[epclust.git] / epclust / R / clustering.R

#' @name clustering
#' @rdname clustering
#' @aliases clusteringTask1 clusteringTask2 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
#'   first clustering algorithm on a contributions matrix, while the latter clusters
#'   a set of series inside one task (~nb_items_clust)
#'
#' @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_per_chunk, nb_items_clust, ncores_clust=1,
	verbose=FALSE, parll=TRUE)
{
	if (verbose)
		cat(paste("*** Clustering task 1 on ",length(indices)," lines\n", sep=""))


##TODO: reviser le spreadIndices, tenant compte de nb_items_clust

	##TODO: reviser / harmoniser avec getContribs qui en récupère pt'et + pt'et - !!


	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)
			{
				unlist( parallel::parLapply(cl, indices_workers, function(inds) {
					require("epclust", quietly=TRUE)
					inds[ computeClusters1(getContribs(inds), K1, verbose) ]
				}) )
			}
			else
			{
				unlist( lapply(indices_workers, function(inds)
					inds[ computeClusters1(getContribs(inds), K1, verbose) ]
				) )
			}
	}
	if (parll)
		parallel::stopCluster(cl)

	indices #medoids
}

#' @rdname clustering
#' @export
clusteringTask2 = function(medoids, K2, getRefSeries, nb_ref_curves,
	nb_series_per_chunk, nbytes,endian,ncores_clust=1,verbose=FALSE,parll=TRUE)
{
	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, nbytes, endian, ncores_clust, verbose, parll)
	medoids[ computeClusters2(distances,K2,verbose), ]
}

#' @rdname clustering
#' @export
computeClusters1 = function(contribs, K1, verbose=FALSE)
{
	if (verbose)
		cat(paste("   computeClusters1() on ",nrow(contribs)," lines\n", sep=""))
	cluster::pam(        t(contribs)       , K1, diss=FALSE)$id.med
}

#' @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
#'
#' 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 L x K1 where L = length of a serie
#'
#' @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? ...or as arg?!
			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=L, ncol=K, 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]
	#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
	bigmemory::as.big.matrix(synchrones[,noNA_rows])
}

#' 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, nbytes,endian,ncores_clust=1,verbose=FALSE,parll=TRUE)
{
	if (verbose)
		cat(paste("--- Compute WER dists\n", sep=""))

	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(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 + 1)
	#condition: ( log2(s0*w0/(2*pi)) - 1 ) * nvoice + 1.5 >= 1
	s0 = 2
	w0 = 2*pi
	scaled=FALSE
	s0log = as.integer( (log2( s0*w0/(2*pi) ) - 1) * nvoice + 1.5 )
	totnoct = noctave + as.integer(s0log/nvoice) + 1

	Xwer_dist <- bigmemory::big.matrix(nrow=n, ncol=n, type="double")

	cwt_file = ".epclust_bin/cwt"
	#TODO: args, nb_per_chunk, nbytes, endian

	# 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)))
	}

	computeSaveCWT = 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,'*')
		res <- sqres / max(Mod(sqres))
		#TODO: serializer les CWT, les récupérer via getDataInFile ;
		#--> OK, faut juste stocker comme séries simples de taille delta*ncol (53*17519)
		binarize(c(as.double(Re(res)),as.double(Im(res))), cwt_file, ncol(res), ",", nbytes, endian)
	}

	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","getCWT"), envir=environment())
	}

	#precompute and serialize all CWT
	ignored <-
		if (parll)
			parallel::parLapply(cl, 1:n, computeSaveCWT)
		else
			lapply(1:n, computeSaveCWT)

	getCWT = function(index)
	{
		#from cwt_file ...
		res <- getDataInFile(c(2*index-1,2*index), cwt_file, nbytes, endian)
	###############TODO:
	}

	# 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)
		}

		i = pair[1] ; j = pair[2]
		if (verbose && j==i+1)
			cat(paste("   Distances (",i,",",j,"), (",i,",",j+1,") ...\n", sep=""))
		cwt_i <- getCWT(i)
		cwt_j <- getCWT(j)

		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.
	}

	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
	{
		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)
		{
			index = rem%%nb_workers + 1
			indices_workers[[index]] = c(indices_workers[[index]], indices[L-rem+1])
			rem = rem - 1
		}
	}
	indices_workers
}
Commit	Line	Data
4bcfdbee BA	1	#' @name clustering
4bcfdbee BA	2	#' @rdname clustering
eef6f6c9	3	#' @aliases clusteringTask1 clusteringTask2 computeClusters1 computeClusters2
4bcfdbee	4	#'
492cd9e7	5	#' @title Two-stage clustering, withing one task (see \code{claws()})
4bcfdbee	6	#'
492cd9e7 BA	7	#' @description \code{clusteringTask1()} runs one full stage-1 task, which consists in
492cd9e7 BA	8	#' iterated stage 1 clustering (on nb_curves / ntasks energy contributions, computed
bf5c0844 BA	9	#' through discrete wavelets coefficients).
	10	#' \code{clusteringTask2()} runs a full stage-2 task, which consists in synchrones
	11	#' and then WER distances computations, before applying the clustering algorithm.
	12	#' \code{computeClusters1()} and \code{computeClusters2()} correspond to the atomic
	13	#' clustering procedures respectively for stage 1 and 2. The former applies the
2b9f5356 BA	14	#' first clustering algorithm on a contributions matrix, while the latter clusters
2b9f5356 BA	15	#' a set of series inside one task (~nb_items_clust)
4bcfdbee BA	16	#'
	17	#' @param indices Range of series indices to cluster in parallel (initial data)
	18	#' @param getContribs Function to retrieve contributions from initial series indices:
	19	#' \code{getContribs(indices)} outpus a contributions matrix
	20	#' @param contribs matrix of contributions (e.g. output of \code{curvesToContribs()})
777c4b02	21	#' @param distances matrix of K1 x K1 (WER) distances between synchrones
4bcfdbee BA	22	#' @inheritParams computeSynchrones
	23	#' @inheritParams claws
	24	#'
492cd9e7	25	#' @return For \code{clusteringTask1()} and \code{computeClusters1()}, the indices of the
4bcfdbee	26	#' computed (K1) medoids. Indices are irrelevant for stage 2 clustering, thus
24ed5d83	27	#' \code{computeClusters2()} outputs a big.matrix of medoids
4bcfdbee BA	28	#' (of size limited by nb_series_per_chunk)
	29	NULL
	30
	31	#' @rdname clustering
	32	#' @export
492cd9e7	33	clusteringTask1 = function(
2b9f5356 BA	34	indices, getContribs, K1, nb_per_chunk, nb_items_clust, ncores_clust=1,
2b9f5356 BA	35	verbose=FALSE, parll=TRUE)
5c652979	36	{
492cd9e7	37	if (verbose)
e161499b	38	cat(paste("*** Clustering task 1 on ",length(indices)," lines\n", sep=""))
4bcfdbee	39
2b9f5356 BA	40
	41
	42
	43
	44
	45	##TODO: reviser le spreadIndices, tenant compte de nb_items_clust
	46
	47	##TODO: reviser / harmoniser avec getContribs qui en récupère pt'et + pt'et - !!
	48
	49
	50
492cd9e7	51	if (parll)
7b13d0c2	52	{
492cd9e7 BA	53	cl = parallel::makeCluster(ncores_clust)
492cd9e7 BA	54	parallel::clusterExport(cl, varlist=c("getContribs","K1","verbose"), envir=environment())
7b13d0c2	55	}
492cd9e7 BA	56	while (length(indices) > K1)
	57	{
	58	indices_workers = .spreadIndices(indices, nb_series_per_chunk)
e161499b BA	59	indices <-
	60	if (parll)
	61	{
	62	unlist( parallel::parLapply(cl, indices_workers, function(inds) {
	63	require("epclust", quietly=TRUE)
	64	inds[ computeClusters1(getContribs(inds), K1, verbose) ]
	65	}) )
	66	}
	67	else
	68	{
	69	unlist( lapply(indices_workers, function(inds)
	70	inds[ computeClusters1(getContribs(inds), K1, verbose) ]
	71	) )
	72	}
492cd9e7 BA	73	}
	74	if (parll)
	75	parallel::stopCluster(cl)
	76
56857861	77	indices #medoids
5c652979 BA	78	}
5c652979 BA	79
4bcfdbee BA	80	#' @rdname clustering
4bcfdbee BA	81	#' @export
a174b8ea BA	82	clusteringTask2 = function(medoids, K2, getRefSeries, nb_ref_curves,
a174b8ea BA	83	nb_series_per_chunk, nbytes,endian,ncores_clust=1,verbose=FALSE,parll=TRUE)
5c652979	84	{
e161499b BA	85	if (verbose)
	86	cat(paste("*** Clustering task 2 on ",nrow(medoids)," lines\n", sep=""))
	87
bf5c0844 BA	88	if (nrow(medoids) <= K2)
bf5c0844 BA	89	return (medoids)
492cd9e7 BA	90	synchrones = computeSynchrones(medoids,
492cd9e7 BA	91	getRefSeries, nb_ref_curves, nb_series_per_chunk, ncores_clust, verbose, parll)
a174b8ea	92	distances = computeWerDists(synchrones, nbytes, endian, ncores_clust, verbose, parll)
777c4b02	93	medoids[ computeClusters2(distances,K2,verbose), ]
5c652979 BA	94	}
5c652979 BA	95
bf5c0844 BA	96	#' @rdname clustering
bf5c0844 BA	97	#' @export
e161499b BA	98	computeClusters1 = function(contribs, K1, verbose=FALSE)
	99	{
	100	if (verbose)
	101	cat(paste(" computeClusters1() on ",nrow(contribs)," lines\n", sep=""))
eef6f6c9	102	cluster::pam( t(contribs) , K1, diss=FALSE)$id.med
e161499b	103	}
bf5c0844 BA	104
	105	#' @rdname clustering
	106	#' @export
e161499b BA	107	computeClusters2 = function(distances, K2, verbose=FALSE)
	108	{
	109	if (verbose)
	110	cat(paste(" computeClusters2() on ",nrow(distances)," lines\n", sep=""))
eef6f6c9	111	cluster::pam( distances , K2, diss=TRUE)$id.med
e161499b	112	}
bf5c0844	113
4bcfdbee BA	114	#' computeSynchrones
	115	#'
	116	#' Compute the synchrones curves (sum of clusters elements) from a matrix of medoids,
	117	#' using L2 distances.
	118	#'
24ed5d83	119	#' @param medoids big.matrix of medoids (curves of same length as initial series)
4bcfdbee BA	120	#' @param getRefSeries Function to retrieve initial series (e.g. in stage 2 after series
4bcfdbee BA	121	#' have been replaced by stage-1 medoids)
492cd9e7	122	#' @param nb_ref_curves How many reference series? (This number is known at this stage)
4bcfdbee BA	123	#' @inheritParams claws
4bcfdbee BA	124	#'
eef6f6c9	125	#' @return A big.matrix of size L x K1 where L = length of a serie
24ed5d83	126	#'
4bcfdbee	127	#' @export
492cd9e7 BA	128	computeSynchrones = function(medoids, getRefSeries,
492cd9e7 BA	129	nb_ref_curves, nb_series_per_chunk, ncores_clust=1,verbose=FALSE,parll=TRUE)
e205f218	130	{
e161499b BA	131	if (verbose)
	132	cat(paste("--- Compute synchrones\n", sep=""))
	133
492cd9e7	134	computeSynchronesChunk = function(indices)
3eef8d3d	135	{
363ae134 BA	136	if (parll)
	137	{
	138	require("bigmemory", quietly=TRUE)
6ad3f3fd	139	requireNamespace("synchronicity", quietly=TRUE)
363ae134 BA	140	require("epclust", quietly=TRUE)
363ae134 BA	141	synchrones <- bigmemory::attach.big.matrix(synchrones_desc)
6ad3f3fd	142	counts <- bigmemory::attach.big.matrix(counts_desc)
363ae134 BA	143	medoids <- bigmemory::attach.big.matrix(medoids_desc)
	144	m <- synchronicity::attach.mutex(m_desc)
	145	}
	146
6ad3f3fd BA	147	ref_series = getRefSeries(indices)
	148	nb_series = nrow(ref_series)
	149
363ae134	150	#get medoids indices for this chunk of series
2c14dbea	151	mi = computeMedoidsIndices(medoids@address, ref_series)
e161499b BA	152
e161499b BA	153	for (i in seq_len(nb_series))
56857861	154	{
492cd9e7 BA	155	if (parll)
492cd9e7 BA	156	synchronicity::lock(m)
eef6f6c9 BA	157	synchrones[, mi[i] ] = synchrones[, mi[i] ] + ref_series[,i]
eef6f6c9 BA	158	counts[ mi[i] ] = counts[ mi[i] ] + 1 #TODO: remove counts? ...or as arg?!
492cd9e7 BA	159	if (parll)
	160	synchronicity::unlock(m)
	161	}
	162	}
	163
e161499b	164	K = nrow(medoids) ; L = ncol(medoids)
492cd9e7	165	# Use bigmemory (shared==TRUE by default) + synchronicity to fill synchrones in //
24ed5d83	166	# TODO: if size > RAM (not our case), use file-backed big.matrix
eef6f6c9	167	synchrones = bigmemory::big.matrix(nrow=L, ncol=K, type="double", init=0.)
e161499b	168	counts = bigmemory::big.matrix(nrow=K, ncol=1, type="double", init=0)
24ed5d83	169	# synchronicity is only for Linux & MacOS; on Windows: run sequentially
492cd9e7 BA	170	parll = (requireNamespace("synchronicity",quietly=TRUE)
	171	&& parll && Sys.info()['sysname'] != "Windows")
	172	if (parll)
363ae134	173	{
492cd9e7	174	m <- synchronicity::boost.mutex()
363ae134 BA	175	m_desc <- synchronicity::describe(m)
363ae134 BA	176	synchrones_desc = bigmemory::describe(synchrones)
6ad3f3fd	177	counts_desc = bigmemory::describe(counts)
363ae134	178	medoids_desc = bigmemory::describe(medoids)
24ed5d83	179	cl = parallel::makeCluster(ncores_clust)
6ad3f3fd BA	180	parallel::clusterExport(cl, varlist=c("synchrones_desc","counts_desc","counts",
6ad3f3fd BA	181	"verbose","m_desc","medoids_desc","getRefSeries"), envir=environment())
24ed5d83 BA	182	}
24ed5d83 BA	183
492cd9e7	184	indices_workers = .spreadIndices(seq_len(nb_ref_curves), nb_series_per_chunk)
c45fd663	185	ignored <-
492cd9e7	186	if (parll)
e161499b	187	parallel::parLapply(cl, indices_workers, computeSynchronesChunk)
492cd9e7	188	else
c45fd663	189	lapply(indices_workers, computeSynchronesChunk)
492cd9e7	190
24ed5d83 BA	191	if (parll)
	192	parallel::stopCluster(cl)
	193
	194	#TODO: can we avoid this loop? ( synchrones = sweep(synchrones, 1, counts, '/') )
492cd9e7	195	for (i in seq_len(K))
eef6f6c9	196	synchrones[,i] = synchrones[,i] / counts[i]
3eef8d3d	197	#NOTE: odds for some clusters to be empty? (when series already come from stage 2)
8702eb86	198	# ...maybe; but let's hope resulting K1' be still quite bigger than K2
eef6f6c9	199	noNA_rows = sapply(seq_len(K), function(i) all(!is.nan(synchrones[,i])))
24ed5d83 BA	200	if (all(noNA_rows))
	201	return (synchrones)
	202	# Else: some clusters are empty, need to slice synchrones
eef6f6c9	203	bigmemory::as.big.matrix(synchrones[,noNA_rows])
e205f218	204	}
1c6f223e	205
4bcfdbee BA	206	#' computeWerDists
	207	#'
	208	#' Compute the WER distances between the synchrones curves (in rows), which are
	209	#' returned (e.g.) by \code{computeSynchrones()}
	210	#'
24ed5d83 BA	211	#' @param synchrones A big.matrix of synchrones, in rows. The series have same length
24ed5d83 BA	212	#' as the series in the initial dataset
492cd9e7	213	#' @inheritParams claws
4bcfdbee	214	#'
777c4b02	215	#' @return A matrix of size K1 x K1
24ed5d83	216	#'
4bcfdbee	217	#' @export
a174b8ea	218	computeWerDists = function(synchrones, nbytes,endian,ncores_clust=1,verbose=FALSE,parll=TRUE)
d03c0621	219	{
e161499b BA	220	if (verbose)
e161499b BA	221	cat(paste("--- Compute WER dists\n", sep=""))
24ed5d83	222
4bcfdbee BA	223	n <- nrow(synchrones)
4bcfdbee BA	224	delta <- ncol(synchrones)
db6fc17d	225	#TODO: automatic tune of all these parameters ? (for other users)
d03c0621	226	nvoice <- 4
4bcfdbee	227	# noctave = 2^13 = 8192 half hours ~ 180 days ; ~log2(ncol(synchrones))
d7d55bc1 BA	228	noctave = 13
d7d55bc1 BA	229	# 4 here represent 2^5 = 32 half-hours ~ 1 day
db6fc17d	230	#NOTE: default scalevector == 2^(0:(noctave * nvoice) / nvoice) * s0 (?)
24ed5d83	231	scalevector <- 2^(4:(noctave * nvoice) / nvoice + 1)
db6fc17d	232	#condition: ( log2(s0w0/(2pi)) - 1 ) * nvoice + 1.5 >= 1
a174b8ea BA	233	s0 = 2
a174b8ea BA	234	w0 = 2*pi
db6fc17d BA	235	scaled=FALSE
	236	s0log = as.integer( (log2( s0w0/(2pi) ) - 1) * nvoice + 1.5 )
	237	totnoct = noctave + as.integer(s0log/nvoice) + 1
	238
e161499b	239	Xwer_dist <- bigmemory::big.matrix(nrow=n, ncol=n, type="double")
e161499b	240
4204e877 BA	241	cwt_file = ".epclust_bin/cwt"
	242	#TODO: args, nb_per_chunk, nbytes, endian
	243
e161499b BA	244	# Generate n(n-1)/2 pairs for WER distances computations
e161499b BA	245	pairs = list()
4204e877 BA	246	V = seq_len(n)
4204e877 BA	247	for (i in 1:n)
e161499b BA	248	{
e161499b BA	249	V = V[-1]
4204e877 BA	250	pairs = c(pairs, lapply(V, function(v) c(i,v)))
4204e877 BA	251	}
a174b8ea	252
4204e877 BA	253	computeSaveCWT = function(index)
	254	{
	255	ts <- scale(ts(synchrones[index,]), center=TRUE, scale=scaled)
	256	totts.cwt = Rwave::cwt(ts, totnoct, nvoice, w0, plot=FALSE)
	257	ts.cwt = totts.cwt[,s0log:(s0log+noctave*nvoice)]
	258	#Normalization
	259	sqs <- sqrt(2^(0:(noctavenvoice)/nvoice)s0)
	260	sqres <- sweep(ts.cwt,2,sqs,'*')
	261	res <- sqres / max(Mod(sqres))
	262	#TODO: serializer les CWT, les récupérer via getDataInFile ;
	263	#--> OK, faut juste stocker comme séries simples de taille deltancol (5317519)
a174b8ea	264	binarize(c(as.double(Re(res)),as.double(Im(res))), cwt_file, ncol(res), ",", nbytes, endian)
4204e877 BA	265	}
	266
	267	if (parll)
	268	{
	269	cl = parallel::makeCluster(ncores_clust)
	270	synchrones_desc <- bigmemory::describe(synchrones)
	271	Xwer_dist_desc <- bigmemory::describe(Xwer_dist)
	272	parallel::clusterExport(cl, varlist=c("synchrones_desc","Xwer_dist_desc","totnoct",
	273	"nvoice","w0","s0log","noctave","s0","verbose","getCWT"), envir=environment())
	274	}
	275
	276	#precompute and serialize all CWT
	277	ignored <-
	278	if (parll)
	279	parallel::parLapply(cl, 1:n, computeSaveCWT)
	280	else
	281	lapply(1:n, computeSaveCWT)
	282
	283	getCWT = function(index)
	284	{
	285	#from cwt_file ...
a174b8ea	286	res <- getDataInFile(c(2index-1,2index), cwt_file, nbytes, endian)
eef6f6c9	287	###############TODO:
4204e877	288	}
e161499b	289
777c4b02	290	# Distance between rows i and j
e161499b BA	291	computeDistancesIJ = function(pair)
e161499b BA	292	{
2c14dbea	293	if (parll)
363ae134	294	{
2c14dbea BA	295	require("bigmemory", quietly=TRUE)
	296	require("epclust", quietly=TRUE)
	297	synchrones <- bigmemory::attach.big.matrix(synchrones_desc)
	298	Xwer_dist <- bigmemory::attach.big.matrix(Xwer_dist_desc)
	299	}
	300
e161499b BA	301	i = pair[1] ; j = pair[2]
	302	if (verbose && j==i+1)
	303	cat(paste(" Distances (",i,",",j,"), (",i,",",j+1,") ...\n", sep=""))
4204e877 BA	304	cwt_i <- getCWT(i)
4204e877 BA	305	cwt_j <- getCWT(j)
2c14dbea	306
363ae134 BA	307	num <- epclustFilter(Mod(cwt_i * Conj(cwt_j)))
363ae134 BA	308	WX <- epclustFilter(Mod(cwt_i * Conj(cwt_i)))
4204e877	309	WY <- epclustFilter(Mod(cwt_j * Conj(cwt_j)))
e161499b	310	wer2 <- sum(colSums(num)^2) / sum(colSums(WX) * colSums(WY))
2c14dbea	311	Xwer_dist[i,j] <- sqrt(delta * ncol(cwt_i) * max(1 - wer2, 0.)) #FIXME: wer2 should be < 1
e161499b BA	312	Xwer_dist[j,i] <- Xwer_dist[i,j]
	313	Xwer_dist[i,i] = 0.
	314	}
	315
e161499b	316	ignored <-
492cd9e7	317	if (parll)
e161499b	318	parallel::parLapply(cl, pairs, computeDistancesIJ)
492cd9e7	319	else
e161499b	320	lapply(pairs, computeDistancesIJ)
492cd9e7 BA	321
	322	if (parll)
	323	parallel::stopCluster(cl)
6ad3f3fd	324
492cd9e7	325	Xwer_dist[n,n] = 0.
777c4b02 BA	326	distances <- Xwer_dist[,]
	327	rm(Xwer_dist) ; gc()
	328	distances #~small matrix K1 x K1
492cd9e7 BA	329	}
	330
	331	# Helper function to divide indices into balanced sets
	332	.spreadIndices = function(indices, nb_per_chunk)
	333	{
	334	L = length(indices)
	335	nb_workers = floor( L / nb_per_chunk )
	336	if (nb_workers == 0)
	337	{
	338	# L < nb_series_per_chunk, simple case
	339	indices_workers = list(indices)
	340	}
	341	else
	342	{
	343	indices_workers = lapply( seq_len(nb_workers), function(i)
	344	indices[(nb_per_chunk(i-1)+1):(nb_per_chunki)] )
	345	# Spread the remaining load among the workers
	346	rem = L %% nb_per_chunk
	347	while (rem > 0)
	348	{
	349	index = rem%%nb_workers + 1
	350	indices_workers[[index]] = c(indices_workers[[index]], indices[L-rem+1])
	351	rem = rem - 1
d03c0621	352	}
1c6f223e	353	}
492cd9e7	354	indices_workers
1c6f223e	355	}