE) |
4efef8cc BA |
100 | #' |
101 | #' # Same example, from CSV file | |
102 | #' csv_file = "/tmp/epclust_series.csv" | |
103 | #' write.table(series, csv_file, sep=",", row.names=FALSE, col.names=FALSE) | |
40f12a2f | 104 | #' res_csv = claws(csv_file, K1=60, K2=6, 200) |
4efef8cc BA |
105 | #' |
106 | #' # Same example, from binary file | |
eef6f6c9 BA |
107 | #' bin_file <- "/tmp/epclust_series.bin" |
108 | #' nbytes <- 8 | |
109 | #' endian <- "little" | |
110 | #' binarize(csv_file, bin_file, 500, nbytes, endian) | |
111 | #' getSeries <- function(indices) getDataInFile(indices, bin_file, nbytes, endian) | |
40f12a2f | 112 | #' res_bin <- claws(getSeries, K1=60, K2=6, 200) |
4efef8cc BA |
113 | #' unlink(csv_file) |
114 | #' unlink(bin_file) | |
115 | #' | |
116 | #' # Same example, from SQLite database | |
117 | #' library(DBI) | |
118 | #' series_db <- dbConnect(RSQLite::SQLite(), "file::memory:") | |
119 | #' # Prepare data.frame in DB-format | |
eef6f6c9 BA |
120 | #' n <- nrow(series) |
121 | #' time_values <- data.frame( | |
4bcfdbee BA |
122 | #' id = rep(1:n,each=L), |
123 | #' time = rep( as.POSIXct(1800*(0:n),"GMT",origin="2001-01-01"), L ), | |
124 | #' value = as.double(t(series)) ) | |
4efef8cc | 125 | #' dbWriteTable(series_db, "times_values", times_values) |
4bcfdbee BA |
126 | #' # Fill associative array, map index to identifier |
127 | #' indexToID_inDB <- as.character( | |
128 | #' dbGetQuery(series_db, 'SELECT DISTINCT id FROM time_values')[,"id"] ) | |
eef6f6c9 BA |
129 | #' serie_length <- as.integer( dbGetQuery(series_db, |
130 | #' paste("SELECT COUNT * FROM time_values WHERE id == ",indexToID_inDB[1],sep="")) ) | |
131 | #' getSeries <- function(indices) { | |
132 | #' request <- "SELECT id,value FROM times_values WHERE id in (" | |
4bcfdbee | 133 | #' for (i in indices) |
eef6f6c9 BA |
134 | #' request <- paste(request, indexToID_inDB[i], ",", sep="") |
135 | #' request <- paste(request, ")", sep="") | |
136 | #' df_series <- dbGetQuery(series_db, request) | |
a52836b2 BA |
137 | #' if (length(df_series) >= 1) |
138 | #' as.matrix(df_series[,"value"], nrow=serie_length) | |
139 | #' else | |
140 | #' NULL | |
4efef8cc | 141 | #' } |
40f12a2f | 142 | #' res_db = claws(getSeries, K1=60, K2=6, 200)) |
4bcfdbee BA |
143 | #' dbDisconnect(series_db) |
144 | #' | |
40f12a2f BA |
145 | #' # All results should be the same: |
146 | #' library(digest) | |
147 | #' digest::sha1(res_ascii) | |
148 | #' digest::sha1(res_csv) | |
149 | #' digest::sha1(res_bin) | |
150 | #' digest::sha1(res_db) | |
1c6f223e | 151 | #' } |
1c6f223e | 152 | #' @export |
40f12a2f BA |
153 | claws <- function(getSeries, K1, K2, nb_series_per_chunk, |
154 | algoClust1=function(data,K) cluster::pam(t(data),K,diss=FALSE,pamonce=1)$id.med, | |
155 | algoClust2=function(dists,K) cluster::pam(dists,K,diss=TRUE,pamonce=1)$id.med, | |
a52836b2 | 156 | wav_filt="d8", contrib_type="absolute", WER="end", nvoice=4, random=TRUE, |
d9bb53c5 BA |
157 | ntasks=1, ncores_tasks=1, ncores_clust=4, sep=",", nbytes=4, |
158 | endian=.Platform$endian, verbose=FALSE, parll=TRUE) | |
ac1d4231 | 159 | { |
0e2dce80 | 160 | # Check/transform arguments |
40f12a2f BA |
161 | if (!is.matrix(series) && !bigmemory::is.big.matrix(series) |
162 | && !is.function(series) | |
163 | && !methods::is(series,"connection") && !is.character(series)) | |
0e2dce80 | 164 | { |
40f12a2f | 165 | stop("'series': [big]matrix, function, file or valid connection (no NA)") |
5c652979 | 166 | } |
eef6f6c9 BA |
167 | K1 <- .toInteger(K1, function(x) x>=2) |
168 | K2 <- .toInteger(K2, function(x) x>=2) | |
37c82bba BA |
169 | nb_series_per_chunk <- .toInteger(nb_series_per_chunk, function(x) x>=1) |
170 | # K1 (number of clusters at step 1) cannot exceed nb_series_per_chunk, because we will need | |
171 | # to load K1 series in memory for clustering stage 2. | |
172 | if (K1 > nb_series_per_chunk) | |
173 | stop("'K1' cannot exceed 'nb_series_per_chunk'") | |
eef6f6c9 | 174 | random <- .toLogical(random) |
0486fbad BA |
175 | tryCatch( {ignored <- wavelets::wt.filter(wav_filt)}, |
176 | error = function(e) stop("Invalid wavelet filter; see ?wavelets::wt.filter") ) | |
eef6f6c9 BA |
177 | ctypes = c("relative","absolute","logit") |
178 | contrib_type = ctypes[ pmatch(contrib_type,ctypes) ] | |
179 | if (is.na(contrib_type)) | |
180 | stop("'contrib_type' in {'relative','absolute','logit'}") | |
7f0781b7 | 181 | if (WER!="end" && WER!="mix") |
eef6f6c9 BA |
182 | stop("'WER': in {'end','mix'}") |
183 | random <- .toLogical(random) | |
184 | ntasks <- .toInteger(ntasks, function(x) x>=1) | |
185 | ncores_tasks <- .toInteger(ncores_tasks, function(x) x>=1) | |
186 | ncores_clust <- .toInteger(ncores_clust, function(x) x>=1) | |
56857861 BA |
187 | if (!is.character(sep)) |
188 | stop("'sep': character") | |
eef6f6c9 BA |
189 | nbytes <- .toInteger(nbytes, function(x) x==4 || x==8) |
190 | verbose <- .toLogical(verbose) | |
191 | parll <- .toLogical(parll) | |
56857861 | 192 | |
40f12a2f | 193 | # Binarize series if it is not a function; the aim is to always use a function, |
2b9f5356 | 194 | # to uniformize treatments. An equally good alternative would be to use a file-backed |
d9bb53c5 | 195 | # bigmemory::big.matrix, but it would break the "all-is-function" pattern. |
40f12a2f | 196 | if (!is.function(series)) |
56857861 | 197 | { |
4bcfdbee | 198 | if (verbose) |
a52836b2 | 199 | cat("...Serialize time-series (or retrieve past binary file)\n") |
40f12a2f | 200 | series_file = ".series.epclust.bin" |
a52836b2 | 201 | if (!file.exists(series_file)) |
40f12a2f | 202 | binarize(series, series_file, nb_series_per_chunk, sep, nbytes, endian) |
4bcfdbee | 203 | getSeries = function(inds) getDataInFile(inds, series_file, nbytes, endian) |
56857861 | 204 | } |
40f12a2f BA |
205 | else |
206 | getSeries = series | |
ac1d4231 | 207 | |
95b5c2e6 | 208 | # Serialize all computed wavelets contributions into a file |
40f12a2f | 209 | contribs_file = ".contribs.epclust.bin" |
7f0781b7 | 210 | index = 1 |
cea14f3a | 211 | nb_curves = 0 |
4bcfdbee | 212 | if (verbose) |
a52836b2 BA |
213 | cat("...Compute contributions and serialize them (or retrieve past binary file)\n") |
214 | if (!file.exists(contribs_file)) | |
215 | { | |
216 | nb_curves = binarizeTransform(getSeries, | |
40f12a2f | 217 | function(curves) curvesToContribs(curves, wav_filt, contrib_type), |
a52836b2 BA |
218 | contribs_file, nb_series_per_chunk, nbytes, endian) |
219 | } | |
220 | else | |
221 | { | |
222 | # TODO: duplicate from getDataInFile() in de_serialize.R | |
223 | contribs_size = file.info(contribs_file)$size #number of bytes in the file | |
224 | contrib_length = readBin(contribs_file, "integer", n=1, size=8, endian=endian) | |
225 | nb_curves = (contribs_size-8) / (nbytes*contrib_length) | |
226 | } | |
4bcfdbee | 227 | getContribs = function(indices) getDataInFile(indices, contribs_file, nbytes, endian) |
8e6accca | 228 | |
2b9f5356 | 229 | # A few sanity checks: do not continue if too few data available. |
eef6f6c9 BA |
230 | if (nb_curves < K2) |
231 | stop("Not enough data: less series than final number of clusters") | |
5c652979 | 232 | nb_series_per_task = round(nb_curves / ntasks) |
eef6f6c9 BA |
233 | if (nb_series_per_task < K2) |
234 | stop("Too many tasks: less series in one task than final number of clusters") | |
ac1d4231 | 235 | |
d9bb53c5 BA |
236 | # Generate a random permutation of 1:N (if random==TRUE); |
237 | # otherwise just use arrival (storage) order. | |
2b9f5356 BA |
238 | indices_all = if (random) sample(nb_curves) else seq_len(nb_curves) |
239 | # Split (all) indices into ntasks groups of ~same size | |
240 | indices_tasks = lapply(seq_len(ntasks), function(i) { | |
241 | upper_bound = ifelse( i<ntasks, min(nb_series_per_task*i,nb_curves), nb_curves ) | |
242 | indices_all[((i-1)*nb_series_per_task+1):upper_bound] | |
243 | }) | |
244 | ||
245 | if (parll && ntasks>1) | |
246 | { | |
247 | # Initialize parallel runs: outfile="" allow to output verbose traces in the console | |
248 | # under Linux. All necessary variables are passed to the workers. | |
249 | cl = parallel::makeCluster(ncores_tasks, outfile="") | |
40f12a2f BA |
250 | parallel::clusterExport(cl, envir = environment(), |
251 | varlist = c("getSeries","getContribs","K1","K2","algoClust1","algoClust2", | |
252 | "nb_series_per_chunk","ncores_clust","nvoice","nbytes","endian","verbose","parll")) | |
2b9f5356 BA |
253 | } |
254 | ||
255 | # This function achieves one complete clustering task, divided in stage 1 + stage 2. | |
256 | # stage 1: n indices --> clusteringTask1(...) --> K1 medoids | |
257 | # stage 2: K1 medoids --> clusteringTask2(...) --> K2 medoids, | |
258 | # where n = N / ntasks, N being the total number of curves. | |
492cd9e7 BA |
259 | runTwoStepClustering = function(inds) |
260 | { | |
d9bb53c5 | 261 | # When running in parallel, the environment is blank: we need to load the required |
2b9f5356 | 262 | # packages, and pass useful variables. |
bf5c0844 | 263 | if (parll && ntasks>1) |
492cd9e7 | 264 | require("epclust", quietly=TRUE) |
40f12a2f BA |
265 | indices_medoids = clusteringTask1(inds, getContribs, K1, algoClust1, |
266 | nb_series_per_chunk, ncores_clust, verbose, parll) | |
267 | if (WER=="mix") | |
56857861 | 268 | { |
40f12a2f | 269 | indices_medoids = clusteringTask2(indices_medoids, getSeries, K2, algoClust2, |
a52836b2 | 270 | nb_series_per_chunk, nvoice, nbytes, endian, ncores_clust, verbose, parll) |
56857861 BA |
271 | } |
272 | indices_medoids | |
492cd9e7 BA |
273 | } |
274 | ||
c45fd663 | 275 | if (verbose) |
e161499b BA |
276 | { |
277 | message = paste("...Run ",ntasks," x stage 1", sep="") | |
278 | if (WER=="mix") | |
279 | message = paste(message," + stage 2", sep="") | |
280 | cat(paste(message,"\n", sep="")) | |
281 | } | |
c45fd663 | 282 | |
2b9f5356 | 283 | # As explained above, indices will be assigned to ntasks*K1 medoids indices [if WER=="end"], |
d9bb53c5 | 284 | # or nothing (empty vector) if WER=="mix"; in this case, synchrones are stored in a file. |
40f12a2f | 285 | indices_medoids_all <- |
eef6f6c9 BA |
286 | if (parll && ntasks>1) |
287 | unlist( parallel::parLapply(cl, indices_tasks, runTwoStepClustering) ) | |
288 | else | |
289 | unlist( lapply(indices_tasks, runTwoStepClustering) ) | |
40f12a2f | 290 | |
bf5c0844 | 291 | if (parll && ntasks>1) |
492cd9e7 | 292 | parallel::stopCluster(cl) |
3465b246 | 293 | |
40f12a2f BA |
294 | # Right before the final stage, input data still is the initial set of curves, referenced |
295 | # by the ntasks*[K1 or K2] medoids indices. | |
0e2dce80 | 296 | |
40f12a2f BA |
297 | # Run last clustering tasks to obtain only K2 medoids indices, from ntasks*[K1 or K2] |
298 | # indices, depending wether WER=="end" or WER=="mix" | |
4bcfdbee BA |
299 | if (verbose) |
300 | cat("...Run final // stage 1 + stage 2\n") | |
40f12a2f | 301 | indices_medoids = clusteringTask1(indices_medoids_all, getContribs, K1, algoClust1, |
0486fbad | 302 | nb_series_per_chunk, ncores_tasks*ncores_clust, verbose, parll) |
40f12a2f | 303 | indices_medoids = clusteringTask2(indices_medoids, getContribs, K2, algoClust2, |
a52836b2 | 304 | nb_series_per_chunk, nvoice, nbytes, endian, ncores_tasks*ncores_clust, verbose, parll) |
4bcfdbee | 305 | |
40f12a2f BA |
306 | # Compute synchrones |
307 | medoids = getSeries(indices_medoids) | |
308 | synchrones = computeSynchrones(medoids, getSeries, nb_curves, nb_series_per_chunk, | |
309 | ncores_tasks*ncores_clust, verbose, parll) | |
4bcfdbee | 310 | |
40f12a2f BA |
311 | # NOTE: no need to use big.matrix here, since there are only K2 << K1 << N remaining curves |
312 | list("medoids"=medoids, "ranks"=indices_medoids, "synchrones"=synchrones) | |
cea14f3a | 313 | } |