code/draft_R_pkg/R/main.R

   1 #' @include defaults.R
   2
   3 #' @title Cluster power curves with PAM in parallel
   4 #'
   5 #' @description Groups electricity power curves (or any series of similar nature) by applying PAM
   6 #' algorithm in parallel to chunks of size \code{nb_series_per_chunk}
   7 #'
   8 #' @param data Access to the data, which can be of one of the three following types:
   9 #' \itemize{
  10 #'   \item data.frame: each line contains its ID in the first cell, and all values after
  11 #'   \item connection: any R connection object (e.g. a file) providing lines as described above
  12 #'   \item function: a custom way to retrieve the curves; it has two arguments: the start index
  13 #'     (start) and number of curves (n); see example in package vignette.
  14 #' }
  15 #' @param K Number of clusters
  16 #' @param nb_series_per_chunk (Maximum) number of series in each group
  17 #' @param min_series_per_chunk Minimum number of series in each group
  18 #' @param writeTmp Function to write temporary wavelets coefficients (+ identifiers);
  19 #'   see defaults in defaults.R
  20 #' @param readTmp Function to read temporary wavelets coefficients (see defaults.R)
  21 #' @param WER "end" to apply stage 2 after stage 1 has iterated and finished, or "mix"
  22 #'   to apply it after every stage 1
  23 #' @param ncores number of parallel processes; if NULL, use parallel::detectCores()
  24 #'
  25 #' @return A data.frame of the final medoids curves (identifiers + values)
  26 epclust = function(data, K, nb_series_per_chunk, min_series_per_chunk=10*K,
  27     writeTmp=defaultWriteTmp, readTmp=defaultReadTmp, WER="end", ncores=NULL)
  28 {
  29     #TODO: setRefClass(...) to avoid copy data:
  30     #http://stackoverflow.com/questions/2603184/r-pass-by-reference
  31
  32     #0) check arguments
  33     if (!is.data.frame(data) && !is.function(data))
  34         tryCatch(
  35             {
  36                 if (is.character(data))
  37                 {
  38                     data_con = file(data, open="r")
  39                 } else if (!isOpen(data))
  40                 {
  41                     open(data)
  42                     data_con = data
  43                 }
  44             },
  45             error="data should be a data.frame, a function or a valid connection")
  46     if (!is.integer(K) || K < 2)
  47         stop("K should be an integer greater or equal to 2")
  48     if (!is.integer(nb_series_per_chunk) || nb_series_per_chunk < K)
  49         stop("nb_series_per_chunk should be an integer greater or equal to K")
  50     if (!is.function(writeTmp) || !is.function(readTmp))
  51         stop("read/writeTmp should be functional (see defaults.R)")
  52     if (WER!="end" && WER!="mix")
  53         stop("WER takes values in {'end','mix'}")
  54     #concerning ncores, any non-integer type will be treated as "use parallel:detectCores()"
  55
  56     #1) acquire data (process curves, get as coeffs)
  57     index = 1
  58     nb_curves = 0
  59     repeat
  60     {
  61         coeffs_chunk = NULL
  62         if (is.data.frame(data))
  63         {
  64             #full data matrix
  65             if (index < nrow(data))
  66             {
  67                 coeffs_chunk = curvesToCoeffs(
  68                     data[index:(min(index+nb_series_per_chunk-1,nrow(data))),])
  69             }
  70         } else if (is.function(data))
  71         {
  72             #custom user function to retrieve next n curves, probably to read from DB
  73             coeffs_chunk = curvesToCoeffs( data(index, nb_series_per_chunk) )
  74         } else
  75         {
  76             #incremental connection
  77             #TODO: find a better way to parse than using a temp file
  78             ascii_lines = readLines(data_con, nb_series_per_chunk)
  79             if (length(ascii_lines > 0))
  80             {
  81                 series_chunk_file = ".tmp/series_chunk"
  82                 writeLines(ascii_lines, series_chunk_file)
  83                 coeffs_chunk = curvesToCoeffs( read.csv(series_chunk_file) )
  84             }
  85         }
  86         if (is.null(coeffs_chunk))
  87             break
  88         writeTmp(coeffs_chunk)
  89         nb_curves = nb_curves + nrow(coeffs_chunk)
  90         index = index + nb_series_per_chunk
  91     }
  92     if (exists(data_con))
  93         close(data_con)
  94     if (nb_curves < min_series_per_chunk)
  95         stop("Not enough data: less rows than min_series_per_chunk!")
  96
  97     #2) process coeffs (by nb_series_per_chunk) and cluster them in parallel
  98     library(parallel)
  99     ncores = ifelse(is.integer(ncores), ncores, parallel::detectCores())
 100     cl = parallel::makeCluster(ncores)
 101     parallel::clusterExport(cl=cl, varlist=c("X", "Y", "K", "p"), envir=environment())
 102     library(cluster)
 103     #TODO: be careful of writing to a new temp file, then flush initial one, then re-use it...
 104     repeat
 105     {
 106         #while there is jobs to do (i.e. size of tmp "file" is greater than nb_series_per_chunk)
 107         nb_workers = nb_curves %/% nb_series_per_chunk
 108         indices = list()
 109         #incides[[i]] == (start_index,number_of_elements)
 110         for (i in 1:nb_workers)
 111             indices[[i]] = c(nb_series_per_chunk*(i-1)+1, nb_series_per_chunk)
 112         remainder = nb_curves %% nb_series_per_chunk
 113         if (remainder >= min_series_per_chunk)
 114         {
 115             nb_workers = nb_workers + 1
 116             indices[[nb_workers]] = c(nb_curves-remainder+1, nb_curves)
 117         } else if (remainder > 0)
 118         {
 119             #spread the load among other workers
 120
 121         }
 122         li = parallel::parLapply(cl, indices, processChunk, WER=="mix")
 123         #C) flush tmp file (current parallel processes will write in it)
 124     }
 125     parallel::stopCluster(cl)
 126
 127     #3) readTmp last results, apply PAM on it, and return medoids + identifiers
 128     final_coeffs = readTmp(1, nb_series_per_chunk)
 129     if (nrow(final_coeffs) == K)
 130     {
 131         return ( list( medoids=coeffsToCurves(final_coeffs[,2:ncol(final_coeffs)]),
 132             ids=final_coeffs[,1] ) )
 133     }
 134     pam_output = getClusters(as.matrix(final_coeffs[,2:ncol(final_coeffs)]), K)
 135     medoids = coeffsToCurves(pam_output$medoids)
 136     ids = final_coeffs[,1] [pam_output$ranks]
 137     return (list(medoids=medoids, ids=ids))
 138
 139     #4) apply stage 2 (in parallel ? inside task 2) ?)
 140     if (WER == "end")
 141     {
 142         #from center curves, apply stage 2...
 143     }
 144 }
 145
 146 processChunk = function(indice, WER)
 147 {
 148     #1) retrieve data
 149     #2) cluster
 150     #3) WER (optional)
 151 }