Newer
Older
array_piece[[i]] <- Subset(data_array_seperate[[which_chunk[i]]],
across_inner_dim,
ind_in_array_seperate[[which_chunk[i]]]:(ind_in_array_seperate[[which_chunk[i]]] + how_many_indices[i] - 1))
ind_in_array_seperate[[which_chunk[i]]] <- ind_in_array_seperate[[which_chunk[i]]] + how_many_indices[i]
}
# re-build the array: paste
data_array_tmp <- array_piece[[1]]
along_pos <- which(names(dim(data_array_tmp)) == across_inner_dim)
if (length(array_piece) > 1) {
for (i in 2:length(array_piece)) {
data_array_tmp <- abind::abind(data_array_tmp, array_piece[[i]],
along = along_pos)
}
}
}
}
4019
4020
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
4077
4078
4079
4080
4081
4082
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
4098
4099
4100
4101
4102
4103
4104
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
4163
4164
4165
4166
4167
4168
4169
4170
4171
4172
4173
4174
4175
4176
4177
4178
4179
4180
4181
4182
4183
4184
4185
4186
4187
4188
4189
4190
4191
4192
4193
4194
4195
4196
4197
4198
4199
4200
4201
4202
4203
4204
4205
4206
4207
4208
4209
4210
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
4245
4246
4247
4248
4249
4250
4251
4252
4253
4254
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4266
4267
4268
4269
4270
4271
4272
4273
4274
4275
4276
4277
4278
4279
4280
4281
4282
4283
4284
4285
4286
4287
4288
4289
4290
4291
4292
4293
4294
4295
4296
4297
4298
4299
4300
4301
4302
4303
4304
4305
4306
4307
4308
4309
4310
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4330
4331
4332
4333
4334
4335
4336
4337
4338
4339
data_array <- array(data_array_tmp, dim = final_dims_fake)
} else { # merge_across_dims_narm = F (old version)
data_array <- array(bigmemory::as.matrix(data_array), dim = final_dims_fake)
}
# NOTE: If split_multiselected_dims + merge_across_dims, the dimension order may change above.
# To get the user-required dim order, we need to reorder the array again.
if (split_multiselected_dims & merge_across_dims) {
if (inner_dim_pos_in_split_dims != 1) {
correct_order <- match(names(final_dims_fake_output), names(final_dims_fake))
data_array <- .aperm2(data_array, correct_order)
}
}
gc()
# Load metadata and remove the metadata folder
if (!is.null(metadata_dims)) {
loaded_metadata_files <- list.files(metadata_folder)
loaded_metadata <- lapply(paste0(metadata_folder, '/', loaded_metadata_files), readRDS)
unlink(metadata_folder, recursive = TRUE)
return_metadata <- vector('list', length = prod(dim(array_of_metadata_flags)[metadata_dims]))
return_metadata[as.numeric(loaded_metadata_files)] <- loaded_metadata
dim(return_metadata) <- dim(array_of_metadata_flags[metadata_dims])
attr(data_array, 'Variables') <- return_metadata
# TODO: Try to infer data type from loaded_metadata
# as.integer(data_array)
}
failed_pieces <- work_pieces[which(unlist(found_files))]
for (failed_piece in failed_pieces) {
array_of_not_found_files <- do.call('[<-',
c(list(array_of_not_found_files),
as.list(failed_piece[['file_indices_in_array_of_files']]),
list(value = TRUE)))
}
if (any(array_of_not_found_files)) {
for (i in 1:prod(dim(array_of_files_to_load))) {
if (is.na(array_of_not_found_files[i])) {
array_of_files_to_load[i] <- NA
} else {
if (array_of_not_found_files[i]) {
array_of_not_found_files[i] <- array_of_files_to_load[i]
array_of_files_to_load[i] <- NA
} else {
array_of_not_found_files[i] <- NA
}
}
}
} else {
array_of_not_found_files <- NULL
}
} # End if (retrieve)
# Change final_dims_fake back because retrieve = FALSE will use it for attributes later
if (exists("final_dims_fake_output")) {
final_dims_fake <- final_dims_fake_output
}
# Replace the vars and common vars by the transformed vars and common vars
for (i in 1:length(dat)) {
if (length(names(transformed_vars[[i]])) > 0) {
picked_vars[[i]][names(transformed_vars[[i]])] <- transformed_vars[[i]]
} else if (length(names(picked_vars_ordered[[i]])) > 0) {
picked_vars[[i]][names(picked_vars_ordered[[i]])] <- picked_vars_ordered[[i]]
}
}
if (length(names(transformed_common_vars)) > 0) {
picked_common_vars[names(transformed_common_vars)] <- transformed_common_vars
} else if (length(names(picked_common_vars_ordered)) > 0) {
picked_common_vars[names(picked_common_vars_ordered)] <- picked_common_vars_ordered
}
if (debug) {
print("-> THE TRANSFORMED VARS:")
print(str(transformed_vars))
print("-> THE PICKED VARS:")
print(str(picked_vars))
}
file_selectors <- NULL
for (i in 1:length(dat)) {
file_selectors[[dat[[i]][['name']]]] <- dat[[i]][['selectors']][which(names(dat[[i]][['selectors']]) %in% found_file_dims[[i]])]
}
if (retrieve) {
if (!silent) {
.message("Successfully retrieved data.")
}
var_backup <- attr(data_array, 'Variables')[[1]]
attr(data_array, 'Variables') <- NULL
attributes(data_array) <- c(attributes(data_array),
list(Variables = c(list(common = c(picked_common_vars, var_backup)),
picked_vars),
Files = array_of_files_to_load,
NotFoundFiles = array_of_not_found_files,
FileSelectors = file_selectors,
PatternDim = found_pattern_dim)
)
attr(data_array, 'class') <- c('startR_array', attr(data_array, 'class'))
data_array
} else {
if (!silent) {
.message("Successfully discovered data dimensions.")
}
start_call <- match.call()
for (i in 2:length(start_call)) {
if (class(start_call[[i]]) %in% c('name', 'call')) {
start_call[[i]] <- eval.parent(start_call[[i]])
}
}
start_call[['retrieve']] <- TRUE
attributes(start_call) <- c(attributes(start_call),
list(Dimensions = final_dims_fake,
Variables = c(list(common = picked_common_vars), picked_vars),
ExpectedFiles = array_of_files_to_load,
FileSelectors = file_selectors,
PatternDim = found_pattern_dim,
MergedDims = if (merge_across_dims) {
inner_dims_across_files
} else {
NULL
},
SplitDims = if (split_multiselected_dims) {
all_split_dims
} else {
NULL
})
)
attr(start_call, 'class') <- c('startR_cube', attr(start_call, 'class'))
start_call
}
}
# This function is the responsible for loading the data of each work
# piece.
.LoadDataFile <- function(work_piece, shared_matrix_pointer,
file_data_reader, synonims,
transform, transform_params,
silent = FALSE, debug = FALSE) {
# suppressPackageStartupMessages({library(bigmemory)})
### TODO: Specify dependencies as parameter
# suppressPackageStartupMessages({library(ncdf4)})
#print("1")
store_indices <- as.list(work_piece[['store_position']])
first_round_indices <- work_piece[['first_round_indices']]
second_round_indices <- work_piece[['second_round_indices']]
#print("2")
file_to_open <- work_piece[['file_path']]
sub_array <- file_data_reader(file_to_open, NULL,
work_piece[['file_selectors']],
first_round_indices, synonims)
if (debug) {
if (all(unlist(store_indices[1:6]) == 1)) {
print("-> LOADING A WORK PIECE")
print("-> STRUCTURE OF READ UNTRANSFORMED DATA:")
print(str(sub_array))
print("-> STRUCTURE OF VARIABLES TO TRANSFORM:")
print(str(work_piece[['vars_to_transform']]))
print("-> COMMON ARRAY DIMENSIONS:")
print(str(work_piece[['store_dims']]))
}
}
if (!is.null(sub_array)) {
# Apply data transformation once we have the data arrays.
if (!is.null(transform)) {
if (debug) {
if (all(unlist(store_indices[1:6]) == 1)) {
print("-> PROCEEDING TO TRANSFORM ARRAY")
print("-> DIMENSIONS OF ARRAY RIGHT BEFORE TRANSFORMING:")
print(dim(sub_array))
}
}
sub_array <- do.call(transform, c(list(data_array = sub_array,
variables = work_piece[['vars_to_transform']],
file_selectors = work_piece[['file_selectors']]),
transform_params))
if (debug) {
if (all(unlist(store_indices[1:6]) == 1)) {
print("-> STRUCTURE OF ARRAY AND VARIABLES RIGHT AFTER TRANSFORMING:")
print(str(sub_array))
print("-> DIMENSIONS OF ARRAY RIGHT AFTER TRANSFORMING:")
print(dim(sub_array$data_array))
}
}
sub_array <- sub_array$data_array
# Subset with second round of indices
dims_to_crop <- which(!sapply(second_round_indices, is.null))
if (length(dims_to_crop) > 0) {
dimnames_to_crop <- names(second_round_indices)[dims_to_crop]
sub_array <- Subset(sub_array, dimnames_to_crop,
second_round_indices[dimnames_to_crop])
}
if (debug) {
if (all(unlist(store_indices[1:6]) == 1)) {
print("-> STRUCTURE OF ARRAY AND VARIABLES RIGHT AFTER SUBSETTING WITH 2nd ROUND INDICES:")
print(str(sub_array))
}
}
}
metadata <- attr(sub_array, 'variables')
names_bk <- names(store_indices)
store_indices <- lapply(names(store_indices),
function (x) {
if (!(x %in% names(first_round_indices))) {
store_indices[[x]]
} else if (is.null(second_round_indices[[x]])) {
1:dim(sub_array)[x]
} else {
if (is.numeric(second_round_indices[[x]])) {
## TODO: Review carefully this line. Inner indices are all
## aligned to the left-most positions. If dataset A has longitudes
## 1, 2, 3, 4 but dataset B has only longitudes 3 and 4, then
## they will be stored as follows:
## 1, 2, 3, 4
## 3, 4, NA, NA
##x - min(x) + 1
1:length(second_round_indices[[x]])
} else {
1:length(second_round_indices[[x]])
}
}
})
names(store_indices) <- names_bk
if (debug) {
if (all(unlist(store_indices) == 1)) {
print("-> STRUCTURE OF FIRST ROUND INDICES FOR THIS WORK PIECE:")
print(str(first_round_indices))
print("-> STRUCTURE OF SECOND ROUND INDICES FOR THIS WORK PIECE:")
print(str(second_round_indices))
print("-> STRUCTURE OF STORE INDICES FOR THIS WORK PIECE:")
print(str(store_indices))
}
}
store_indices <- lapply(store_indices, as.integer)
store_dims <- work_piece[['store_dims']]
# split the storage work of the loaded subset in parts
largest_dim_name <- names(dim(sub_array))[which.max(dim(sub_array))]
max_parts <- length(store_indices[[largest_dim_name]])
# Indexing a data file of N MB with expand.grid takes 30*N MB
# The peak ram of Start is, minimum, 2 * total data to load from all files
# due to inefficiencies in other regions of the code
# The more parts we split the indexing done below in, the lower
# the memory footprint of the indexing and the fast.
# But more than 10 indexing iterations (parts) for each MB processed
# makes the iteration slower (tested empirically on BSC workstations).
subset_size_in_mb <- prod(dim(sub_array)) * 8 / 1024 / 1024
best_n_parts <- ceiling(subset_size_in_mb * 10)
# We want to set n_parts to a greater value than the one that would
# result in a memory footprint (of the subset indexing code below) equal
# to 2 * total data to load from all files.
# s = subset size in MB
# p = number of parts to break it in
# T = total size of data to load
# then, s / p * 30 = 2 * T
# then, p = s * 15 / T
min_n_parts <- ceiling(prod(dim(sub_array)) * 15 / prod(store_dims))
# Make sure we pick n_parts much greater than the minimum calculated
n_parts <- min_n_parts * 10
if (n_parts > best_n_parts) {
n_parts <- best_n_parts
}
# Boundary checks
if (n_parts < 1) {
n_parts <- 1
}
if (n_parts > max_parts) {
n_parts <- max_parts
}
if (n_parts > 1) {
make_parts <- function(length, n) {
clusters <- cut(1:length, n, labels = FALSE)
lapply(1:n, function(y) which(clusters == y))
}
part_indices <- make_parts(max_parts, n_parts)
parts <- lapply(part_indices,
function(x) {
store_indices[[largest_dim_name]][x]
})
} else {
part_indices <- list(1:max_parts)
parts <- store_indices[largest_dim_name]
}
# do the storage work
weights <- sapply(1:length(store_dims),
function(i) prod(c(1, store_dims)[1:i]))
part_indices_in_sub_array <- as.list(rep(TRUE, length(dim(sub_array))))
names(part_indices_in_sub_array) <- names(dim(sub_array))
data_array <- bigmemory::attach.big.matrix(shared_matrix_pointer)
for (i in 1:n_parts) {
store_indices[[largest_dim_name]] <- parts[[i]]
# Converting array indices to vector indices
matrix_indices <- do.call("expand.grid", store_indices)
# Given a matrix where each row is a set of array indices of an element
# the vector indices are computed
matrix_indices <- 1 + colSums(t(matrix_indices - 1) * weights)
part_indices_in_sub_array[[largest_dim_name]] <- part_indices[[i]]
data_array[matrix_indices] <- as.vector(do.call('[',
c(list(x = sub_array),
part_indices_in_sub_array)))
}
rm(data_array)
gc()
if (!is.null(work_piece[['save_metadata_in']])) {
saveRDS(metadata, file = work_piece[['save_metadata_in']])
}
}
if (!is.null(work_piece[['progress_amount']]) && !silent) {
message(work_piece[['progress_amount']], appendLF = FALSE)
}
is.null(sub_array)
}