Newer
Older
Start <- function(..., # dim = indices/selectors,
# dim_var = 'var',
# dim_reorder = Sort/CircularSort,
# dim_tolerance = number,
# dim_depends = 'file_dim',
# dim_across = 'file_dim',
return_vars = NULL,
file_opener = NcOpener,
file_var_reader = NcVarReader,
file_dim_reader = NcDimReader,
file_data_reader = NcDataReader,
file_closer = NcCloser,
transform = NULL,
transform_params = NULL,
transform_vars = NULL,
transform_extra_cells = 0,
apply_indices_after_transform = FALSE,
pattern_dims = NULL,
metadata_dims = NULL,
#, config_file = NULL
#dictionary_dim_names = ,
#dictionary_var_names =
dim_params <- list(...)
# Take *_var parameters apart
var_params_ind <- grep('_var$', names(dim_params))
var_params <- dim_params[var_params_ind]
# Check all *_var are NULL or vectors of character strings, and
# that they all have a matching dimension param.
i <- 1
for (var_param in var_params) {
if (!is.character(var_param)) {
stop("All '*_var' parameters must be character strings.")
} else if (!any(grepl(paste0('^', strsplit(names(var_params)[i],
'_var$')[[1]][1], '$'),
names(dim_params)))) {
stop(paste0("All '*_var' parameters must be associated to a dimension parameter. Found parameter '",
names(var_params)[i], "' but no parameter '",
strsplit(names(var_params)[i], '_var$')[[1]][1], "'."))
}
i <- i + 1
}
# Make the keys of 'var_params' to be the name of
# the corresponding dimension.
if (length(var_params) < 1) {
var_params <- NULL
} else {
names(var_params) <- gsub('_var$', '', names(var_params))
}
# Take *_reorder parameters apart
dim_reorder_params_ind <- grep('_reorder$', names(dim_params))
dim_reorder_params <- dim_params[dim_reorder_params_ind]
# Check all *_reorder are NULL or functions, and that they all have
# a matching dimension param.
i <- 1
for (dim_reorder_param in dim_reorder_params) {
if (!is.function(dim_reorder_param)) {
stop("All '*_reorder' parameters must be functions.")
} else if (!any(grepl(paste0('^', strsplit(names(dim_reorder_params)[i],
'_reorder$')[[1]][1], '$'),
names(dim_params)))) {
stop(paste0("All '*_reorder' parameters must be associated to a dimension parameter. Found parameter '",
names(dim_reorder_params)[i], "' but no parameter '",
strsplit(names(dim_reorder_params)[i], '_reorder$')[[1]][1], "'."))
} else if (!any(grepl(paste0('^', strsplit(names(dim_reorder_params)[i],
'_reorder$')[[1]][1], '$'),
names(var_params)))) {
stop(paste0("All '*_reorder' parameters must be associated to a dimension parameter associated to a ",
"variable. Found parameter '", names(dim_reorder_params)[i], "' and dimension parameter '",
strsplit(names(dim_reorder_params)[i], '_reorder$')[[1]][1], "' but did not find variable ",
"parameter '", strsplit(names(dim_reorder_params)[i], '_reorder$')[[1]][1], "_var'."))
}
i <- i + 1
}
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
# Make the keys of 'dim_reorder_params' to be the name of
# the corresponding dimension.
if (length(dim_reorder_params) < 1) {
dim_reorder_params <- NULL
} else {
names(dim_reorder_params) <- gsub('_reorder$', '', names(dim_reorder_params))
}
# Take *_tolerance parameters apart
tolerance_params_ind <- grep('_tolerance$', names(dim_params))
tolerance_params <- dim_params[tolerance_params_ind]
# Check all *_tolerance are NULL or vectors of character strings, and
# that they all have a matching dimension param.
i <- 1
for (tolerance_param in tolerance_params) {
if (!any(grepl(paste0('^', strsplit(names(tolerance_params)[i],
'_tolerance$')[[1]][1], '$'),
names(dim_params)))) {
stop(paste0("All '*_tolerance' parameters must be associated to a dimension parameter. Found parameter '",
names(tolerance_params)[i], "' but no parameter '",
strsplit(names(tolerance_params)[i], '_tolerance$')[[1]][1], "'."))
} else if (!any(grepl(paste0('^', strsplit(names(tolerance_params)[i],
'_tolerance$')[[1]][1], '$'),
names(var_params)))) {
stop(paste0("All '*_tolerance' parameters must be associated to a dimension parameter associated to a ",
"variable. Found parameter '", names(tolerance_params)[i], "' and dimension parameter '",
strsplit(names(tolerance_params)[i], '_tolerance$')[[1]][1], "' but did not find variable ",
"parameter '", strsplit(names(tolerance_params)[i], '_tolerance$')[[1]][1], "_var'."))
}
i <- i + 1
}
# Make the keys of 'tolerance_params' to be the name of
# the corresponding dimension.
if (length(tolerance_params) < 1) {
tolerance_params <- NULL
} else {
names(tolerance_params) <- gsub('_tolerance$', '', names(tolerance_params))
}
# Take *_depends parameters apart
depends_params_ind <- grep('_depends$', names(dim_params))
depends_params <- dim_params[depends_params_ind]
# Check all *_depends are NULL or vectors of character strings, and
# that they all have a matching dimension param.
i <- 1
for (depends_param in depends_params) {
if (!is.character(depends_param) || (length(depends_param) > 1)) {
stop("All '*_depends' parameters must be single character strings.")
} else if (!any(grepl(paste0('^', strsplit(names(depends_params)[i],
'_depends$')[[1]][1], '$'),
names(dim_params)))) {
stop(paste0("All '*_depends' parameters must be associated to a dimension parameter. Found parameter '",
names(depends_params)[i], "' but no parameter '",
strsplit(names(depends_params)[i], '_depends$')[[1]][1], "'."))
}
i <- i + 1
}
# Make the keys of 'depends_params' to be the name of
# the corresponding dimension.
if (length(depends_params) < 1) {
depends_params <- NULL
} else {
names(depends_params) <- gsub('_depends$', '', names(depends_params))
}
# Change name to depending_file_dims
depending_file_dims <- depends_params
# Take *_across parameters apart
across_params_ind <- grep('_across$', names(dim_params))
across_params <- dim_params[across_params_ind]
# Check all *_across are NULL or vectors of character strings, and
# that they all have a matching dimension param.
i <- 1
for (across_param in across_params) {
if (!is.character(across_param) || (length(across_param) > 1)) {
stop("All '*_across' parameters must be single character strings.")
} else if (!any(grepl(paste0('^', strsplit(names(across_params)[i],
'_across$')[[1]][1], '$'),
names(dim_params)))) {
stop(paste0("All '*_across' parameters must be associated to a dimension parameter. Found parameter '",
names(across_params)[i], "' but no parameter '",
strsplit(names(across_params)[i], '_across$')[[1]][1], "'."))
}
i <- i + 1
}
# Make the keys of 'across_params' to be the name of
# the corresponding dimension.
if (length(across_params) < 1) {
across_params <- NULL
} else {
names(across_params) <- gsub('_across$', '', names(across_params))
}
# Change name to inner_dims_across_files
inner_dims_across_files <- across_params
# Leave alone the dimension parameters in the variable dim_params
if (length(c(var_params_ind, dim_reorder_params_ind, tolerance_params_ind,
depends_params_ind, across_params_ind)) > 0) {
dim_params <- dim_params[-c(var_params_ind, dim_reorder_params_ind,
tolerance_params_ind, depends_params_ind,
across_params_ind)]
}
dim_names <- names(dim_params)
if (is.null(dim_names)) {
stop("At least one pattern dim must be specified.")
}
# Look for chunked dims
chunks <- vector('list', length(dim_names))
names(chunks) <- dim_names
for (dim_name in dim_names) {
if (!is.null(attr(dim_params[[dim_name]], 'chunk'))) {
chunks[[dim_name]] <- attr(dim_params[[dim_name]], 'chunk')
attributes(dim_params[[dim_name]]) <- attributes(dim_params[[dim_name]])[-which(names(attributes(dim_params[[dim_name]])) == 'chunk')]
} else {
chunks[[dim_name]] <- c(chunk = 1, n_chunks = 1)
}
}
# This is a helper function to compute the chunk indices to take once the total
# number of indices for a dimension has been discovered.
chunk_indices <- function(n_indices, chunk, n_chunks, dim_name) {
if (n_chunks > n_indices) {
stop("Requested to divide dimension '", dim_name, "' of length ",
n_indices, " in ", n_chunks, " chunks, which is not possible.")
}
chunks_to_extend <- n_indices - chunk_sizes[1] * n_chunks
if (chunks_to_extend > 0) {
chunk_sizes[1:chunks_to_extend] <- chunk_sizes[1:chunks_to_extend] + 1
}
chunk_size <- chunk_sizes[chunk]
offset <- 0
if (chunk > 1) {
offset <- sum(chunk_sizes[1:(chunk - 1)])
}
indices <- 1:chunk_sizes[chunk] + offset
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
# Check pattern_dims
if (is.null(pattern_dims)) {
.warning(paste0("Parameter 'pattern_dims' not specified. Taking the first dimension, '",
dim_names[1], "' as 'pattern_dims'."))
pattern_dims <- dim_names[1]
} else if (is.character(pattern_dims) && (length(pattern_dims) > 0)) {
pattern_dims <- unique(pattern_dims)
} else {
stop("Parameter 'pattern_dims' must be a vector of character strings.")
}
if (any(names(var_params) %in% pattern_dims)) {
stop("'*_var' parameters specified for pattern dimensions. Remove or fix them.")
}
# Find the pattern dimension with the pattern specifications
found_pattern_dim <- NULL
for (pattern_dim in pattern_dims) {
# Check all specifications in pattern_dim are valid
dat <- datasets <- dim_params[[pattern_dim]]
if (is.null(dat) || !(is.character(dat) && all(nchar(dat) > 0)) && !is.list(dat)) {
stop(paste0("Parameter '", pattern_dim,
"' must be a list of lists with pattern specifications or a vector of character strings."))
}
if (!is.null(dim_reorder_params[[pattern_dim]])) {
.warning(paste0("A reorder for the selectors of '", pattern_dim,
"' has been specified, but it is a pattern dimension and the reorder will be ignored."))
}
if (is.null(found_pattern_dim)) {
found_pattern_dim <- pattern_dim
} else {
stop("Found more than one pattern dim with pattern specifications (list of lists). One and only one pattern dim must contain pattern specifications.")
}
}
}
if (is.null(found_pattern_dim)) {
.warning(paste0("Could not find any pattern dim with explicit data set descriptions (in the form of list of lists). Taking the first pattern dim, '", pattern_dims[1], "', as dimension with pattern specifications."))
found_pattern_dim <- pattern_dims[1]
}
# Check metadata_dims
if (!is.null(metadata_dims)) {
if (is.na(metadata_dims)) {
metadata_dims <- NULL
} else if (!is.character(metadata_dims) || (length(metadata_dims) < 1)) {
stop("Parameter 'metadata' dims must be a vector of at least one character string.")
}
} else {
metadata_dims <- pattern_dims
}
# Once the pattern dimension with dataset specifications is found,
# the variable 'dat' is mounted with the information of each
# dataset.
# Take only the datasets for the requested chunk
dats_to_take <- chunk_indices(length(dim_params[[found_pattern_dim]]),
chunks[[found_pattern_dim]]['chunk'],
chunks[[found_pattern_dim]]['n_chunks'],
found_pattern_dim)
dim_params[[found_pattern_dim]] <- dim_params[[found_pattern_dim]][dats_to_take]
dat <- datasets <- dim_params[[found_pattern_dim]]
dat_info_names <- c('name', 'path')#, 'nc_var_name', 'suffix', 'var_min', 'var_max', 'dimnames')
dat_to_fetch <- c()
dat_names <- c()
if (!is.list(dat)) {
dat <- as.list(dat)
} else {
if (!any(sapply(dat, is.list))) {
dat <- list(dat)
}
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
}
for (i in 1:length(dat)) {
if (is.character(dat[[i]]) && length(dat[[i]]) == 1 && nchar(dat[[i]]) > 0) {
if (grepl('^(\\./|\\.\\./|/.*/|~/)', dat[[i]])) {
dat[[i]] <- list(path = dat[[i]])
} else {
dat[[i]] <- list(name = dat[[i]])
}
} else if (!is.list(dat[[i]])) {
stop(paste0("Parameter '", pattern_dim,
"' is incorrect. It must be a list of lists or character strings."))
}
#if (!(all(names(dat[[i]]) %in% dat_info_names))) {
# stop("Error: parameter 'dat' is incorrect. There are unrecognized components in the information of some of the datasets. Check 'dat' in ?Load for details.")
#}
if (!('name' %in% names(dat[[i]]))) {
dat[[i]][['name']] <- paste0('dat', i)
if (!('path' %in% names(dat[[i]]))) {
stop(paste0("Parameter '", found_pattern_dim,
"' is incorrect. A 'path' should be provided for each dataset if no 'name' is provided."))
}
} else if (!('path' %in% names(dat[[i]]))) {
dat_to_fetch <- c(dat_to_fetch, i)
}
#if ('path' %in% names(dat[[i]])) {
# if (!('nc_var_name' %in% names(dat[[i]]))) {
# dat[[i]][['nc_var_name']] <- '$var_name$'
# }
# if (!('suffix' %in% names(dat[[i]]))) {
# dat[[i]][['suffix']] <- ''
# }
# if (!('var_min' %in% names(dat[[i]]))) {
# dat[[i]][['var_min']] <- ''
# }
# if (!('var_max' %in% names(dat[[i]]))) {
# dat[[i]][['var_max']] <- ''
# }
#}
dat_names <- c(dat_names, dat[[i]][['name']])
}
if ((length(dat_to_fetch) > 0) && (length(dat_to_fetch) < length(dat))) {
.warning("'path' has been provided for some datasets. Any information in the configuration file related to these will be ignored.")
}
if (length(dat_to_fetch) > 0) {
stop("Specified only the name for some data sets, but not the path ",
"pattern. This option has not been yet implemented.")
}
# Reorder inner_dims_across_files (to make the keys be the file dimensions,
# and the values to be the inner dimensions that go across it).
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
if (!is.null(inner_dims_across_files)) {
# Reorder: example, convert list(ftime = 'chunk', ensemble = 'member', xx = 'chunk')
# to list(chunk = c('ftime', 'xx'), member = 'ensemble')
new_idaf <- list()
for (i in names(inner_dims_across_files)) {
if (!(inner_dims_across_files[[i]] %in% names(new_idaf))) {
new_idaf[[inner_dims_across_files[[i]]]] <- i
} else {
new_idaf[[inner_dims_across_files[[i]]]] <- c(new_idaf[[inner_dims_across_files[[i]]]], i)
}
}
inner_dims_across_files <- new_idaf
}
# Check return_vars
if (is.null(return_vars)) {
return_vars <- list()
# if (length(var_params) > 0) {
# return_vars <- as.list(paste0(names(var_params), '_var'))
# } else {
# return_vars <- list()
# }
}
if (!is.list(return_vars)) {
stop("Parameter 'return_vars' must be a list or NULL.")
}
if (length(return_vars) > 0 && is.null(names(return_vars))) {
# names(return_vars) <- rep('', length(return_vars))
stop("Parameter 'return_vars' must be a named list.")
}
i <- 1
while (i <= length(return_vars)) {
# if (names(return_vars)[i] == '') {
# if (!(is.character(return_vars[[i]]) && (length(return_vars[[i]]) == 1))) {
# stop("The ", i, "th specification in 'return_vars' is malformed.")
# }
# if (!grepl('_var$', return_vars[[i]])) {
# stop("The ", i, "th specification in 'return_vars' is malformed.")
# }
# dim_name <- strsplit(return_vars[[i]], '_var$')[[1]][1]
# if (!(dim_name %in% names(var_params))) {
# stop("'", dim_name, "_var' requested in 'return_vars' but ",
# "no '", dim_name, "_var' specified in the .Load call.")
# }
# names(return_vars)[i] <- var_params[[dim_name]]
# return_vars[[i]] <- found_pattern_dim
# } else
if (length(return_vars[[i]]) > 0) {
if (!is.character(return_vars[[i]])) {
stop("The ", i, "th specification in 'return_vars' is malformed. It ",
"must be a vector of character strings of valid file dimension ",
"names.")
}
}
i <- i + 1
}
# Check synonims
if (!is.null(synonims)) {
error <- FALSE
if (!is.list(synonims)) {
error <- TRUE
}
for (synonim_entry in names(synonims)) {
if (!(synonim_entry %in% names(dim_params)) &&
!(synonim_entry %in% names(return_vars))) {
error <- TRUE
}
if (!is.character(synonims[[synonim_entry]]) ||
length(synonims[[synonim_entry]]) < 1) {
error <- TRUE
}
}
if (error) {
stop("Parameter 'synonims' must be a named list, where the names are ",
"a name of a requested dimension or variable and the values are ",
"vectors of character strings with at least one alternative name ",
" for each dimension or variable in 'synonims'.")
}
}
if (length(unique(names(synonims))) < length(names(synonims))) {
stop("There must not be repeated entries in 'synonims'.")
}
if (length(unique(unlist(synonims))) < length(unlist(synonims))) {
stop("There must not be repeated values in 'synonims'.")
}
# Make that all dims and vars have an entry in synonims, even if only dim_name = dim_name
dim_entries_to_add <- which(!(names(dim_params) %in% names(synonims)))
if (length(dim_entries_to_add) > 0) {
synonims[names(dim_params)[dim_entries_to_add]] <- as.list(names(dim_params)[dim_entries_to_add])
}
var_entries_to_add <- which(!(names(var_params) %in% names(synonims)))
if (length(var_entries_to_add) > 0) {
synonims[names(var_params)[var_entries_to_add]] <- as.list(names(var_params)[var_entries_to_add])
}
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
# Check selector_checker
if (is.null(selector_checker) || !is.function(selector_checker)) {
stop("Parameter 'selector_checker' must be a function.")
}
# Check file_opener
if (is.null(file_opener) || !is.function(file_opener)) {
stop("Parameter 'file_opener' must be a function.")
}
# Check file_var_reader
if (!is.null(file_var_reader) && !is.function(file_var_reader)) {
stop("Parameter 'file_var_reader' must be a function.")
}
# Check file_dim_reader
if (!is.null(file_dim_reader) && !is.function(file_dim_reader)) {
stop("Parameter 'file_dim_reader' must be a function.")
}
# Check file_data_reader
if (is.null(file_data_reader) || !is.function(file_data_reader)) {
stop("Parameter 'file_data_reader' must be a function.")
}
# Check file_closer
if (is.null(file_closer) || !is.function(file_closer)) {
stop("Parameter 'file_closer' must be a function.")
}
# Check transform
if (!is.null(transform)) {
if (!is.function(transform)) {
stop("Parameter 'transform' must be a function.")
}
}
# Check transform_params
if (!is.null(transform_params)) {
if (!is.list(transform_params)) {
stop("Parameter 'transform_params' must be a list.")
}
if (is.null(names(transform_params))) {
stop("Parameter 'transform_params' must be a named list.")
}
}
# Check transform_vars
if (!is.null(transform_vars)) {
if (!is.character(transform_vars)) {
stop("Parameter 'transform_vars' must be a vector of character strings.")
}
}
if (any(!(transform_vars %in% names(return_vars)))) {
stop("All the variables specified in 'transform_vars' must also be specified in 'return_vars'.")
}
# Check apply_indices_after_transform
if (!is.logical(apply_indices_after_transform)) {
stop("Parameter 'apply_indices_after_transform' must be either TRUE or FALSE.")
# Check transform_extra_cells
if (!is.numeric(transform_extra_cells)) {
stop("Parameter 'transform_extra_cells' must be numeric.")
}
transform_extra_cells <- round(transform_extra_cells)
# Check retrieve
if (!is.logical(retrieve)) {
stop("Parameter 'retrieve' must be TRUE or FALSE.")
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
# Check num_procs
if (!is.null(num_procs)) {
if (!is.numeric(num_procs)) {
stop("Parameter 'num_procs' must be numeric.")
} else {
num_procs <- round(num_procs)
}
}
# Check silent
if (!is.logical(silent)) {
stop("Parameter 'silent' must be logical.")
}
dim_params[[found_pattern_dim]] <- dat_names
if (!silent) {
.message(paste0("Exploring files... This will take a variable amount ",
"of time depending on the issued request and the ",
"performance of the file server..."))
}
if (!is.character(debug)) {
dims_to_check <- c('time')
} else {
dims_to_check <- debug
debug <- TRUE
}
############################## READING FILE DIMS ############################
# Check that no unrecognized variables are present in the path patterns
# and also that no file dimensions are requested to THREDDs catalogs.
# And in the mean time, build all the work pieces and look for the
# first available file of each dataset.
array_of_files_to_load <- NULL
array_of_not_found_files <- NULL
indices_of_first_files_with_data <- vector('list', length(dat))
selectors_of_first_files_with_data <- vector('list', length(dat))
found_file_dims <- vector('list', length(dat))
expected_inner_dims <- vector('list', length(dat))
dataset_has_files <- rep(FALSE, length(dat))
#print("A")
for (i in 1:length(dat)) {
#print("B")
dat_selectors <- dim_params
dat_selectors[[found_pattern_dim]] <- dat_selectors[[found_pattern_dim]][i]
dim_vars <- paste0('$', dim_names, '$')
file_dims <- which(sapply(dim_vars, grepl, dat[[i]][['path']], fixed = TRUE))
if (length(file_dims) > 0) {
file_dims <- dim_names[file_dims]
}
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
file_dims <- unique(c(pattern_dims, file_dims))
found_file_dims[[i]] <- file_dims
expected_inner_dims[[i]] <- dim_names[which(!(dim_names %in% file_dims))]
# (Check the depending_file_dims).
if (any(c(names(depending_file_dims), unlist(depending_file_dims)) %in%
expected_inner_dims[[i]])) {
stop(paste0("The dimension dependancies specified in ",
"'depending_file_dims' can only be between file ",
"dimensions, but some inner dimensions found in ",
"dependancies for '", dat[[i]][['name']], "', which ",
"has the following file dimensions: ",
paste(paste0("'", file_dims, "'"), collapse = ', '), "."))
} else {
a <- names(depending_file_dims) %in% file_dims
b <- unlist(depending_file_dims) %in% file_dims
ab <- a & b
if (any(!ab)) {
.warning("Detected some dependancies in 'depending_file_dims' with ",
"non-existing dimension names. These will be disregarded.")
depending_file_dims <- depending_file_dims[-which(!ab)]
}
if (any(names(depending_file_dims) == unlist(depending_file_dims))) {
depending_file_dims <- depending_file_dims[-which(names(depending_file_dims) == unlist(depending_file_dims))]
}
}
# (Check the inner_dims_across_files).
if (any(!(names(inner_dims_across_files) %in% file_dims)) ||
any(!(unlist(inner_dims_across_files) %in% expected_inner_dims[[i]]))) {
stop(paste0("All relationships specified in ",
"'inner_dims_across_files' must be between a inner ",
"dimension and a file dimension. Found wrong ",
"specification for '", dat[[i]][['name']], "', which ",
"has the following file dimensions: ",
paste(paste0("'", file_dims, "'"), collapse = ', '),
", and the following inner dimensions: ",
paste(paste0("'", expected_inner_dims[[i]], "'"),
collapse = ', '), "."))
}
# (Check the return_vars).
j <- 1
while (j <= length(return_vars)) {
if (any(!(return_vars[[j]] %in% file_dims))) {
if (any(return_vars[[j]] %in% expected_inner_dims[[i]])) {
stop("Found variables in 'return_vars' requested ",
"for some inner dimensions (for dataset '",
dat[[i]][['name']], "'), but variables can only be ",
"requested for file dimensions.")
} else {
stop("Found variables in 'return_vars' requested ",
"for non-existing dimensions.")
}
}
j <- j + 1
}
if (!is.null(metadata_dims)) {
if (any(!(metadata_dims %in% file_dims))) {
stop("All dimensions in 'metadata_dims' must be file dimensions.")
}
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
# (Check the *_var parameters).
if (any(!(unlist(var_params) %in% names(return_vars)))) {
stop("All '*_var' params must associate a dimension to one of the ",
"requested variables in 'return_vars'.")
}
# # (Check the circular_dims parameters).
# if (!all(names(circular_dims) %in% expected_inner_dims[[i]])) {
# stop("Only inner dimensions can be requested as 'circular_dims'.")
# }
replace_values <- vector('list', length = length(file_dims))
names(replace_values) <- file_dims
# Take the first selector for all possible file dimensions
for (file_dim in file_dims) {
if (file_dim %in% names(var_params)) {
.warning(paste0("The '", file_dim, "_var' param will be ignored since '",
file_dim, "' is a file dimension (for the dataset with pattern ",
dat[[i]][['path']], ")."))
}
if (!is.list(dat_selectors[[file_dim]]) ||
(is.list(dat_selectors[[file_dim]]) &&
length(dat_selectors[[file_dim]]) == 2 &&
is.null(names(dat_selectors[[file_dim]])))) {
dat_selectors[[file_dim]] <- list(dat_selectors[[file_dim]])
}
first_class <- class(dat_selectors[[file_dim]][[1]])
first_length <- length(dat_selectors[[file_dim]][[1]])
for (j in 1:length(dat_selectors[[file_dim]])) {
sv <- selector_vector <- dat_selectors[[file_dim]][[j]]
if (!identical(first_class, class(sv)) ||
!identical(first_length, length(sv))) {
stop("All provided selectors for depending dimensions must ",
"be vectors of the same length and of the same class.")
}
if (is.character(sv) && !((length(sv) == 1) && (sv[1] %in% c('all', 'first', 'last')))) {
dat_selectors[[file_dim]][[j]] <- selector_checker(selectors = sv,
return_indices = FALSE)
# Take chunk if needed
dat_selectors[[file_dim]][[j]] <- dat_selectors[[file_dim]][[j]][chunk_indices(length(dat_selectors[[file_dim]][[j]]),
chunks[[file_dim]]['chunk'],
chunks[[file_dim]]['n_chunks'],
file_dim)]
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
} else if (!(is.numeric(sv) ||
(is.character(sv) && (length(sv) == 1) && (sv %in% c('all', 'first', 'last'))) ||
(is.list(sv) && (length(sv) == 2) && (all(sapply(sv, is.character)) ||
all(sapply(sv, is.numeric)))))) {
stop("All explicitly provided selectors for file dimensions must be character strings.")
}
}
sv <- dat_selectors[[file_dim]][[1]]
if (is.character(sv) && !((length(sv) == 1) && (sv[1] %in% c('all', 'first', 'last')))) {
replace_values[[file_dim]] <- dat_selectors[[file_dim]][[1]][1]
}
}
#print("C")
# Now we know which dimensions whose selectors are provided non-explicitly.
undefined_file_dims <- file_dims[which(sapply(replace_values, is.null))]
defined_file_dims <- file_dims[which(!(file_dims %in% undefined_file_dims))]
# Quickly check if the depending dimensions are provided properly.
for (file_dim in file_dims) {
if (file_dim %in% names(depending_file_dims)) {
## TODO: Detect multi-dependancies and forbid.
if (all(c(file_dim, depending_file_dims[[file_dim]]) %in% defined_file_dims)) {
if (length(dat_selectors[[file_dim]]) != length(dat_selectors[[depending_file_dims[[file_dim]]]][[1]])) {
stop(paste0("If providing selectors for the depending ",
"dimension '", file_dim, "', a ",
"vector of selectors must be provided for ",
"each selector of the dimension it depends on, '",
depending_file_dims[[file_dim]], "'."))
} else if (!all(names(dat_selectors[[file_dim]]) == dat_selectors[[depending_file_dims[[file_dim]]]][[1]])) {
stop(paste0("If providing selectors for the depending ",
"dimension '", file_dim, "', the name of the ",
"provided vectors of selectors must match ",
"exactly the selectors of the dimension it ",
"depends on, '", depending_file_dims[[file_dim]], "'."))
}
}
}
}
# Find the possible values for the selectors that are provided as
# indices. If the requested file is on server, impossible operation.
if (length(grep("^http", dat[[i]][['path']])) > 0) {
if (length(undefined_file_dims) > 0) {
stop(paste0("All selectors for the file dimensions must be ",
"character strings if requesting data to a remote ",
"server. Found invalid selectors for the file dimensions ",
paste(paste0("'", undefined_file_dims, "'"), collapse = ', '), "."))
}
dataset_has_files[i] <- TRUE
} else {
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
# Iterate over the known dimensions to find the first existing file.
# The path to the first existing file will be used to find the
# values for the non explicitly defined selectors.
first_file <- NULL
first_file_selectors <- NULL
if (length(undefined_file_dims) > 0) {
replace_values[undefined_file_dims] <- '*'
}
## TODO: What if length of defined_file_dims is 0? code might crash (in practice it worked for an example case)
files_to_check <- sapply(dat_selectors[defined_file_dims], function(x) length(x[[1]]))
sub_array_of_files_to_check <- array(1:prod(files_to_check), dim = files_to_check)
j <- 1
#print("D")
while (j <= prod(files_to_check) && is.null(first_file)) {
selector_indices <- which(sub_array_of_files_to_check == j, arr.ind = TRUE)[1, ]
selectors <- sapply(1:length(defined_file_dims),
function (x) {
vector_to_pick <- 1
if (defined_file_dims[x] %in% names(depending_file_dims)) {
vector_to_pick <- selector_indices[which(defined_file_dims == depending_file_dims[[defined_file_dims[x]]])]
}
dat_selectors[defined_file_dims][[x]][[vector_to_pick]][selector_indices[x]]
})
replace_values[defined_file_dims] <- selectors
file_path <- .ReplaceVariablesInString(dat[[i]][['path']], replace_values)
file_path <- Sys.glob(file_path)
if (length(file_path) > 0) {
first_file <- file_path[1]
first_file_selectors <- selectors
}
j <- j + 1
}
#print("E")
# Start looking for values for the non-explicitly defined selectors.
if (is.null(first_file)) {
.warning(paste0("No found files for the datset '", dat[[i]][['name']],
"'. Provide existing selectors for the file dimensions ",
" or check and correct its path pattern: ", dat[[i]][['path']]))
} else {
dataset_has_files[i] <- TRUE
## TODO: Improve message here if no variable found:
if (length(undefined_file_dims) > 0) {
# Looking for the first values, parsed from first_file.
first_values <- vector('list', length = length(undefined_file_dims))
names(first_values) <- undefined_file_dims
found_values <- 0
stop <- FALSE
try_dim <- 1
last_success <- 1
while ((found_values < length(undefined_file_dims)) && !stop) {
u_file_dim <- undefined_file_dims[try_dim]
if (is.null(first_values[[u_file_dim]])) {
path_with_globs_and_tag <- .ReplaceVariablesInString(dat[[i]][['path']],
replace_values[-which(file_dims == u_file_dim)],
allow_undefined_key_vars = TRUE)
found_value <- .FindTagValue(path_with_globs_and_tag,
first_file, u_file_dim)
if (!is.null(found_value)) {
found_values <- found_values + 1
last_success <- try_dim
first_values[[u_file_dim]] <- found_value
replace_values[[u_file_dim]] <- found_value
}
}
try_dim <- (try_dim %% length(undefined_file_dims)) + 1
if (try_dim == last_success) {
stop <- TRUE
}
}
if (found_values < length(undefined_file_dims)) {
stop(paste0("Path pattern of dataset '", dat[[i]][['name']],
"' is too complex. Could not automatically ",
"detect values for all non-explicitly defined ",
"indices. Check its pattern: ", dat[[i]][['path']]))
}
## TODO: Replace ReplaceGlobExpressions by looped call to FindTagValue? As done above
## Maybe it can solve more cases actually. I got warnings in ReplGlobExp with a typical
## cmor case, requesting all members and chunks for fixed var and sdate. Not fixing
## sdate raised 'too complex' error.
# Replace shell globs in path pattern and keep the file_dims as tags
dat[[i]][['path']] <- .ReplaceGlobExpressions(dat[[i]][['path']], first_file, replace_values,
file_dims, dat[[i]][['name']], FALSE)
# Now time to look for the available values for the non
# explicitly defined selectors for the file dimensions.
#print("H")
# Check first the ones that do not depend on others.
ufd <- c(undefined_file_dims[which(!(undefined_file_dims %in% names(depending_file_dims)))],
undefined_file_dims[which(undefined_file_dims %in% names(depending_file_dims))])
for (u_file_dim in ufd) {
replace_values[undefined_file_dims] <- first_values
replace_values[[u_file_dim]] <- '*'
depended_dim <- NULL
depended_dim_values <- NA
selectors <- dat_selectors[[u_file_dim]][[1]]
if (u_file_dim %in% names(depending_file_dims)) {
depended_dim <- depending_file_dims[[u_file_dim]]
depended_dim_values <- dat_selectors[[depended_dim]][[1]]
dat_selectors[[u_file_dim]] <- vector('list', length = length(depended_dim_values))
names(dat_selectors[[u_file_dim]]) <- depended_dim_values
} else {
dat_selectors[[u_file_dim]] <- list()
}
if (u_file_dim %in% unlist(depending_file_dims)) {
depending_dims <- names(depending_file_dims)[which(sapply(depending_file_dims, function(x) u_file_dim %in% x))]
replace_values[depending_dims] <- rep('*', length(depending_dims))
}
for (j in 1:length(depended_dim_values)) {
parsed_values <- c()
if (!is.null(depended_dim)) {
replace_values[[depended_dim]] <- depended_dim_values[j]
}
path_with_globs <- .ReplaceVariablesInString(dat[[i]][['path']], replace_values)
found_files <- Sys.glob(path_with_globs)
## TODO: Enhance this error message, or change by warning.
## Raises if a wrong sdate is specified, for example.
if (length(found_files) == 0) {
.warning(paste0("Could not find files for any '", u_file_dim,
"' for '", depended_dim, "' = '",
depended_dim_values[j], "'."))
dat_selectors[[u_file_dim]][[j]] <- NA
} else {
for (found_file in found_files) {
path_with_globs_and_tag <- .ReplaceVariablesInString(dat[[i]][['path']],
replace_values[-which(file_dims == u_file_dim)],
allow_undefined_key_vars = TRUE)
parsed_values <- c(parsed_values,
.FindTagValue(path_with_globs_and_tag, found_file,
u_file_dim))
}
dat_selectors[[u_file_dim]][[j]] <- selector_checker(selectors = selectors,
var = unique(parsed_values),
return_indices = FALSE)
# Take chunk if needed
dat_selectors[[u_file_dim]][[j]] <- dat_selectors[[u_file_dim]][[j]][chunk_indices(length(dat_selectors[[u_file_dim]][[j]]),
chunks[[u_file_dim]]['chunk'],
chunks[[u_file_dim]]['n_chunks'],
u_file_dim)]
}
}
}
#print("I")
} else {
dat[[i]][['path']] <- .ReplaceGlobExpressions(dat[[i]][['path']], first_file, replace_values,
defined_file_dims, dat[[i]][['name']], FALSE)
}
}
}
# Now fetch for the first available file
if (dataset_has_files[i]) {
known_dims <- file_dims
} else {
known_dims <- defined_file_dims
}
replace_values <- vector('list', length = length(known_dims))
names(replace_values) <- known_dims
files_to_load <- sapply(dat_selectors[known_dims], function(x) length(x[[1]]))
files_to_load[found_pattern_dim] <- 1
sub_array_of_files_to_load <- array(1:prod(files_to_load),
dim = files_to_load)
names(dim(sub_array_of_files_to_load)) <- known_dims
sub_array_of_not_found_files <- array(!dataset_has_files[i],
dim = files_to_load)
names(dim(sub_array_of_not_found_files)) <- known_dims
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
j <- 1
while (j <= prod(files_to_load)) {
selector_indices <- which(sub_array_of_files_to_load == j, arr.ind = TRUE)[1, ]
names(selector_indices) <- known_dims
selectors <- sapply(1:length(known_dims),
function (x) {
vector_to_pick <- 1
if (known_dims[x] %in% names(depending_file_dims)) {
vector_to_pick <- selector_indices[which(known_dims == depending_file_dims[[known_dims[x]]])]
}
dat_selectors[known_dims][[x]][[vector_to_pick]][selector_indices[x]]
})
names(selectors) <- known_dims
replace_values[known_dims] <- selectors
if (!dataset_has_files[i]) {
if (any(is.na(selectors))) {
replace_values <- replace_values[-which(names(replace_values) %in% names(selectors[which(is.na(selectors))]))]
}
file_path <- .ReplaceVariablesInString(dat[[i]][['path']], replace_values, TRUE)
sub_array_of_files_to_load[j] <- file_path
} else {
if (any(is.na(selectors))) {
replace_values <- replace_values[-which(names(replace_values) %in% names(selectors[which(is.na(selectors))]))]
file_path <- .ReplaceVariablesInString(dat[[i]][['path']], replace_values, TRUE)
sub_array_of_files_to_load[j] <- file_path
sub_array_of_not_found_files[j] <- TRUE
} else {
file_path <- .ReplaceVariablesInString(dat[[i]][['path']], replace_values)
sub_array_of_files_to_load[j] <- file_path
if (is.null(indices_of_first_files_with_data[[i]])) {
if (!(length(grep("^http", file_path)) > 0)) {
if (!file.exists(file_path)) {
file_path <- NULL
}
}
if (!is.null(file_path)) {
test_file <- NULL
## TODO: suppress error messages
test_file <- file_opener(file_path)
if (!is.null(test_file)) {
selector_indices[which(known_dims == found_pattern_dim)] <- i
indices_of_first_files_with_data[[i]] <- selector_indices
selectors_of_first_files_with_data[[i]] <- selectors
file_closer(test_file)
}
}
}
}
}
j <- j + 1
}
# Extend array as needed progressively
if (is.null(array_of_files_to_load)) {
array_of_files_to_load <- sub_array_of_files_to_load
array_of_not_found_files <- sub_array_of_not_found_files
} else {
array_of_files_to_load <- .MergeArrays(array_of_files_to_load, sub_array_of_files_to_load,
along = found_pattern_dim)
## TODO: file_dims, and variables like that.. are still ok now? I don't think so
array_of_not_found_files <- .MergeArrays(array_of_not_found_files, sub_array_of_not_found_files,
along = found_pattern_dim)
}
dat[[i]][['selectors']] <- dat_selectors
}
if (all(sapply(indices_of_first_files_with_data, is.null))) {
stop("No data files found for any of the specified datasets.")
}
########################### READING INNER DIMS. #############################
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
#print("J")
## TODO: To be run in parallel (local multi-core)
# Now time to work out the inner file dimensions.
# First pick the requested variables.
dims_to_iterate <- NULL
for (return_var in names(return_vars)) {
dims_to_iterate <- unique(c(dims_to_iterate, return_vars[[return_var]]))
}
if (found_pattern_dim %in% dims_to_iterate) {
dims_to_iterate <- dims_to_iterate[-which(dims_to_iterate == found_pattern_dim)]
}
common_return_vars <- NULL
common_first_found_file <- NULL
common_return_vars_pos <- NULL
if (length(return_vars) > 0) {
common_return_vars_pos <- which(sapply(return_vars, function(x) !(found_pattern_dim %in% x)))
}
if (length(common_return_vars_pos) > 0) {
common_return_vars <- return_vars[common_return_vars_pos]
return_vars <- return_vars[-common_return_vars_pos]
common_first_found_file <- rep(FALSE, length(which(sapply(common_return_vars, length) == 0)))
names(common_first_found_file) <- names(common_return_vars[which(sapply(common_return_vars, length) == 0)])
}
return_vars <- lapply(return_vars,
function(x) {
if (found_pattern_dim %in% x) {
x[-which(x == found_pattern_dim)]
} else {
x
}
})
if (length(common_return_vars) > 0) {
picked_common_vars <- vector('list', length = length(common_return_vars))
names(picked_common_vars) <- names(common_return_vars)
} else {
picked_common_vars <- NULL
}
picked_common_vars_ordered <- picked_common_vars
picked_common_vars_unorder_indices <- picked_common_vars
picked_vars <- vector('list', length = length(dat))
names(picked_vars) <- dat_names
picked_vars_ordered <- picked_vars
picked_vars_unorder_indices <- picked_vars
for (i in 1:length(dat)) {
if (dataset_has_files[i]) {
# Put all selectors in a list of a single list/vector of selectors.
# The dimensions that go across files will later be extended to have
# lists of lists/vectors of selectors.
for (inner_dim in expected_inner_dims[[i]]) {
if (!is.list(dat[[i]][['selectors']][[inner_dim]]) ||
(is.list(dat[[i]][['selectors']][[inner_dim]]) &&
length(dat[[i]][['selectors']][[inner_dim]]) == 2 &&
is.null(names(dat[[i]][['selectors']][[inner_dim]])))) {
dat[[i]][['selectors']][[inner_dim]] <- list(dat[[i]][['selectors']][[inner_dim]])
}
}
if (length(return_vars) > 0) {
picked_vars[[i]] <- vector('list', length = length(return_vars))
names(picked_vars[[i]]) <- names(return_vars)
picked_vars_ordered[[i]] <- picked_vars[[i]]
picked_vars_unorder_indices[[i]] <- picked_vars[[i]]
}
indices_of_first_file <- as.list(indices_of_first_files_with_data[[i]])
array_file_dims <- sapply(dat[[i]][['selectors']][found_file_dims[[i]]], function(x) length(x[[1]]))
names(array_file_dims) <- found_file_dims[[i]]
if (length(dims_to_iterate) > 0) {
indices_of_first_file[dims_to_iterate] <- lapply(array_file_dims[dims_to_iterate], function(x) 1:x)
}
array_of_var_files <- do.call('[', c(list(x = array_of_files_to_load), indices_of_first_file, list(drop = FALSE)))
array_of_var_indices <- array(1:length(array_of_var_files), dim = dim(array_of_var_files))
array_of_not_found_var_files <- do.call('[', c(list(x = array_of_not_found_files), indices_of_first_file, list(drop = FALSE)))
previous_indices <- rep(-1, length(indices_of_first_file))
names(previous_indices) <- names(indices_of_first_file)
first_found_file <- NULL
if (length(return_vars) > 0) {
first_found_file <- rep(FALSE, length(which(sapply(return_vars, length) == 0)))
names(first_found_file) <- names(return_vars[which(sapply(return_vars, length) == 0)])
}
for (j in 1:length(array_of_var_files)) {
current_indices <- which(array_of_var_indices == j, arr.ind = TRUE)[1, ]
names(current_indices) <- names(indices_of_first_file)
if (!is.na(array_of_var_files[j]) && !array_of_not_found_var_files[j]) {
changed_dims <- which(current_indices != previous_indices)
vars_to_read <- NULL
if (length(return_vars) > 0) {
vars_to_read <- names(return_vars)[sapply(return_vars, function(x) any(names(changed_dims) %in% x))]
}
if (!is.null(first_found_file)) {
if (any(!first_found_file)) {
vars_to_read <- c(vars_to_read, names(first_found_file[which(!first_found_file)]))
}
}
if ((i == 1) && (length(common_return_vars) > 0)) {
vars_to_read <- c(vars_to_read, names(common_return_vars)[sapply(common_return_vars, function(x) any(names(changed_dims) %in% x))])
}
if (!is.null(common_first_found_file)) {
if (any(!common_first_found_file)) {
vars_to_read <- c(vars_to_read, names(common_first_found_file[which(!common_first_found_file)]))
}
}
file_object <- file_opener(array_of_var_files[j])
if (!is.null(file_object)) {
for (var_to_read in vars_to_read) {
if (var_to_read %in% unlist(var_params)) {
associated_dim_name <- names(var_params)[which(unlist(var_params) == var_to_read)]
var_name_to_reader <- var_to_read
names(var_name_to_reader) <- 'var'
var_dims <- file_dim_reader(NULL, file_object, var_name_to_reader, NULL,
synonims)
# file_dim_reader returns dimension names as found in the file.
# Need to translate accoridng to synonims:
names(var_dims) <- sapply(names(var_dims),
function(x) {
which_entry <- which(sapply(synonims, function(y) x %in% y))
if (length(which_entry) > 0) {
names(synonims)[which_entry]
} else {
x
}
})
if (!is.null(var_dims)) {
var_file_dims <- NULL
if (var_to_read %in% names(common_return_vars)) {
var_to_check <- common_return_vars[[var_to_read]]
var_to_check <- return_vars[[var_to_read]]
if (any(names(dim(array_of_files_to_load)) %in% var_to_check)) {
var_file_dims <- dim(array_of_files_to_load)[which(names(dim(array_of_files_to_load)) %in%
var_to_check)]
}
if (((var_to_read %in% names(common_return_vars)) &&
is.null(picked_common_vars[[var_to_read]])) ||
((var_to_read %in% names(return_vars)) &&
is.null(picked_vars[[i]][[var_to_read]]))) {
if (any(names(var_file_dims) %in% names(var_dims))) {
stop("Found a requested var in 'return_var' requested for a ",
"file dimension which also appears in the dimensions of ",
"the variable inside the file.\n", array_of_var_files[j])
special_types <- list('POSIXct' = as.POSIXct, 'POSIXlt' = as.POSIXlt,
'Date' = as.Date)
first_sample <- file_var_reader(NULL, file_object, NULL,
var_to_read, synonims)
if (any(class(first_sample) %in% names(special_types))) {
array_size <- prod(c(var_file_dims, var_dims))
new_array <- rep(special_types[[class(first_sample)[1]]](NA), array_size)
dim(new_array) <- c(var_file_dims, var_dims)
} else {
new_array <- array(dim = c(var_file_dims, var_dims))
attr(new_array, 'variables') <- attr(first_sample, 'variables')
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
if (var_to_read %in% names(common_return_vars)) {
picked_common_vars[[var_to_read]] <- new_array
pick_ordered <- FALSE
if (var_to_read %in% unlist(var_params)) {
if (associated_dim_name %in% names(dim_reorder_param) && !aiat) {
picked_common_vars_ordered[[var_to_read]] <- new_array
pick_ordered <- TRUE
}
}
if (!pick_ordered) {
picked_common_vars_ordered[[var_to_read]] <- NULL
}
} else {
picked_vars[[i]][[var_to_read]] <- new_array
pick_ordered <- FALSE
if (var_to_read %in% unlist(var_params)) {
if (associated_dim_name %in% names(dim_reorder_params) && !aiat) {
picked_vars_ordered[[i]][[var_to_read]] <- new_array
pick_ordered <- TRUE
}
}
if (!pick_ordered) {
picked_vars_ordered[[i]][[var_to_read]] <- NULL
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
} else {
if (var_to_read %in% names(common_return_vars)) {
array_var_dims <- dim(picked_common_vars[[var_to_read]])
} else {
array_var_dims <- dim(picked_vars[[i]][[var_to_read]])
}
if (any(names(array_var_dims) %in% names(var_file_dims))) {
array_var_dims <- array_var_dims[-which(names(array_var_dims) %in% names(var_file_dims))]
}
if (names(array_var_dims) != names(var_dims)) {
stop("Error while reading the variable '", var_to_read, "' from ",
"the file. Dimensions do not match.\nExpected ",
paste(paste0("'", names(array_var_dims), "'"),
collapse = ', '), " but found ",
paste(paste0("'", names(var_dims), "'"),
collapse = ', '), ".\n", array_of_var_files[j])
}
if (any(var_dims > array_var_dims)) {
stop("Error while reading the variable '", var_to_read, "' from ",
"the file. Found size (", paste(var_dims, collapse = ' x '),
") is greater than expected maximum size (",
array_var_dims, ").")
}
}
var_store_indices <- c(as.list(current_indices[names(var_file_dims)]), lapply(var_dims, function(x) 1:x))
var_values <- file_var_reader(NULL, file_object, NULL, var_to_read, synonims)
if (var_to_read %in% unlist(var_params)) {
if ((associated_dim_name %in% names(dim_reorder_params)) && !aiat) {
## Is this check really needed?
if (length(dim(var_values)) > 1) {
stop("Requested a '", associated_dim_name, "_reorder' for a dimension ",
"whose coordinate variable that has more than 1 dimension. This is ",
"not supported.")
}
ordered_var_values <- dim_reorder_params[[associated_dim_name]](var_values)
attr(ordered_var_values$x, 'variables') <- attr(var_values, 'variables')
if (!all(c('x', 'ix') %in% names(ordered_var_values))) {
stop("All the dimension reorder functions must return a list with the components 'x' and 'ix'.")
}
# Save the indices to reorder back the ordered variable values.
# This will be used to define the first round indices.
unorder <- sort(ordered_var_values$ix, index.return = TRUE)$ix
if (var_to_read %in% names(common_return_vars)) {
picked_common_vars_ordered[[var_to_read]] <- do.call('[<-',
c(list(x = picked_common_vars_ordered[[var_to_read]]),
var_store_indices,
list(value = ordered_var_values$x)))
picked_common_vars_unorder_indices[[var_to_read]] <- do.call('[<-',
c(list(x = picked_common_vars_unorder_indices[[var_to_read]]),
var_store_indices,
list(value = unorder)))
} else {
picked_vars_ordered[[i]][[var_to_read]] <- do.call('[<-',
c(list(x = picked_vars_ordered[[i]][[var_to_read]]),
var_store_indices,
list(value = ordered_var_values$x)))
picked_vars_unorder_indices[[i]][[var_to_read]] <- do.call('[<-',
c(list(x = picked_vars_unorder_indices[[i]][[var_to_read]]),
var_store_indices,
list(value = unorder)))
}
}
}
if (var_to_read %in% names(common_return_vars)) {
picked_common_vars[[var_to_read]] <- do.call('[<-',
c(list(x = picked_common_vars[[var_to_read]]),
var_store_indices,
list(value = var_values)))
picked_vars[[i]][[var_to_read]] <- do.call('[<-',
c(list(x = picked_vars[[i]][[var_to_read]]),
var_store_indices,
list(value = var_values)))
if (var_to_read %in% names(first_found_file)) {
first_found_file[var_to_read] <- TRUE
if (var_to_read %in% names(common_first_found_file)) {
common_first_found_file[var_to_read] <- TRUE
stop("Could not find variable '", var_to_read,
"' in the file ", array_of_var_files[j])
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
}
}
previous_indices <- current_indices
}
}
}
# Once we have the variable values, we can work out the indices
# for the implicitly defined selectors.
#
# Trnasforms a vector of indices v expressed in a world of
# length N from 1 to N, into a world of length M, from
# 1 to M. Repeated adjacent indices are collapsed.
transform_indices <- function(v, n, m) {
#unique2 turns e.g. 1 1 2 2 2 3 3 1 1 1 into 1 2 3 1
unique2 <- function(v) {
if (length(v) < 2) {
v
} else {
v[c(1, v[2:length(v)] - v[1:(length(v) - 1)]) != 0]
}
}
unique2(round(((v - 1) / (n - 1)) * (m - 1))) + 1 # this rounding may generate 0s. what then?
}
beta <- transform_extra_cells
dims_to_crop <- vector('list')
transformed_vars <- vector('list', length = length(dat))
names(transformed_vars) <- dat_names
transformed_vars_ordered <- transformed_vars
transformed_vars_unorder_indices <- transformed_vars
transformed_common_vars <- NULL
transformed_common_vars_ordered <- NULL
transformed_common_vars_unorder_indices <- NULL
for (i in 1:length(dat)) {
if (dataset_has_files[i]) {
indices <- indices_of_first_files_with_data[[i]]
if (!is.null(indices)) {
file_path <- do.call("[", c(list(array_of_files_to_load), as.list(indices_of_first_files_with_data[[i]])))
# The following 5 lines should go several lines below, but were moved
# here for better performance.
# If any of the dimensions comes without defining variable, then we read
# the data dimensions.
if (length(unlist(var_params[expected_inner_dims[[i]]])) < length(expected_inner_dims[[i]])) {
file_to_open <- file_path
data_dims <- file_dim_reader(file_to_open, NULL, selectors_of_first_files_with_data[[i]],
lapply(dat[[i]][['selectors']][expected_inner_dims[[i]]], '[[', 1),
synonims)
# file_dim_reader returns dimension names as found in the file.
# Need to translate accoridng to synonims:
names(data_dims) <- sapply(names(data_dims),
function(x) {
which_entry <- which(sapply(synonims, function(y) x %in% y))
if (length(which_entry) > 0) {
names(synonims)[which_entry]
} else {
x
}
})
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
}
# Transform the variables if needed and keep them apart.
if (!is.null(transform) && (length(transform_vars) > 0)) {
if (!all(transform_vars %in% c(names(picked_vars[[i]]), names(picked_common_vars)))) {
stop("Could not find all the required variables in 'transform_vars' ",
"for the dataset '", dat[[i]][['name']], "'.")
}
vars_to_transform <- NULL
picked_vars_to_transform <- which(names(picked_vars[[i]]) %in% transform_vars)
if (length(picked_vars_to_transform) > 0) {
picked_vars_to_transform <- names(picked_vars[[i]])[picked_vars_to_transform]
new_vars_to_transform <- picked_vars[[i]][picked_vars_to_transform]
which_are_ordered <- which(!sapply(picked_vars_ordered[[i]][picked_vars_to_transform], is.null))
if (length(which_are_ordered) > 0) {
new_vars_to_transform[which_are_ordered] <- picked_vars_ordered[[i]][which_are_ordered]
}
vars_to_transform <- c(vars_to_transform, new_vars_to_transform)
}
picked_common_vars_to_transform <- which(names(picked_common_vars) %in% transform_vars)
if (length(picked_common_vars_to_transform) > 0) {
picked_common_vars_to_transform <- names(picked_common_vars)[picked_common_vars_to_transform]
new_vars_to_transform <- picked_common_vars[[i]][picked_common_vars_to_transform]
which_are_ordered <- which(!sapply(picked_common_vars_ordered[[i]][picked_common_vars_to_transform], is.null))
if (length(which_are_ordered) > 0) {
new_vars_to_transform[which_are_ordered] <- picked_common_vars_ordered[[i]][which_are_ordered]
}
vars_to_transform <- c(vars_to_transform, new_vars_to_transform)
}
# Transform the variables
transformed_data <- do.call(transform, c(list(data_array = NULL,
variables = vars_to_transform,
file_selectors = selectors_of_first_files_with_data[[i]]),
transform_params))
# Discard the common transformed variables if already transformed before
if (!is.null(transformed_common_vars)) {
common_ones <- which(names(picked_common_vars) %in% names(transformed_data$variables))
if (length(common_ones) > 0) {
transformed_data$variables <- transformed_data$variables[-common_ones]
}
}
transformed_vars[[i]] <- list()
transformed_vars_ordered[[i]] <- list()
transformed_vars_unorder_indices[[i]] <- list()
# Order the transformed variables if needed
# 'var_to_read' should be 'transformed_var', but is kept to reuse the same code as above.
for (var_to_read in names(transformed_data$variables)) {
if (var_to_read %in% unlist(var_params)) {
associated_dim_name <- names(var_params)[which(unlist(var_params) == var_to_read)]
if ((associated_dim_name %in% names(dim_reorder_params)) && aiat) {
## Is this check really needed?
if (length(dim(transformed_data$variables[[associated_dim_name]])) > 1) {
stop("Requested a '", associated_dim_name, "_reorder' for a dimension ",
"whose coordinate variable that has more than 1 dimension (after ",
"transform). This is not supported.")
}
ordered_var_values <- dim_reorder_params[[associated_dim_name]](transformed_data$variables[[associated_dim_name]])
attr(ordered_var_values, 'variables') <- attr(transformed_data$variables[[associated_dim_name]], 'variables')
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
if (!all(c('x', 'ix') %in% names(ordered_var_values))) {
stop("All the dimension reorder functions must return a list with the components 'x' and 'ix'.")
}
# Save the indices to reorder back the ordered variable values.
# This will be used to define the first round indices.
unorder <- sort(ordered_var_values$ix, index.return = TRUE)$ix
if (var_to_read %in% names(picked_common_vars)) {
transformed_common_vars_ordered[[var_to_read]] <- ordered_var_values$x
transformed_common_vars_unorder_indices[[var_to_read]] <- unorder
} else {
transformed_vars_ordered[[i]][[var_to_read]] <- ordered_var_values$x
transformed_vars_unorder_indices[[i]][[var_to_read]] <- unorder
}
}
}
}
transformed_picked_vars <- which(names(picked_vars[[i]]) %in% names(transformed_data$variables))
if (length(transformed_picked_vars) > 0) {
transformed_picked_vars <- names(picked_vars[[i]])[transformed_picked_vars]
transformed_vars[[i]][transformed_picked_vars] <- transformed_data$variables[transformed_picked_vars]
}
if (is.null(transformed_common_vars)) {
transformed_picked_common_vars <- which(names(picked_common_vars) %in% names(transformed_data$variables))
if (length(transformed_picked_common_vars) > 0) {
transformed_picked_common_vars <- names(picked_common_vars)[transformed_picked_common_vars]
transformed_common_vars <- transformed_data$variables[transformed_picked_common_vars]
}
}
}
# Once the variables are transformed, we compute the indices to be
# taken for each inner dimension.
# In all cases, indices will have to be computed to know which data
# values to take from the original data for each dimension (if a
# variable is specified for that dimension, it will be used to
# convert the provided selectors into indices). These indices are
# referred to as 'first round of indices'.
# The taken data will then be transformed if needed, together with
# the dimension variable if specified, and, in that case, indices
# will have to be computed again to know which values to take from the
# transformed data. These are the 'second round of indices'. In the
# case there is no transformation, the second round of indices will
# be all the available indices, i.e. from 1 to the number of taken
# values with the first round of indices.
for (inner_dim in expected_inner_dims[[i]]) {
if (debug) {
print("-> DEFINING INDICES FOR INNER DIMENSION:")
print(inner_dim)
}
file_dim <- NULL
if (inner_dim %in% unlist(inner_dims_across_files)) {
file_dim <- names(inner_dims_across_files)[which(sapply(inner_dims_across_files, function(x) inner_dim %in% x))[1]]
chunk_amount <- length(dat[[i]][['selectors']][[file_dim]][[1]])
names(chunk_amount) <- file_dim
} else {
chunk_amount <- 1
}
# In the special case that the selectors for a dimension are 'all', 'first', ...
# and chunking (dividing in more than 1 chunk) is requested, the selectors are
# replaced for equivalent indices.
if ((dat[[i]][['selectors']][[inner_dim]][[1]] %in% c('all', 'first', 'last')) &&
(chunks[[inner_dim]]['n_chunks'] != 1)) {
selectors <- dat[[i]][['selectors']][[inner_dim]][[1]]
if (selectors == 'all') {
selectors <- indices(1:(data_dims[[inner_dim]] * chunk_amount))
} else if (selectors == 'first') {
selectors <- indices(1)
} else {
selectors <- indices(data_dims[[inner_dim]] * chunk_amount)
}
dat[[i]][['selectors']][[inner_dim]][[1]] <- selectors
}
# The selectors for the inner dimension are taken.
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
selector_array <- dat[[i]][['selectors']][[inner_dim]][[1]]
if (debug) {
if (inner_dim %in% dims_to_check) {
print(paste0("-> DEBUG MESSAGES FOR THE DATASET", i, " AND INNER DIMENSION '", inner_dim, "':"))
print("-> STRUCTURE OF SELECTOR ARRAY:")
print(str(selector_array))
print("-> PICKED VARS:")
print(picked_vars)
print("-> TRANSFORMED VARS:")
print(transformed_vars)
}
}
if (is.null(dim(selector_array))) {
dim(selector_array) <- length(selector_array)
}
if (is.null(names(dim(selector_array)))) {
if (length(dim(selector_array)) == 1) {
names(dim(selector_array)) <- inner_dim
} else {
stop("Provided selector arrays must be provided with dimension ",
"names. Found an array of selectors without dimension names ",
"for the dimension '", inner_dim, "'.")
}
}
selectors_are_indices <- FALSE
if (!is.null(attr(selector_array, 'indices'))) {
if (!is.logical(attr(selector_array, 'indices'))) {
stop("The atribute 'indices' for the selectors for the dimension '",
inner_dim, "' must be TRUE or FALSE.")
}
selectors_are_indices <- attr(selector_array, 'indices')
}
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
taken_chunks <- rep(FALSE, chunk_amount)
selector_file_dims <- 1
if (any(found_file_dims[[i]] %in% names(dim(selector_array)))) {
selector_file_dims <- dim(selector_array)[which(names(dim(selector_array)) %in% found_file_dims[[i]])]
}
selector_inner_dims <- dim(selector_array)[which(!(names(dim(selector_array)) %in% found_file_dims[[i]]))]
var_with_selectors <- NULL
var_with_selectors_name <- var_params[[inner_dim]]
var_ordered <- NULL
var_unorder_indices <- NULL
with_transform <- FALSE
# If the selectors come with an associated variable
if (!is.null(var_with_selectors_name)) {
if ((var_with_selectors_name %in% transform_vars) && (!is.null(transform))) {
with_transform <- TRUE
if (!is.null(file_dim)) {
stop("Requested a transformation over the dimension '",
inner_dim, "', wich goes across files. This feature ",
"is not supported. Either do the request without the ",
"transformation or request it over dimensions that do ",
"not go across files.")
}
}
if (debug) {
if (inner_dim %in% dims_to_check) {
print("-> NAME OF THE VARIABLE WITH THE SELECTOR VALUES FOR THE CURRENT INNER DIMENSION:")
print(var_with_selectors_name)
print("-> NAMES OF THE VARIABLES TO BE TRANSFORMED:")
print(transform_vars)
print("-> STRUCTURE OF THE TRANSFORMATION FUNCTION:")
print(str(transform))
}
}
if (var_with_selectors_name %in% names(picked_vars[[i]])) {
var_with_selectors <- picked_vars[[i]][[var_with_selectors_name]]
var_ordered <- picked_vars_ordered[[i]][[var_with_selectors_name]]
var_unorder_indices <- picked_vars_unorder_indices[[i]][[var_with_selectors_name]]
} else if (var_with_selectors_name %in% names(picked_common_vars)) {
var_with_selectors <- picked_common_vars[[var_with_selectors_name]]
var_ordered <- picked_common_vars_ordered[[var_with_selectors_name]]
var_unorder_indices <- picked_common_vars_unorder_indices[[var_with_selectors_name]]
}
n <- prod(dim(var_with_selectors))
if (is.null(var_unorder_indices)) {
var_unorder_indices <- 1:n
}
if (with_transform) {
if (var_with_selectors_name %in% names(transformed_vars[[i]])) {
m <- prod(dim(transformed_vars[[i]][[var_with_selectors_name]]))
var_with_selectors <- transformed_vars[[i]][[var_with_selectors_name]]
var_ordered <- transformed_vars_ordered[[i]][[var_with_selectors_name]]
var_unorder_indices <- transformed_vars_unorder_indices[[i]][[var_with_selectors_name]]
}
} else if (var_with_selectors_name %in% names(transformed_common_vars)) {
m <- prod(dim(transformed_common_vars[[var_with_selectors_name]]))
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
var_with_selectors <- transformed_common_vars[[var_with_selectors_name]]
var_ordered <- transformed_common_vars_ordered[[var_with_selectors_name]]
var_unorder_indices <- transformed_common_vars_unorder_indices[[var_with_selectors_name]]
}
}
if (is.null(var_unorder_indices)) {
var_unorder_indices <- 1:m
}
}
if (debug) {
if (inner_dim %in% dims_to_check) {
print("-> SIZE OF ORIGINAL VARIABLE:")
print(n)
print("-> SIZE OF TRANSFORMED VARIABLE:")
if (with_transform) print(m)
print("-> STRUCTURE OF ORDERED VAR:")
print(str(var_ordered))
print("-> UNORDER INDICES:")
print(var_unorder_indices)
}
}
var_dims <- dim(var_with_selectors)
var_file_dims <- 1
if (any(names(var_dims) %in% found_file_dims[[i]])) {
if (with_transform) {
stop("Requested transformation for inner dimension '",
inner_dim, "' but provided selectors for such dimension ",
"over one or more file dimensions. This is not ",
"supported. Either request no transformation for the ",
"dimension '", inner_dim, "' or specify the ",
"selectors for this dimension without the file dimensions.")
}
var_file_dims <- var_dims[which(names(var_dims) %in% found_file_dims[[i]])]
var_dims <- var_dims[-which(names(var_dims) %in% found_file_dims[[i]])]
}
## # Keep the selectors if they correspond to a variable that will be transformed.
## if (with_transform) {
## if (var_with_selectors_name %in% names(picked_vars[[i]])) {
## transformed_var_with_selectors <- transformed_vars[[i]][[var_with_selectors_name]]
## } else if (var_with_selectors_name %in% names(picked_common_vars)) {
## transformed_var_with_selectors <- transformed_common_vars[[var_with_selectors_name]]
## }
## transformed_var_dims <- dim(transformed_var_with_selectors)
## transformed_var_file_dims <- 1
## if (any(names(transformed_var_dims) %in% found_file_dims[[i]])) {
## transformed_var_file_dims <- transformed_var_dims[which(names(transformed_var_dims) %in% found_file_dims[[i]])]
## transformed_var_dims <- tranasformed_var_dims[-which(names(transformed_var_dims) %in% found_file_dims[[i]])]
## }
##if (inner_dim %in% dims_to_check) {
##print("111m")
##print(str(transformed_var_dims))
##}
##
## m <- prod(transformed_var_dims)
## }
# Work out var file dims and inner dims.
if (inner_dim %in% unlist(inner_dims_across_files)) {
#TODO: if (chunk_amount != number of chunks in selector_file_dims), crash
if (length(var_dims) > 1) {
stop("Specified a '", inner_dim, "_var' for the dimension '",
inner_dim, "', which goes across files (across '", file_dim,
"'). The specified variable, '", var_with_selectors_name, "', has more ",
"than one dimension and can not be used as selector variable. ",
"Select another variable or fix it in the files.")
}
}
## TODO HERE::
#- indices_of_first_files_with_data may change, because array is now extended
var_full_dims <- dim(var_with_selectors)
if (!(inner_dim %in% names(var_full_dims))) {
stop("Could not find the dimension '", inner_dim, "' in ",
"the file. Either change the dimension name in ",
"your request, adjust the parameter ",
"'dim_names_in_files' or fix the dimension name in ",
"the file.\n", file_path)
}
} else if ((is.numeric(selector_array) && selectors_are_indices) ||
(is.character(selector_array) && (length(selector_array) == 1) &&
(selector_array %in% c('all', 'first', 'last')) &&
!is.null(file_dim_reader))) {
#### TODO HERE::
###- indices_of_first_files_with_data may change, because array is now extended
# Lines moved above for better performance.
##data_dims <- file_dim_reader(file_path, NULL, selectors_of_first_files_with_data[[i]],
## lapply(dat[[i]][['selectors']][expected_inner_dims[[i]]], '[[', 1))
if (!(inner_dim %in% names(data_dims))) {
stop("Could not find the dimension '", inner_dim, "' in ",
"the file. Either change the dimension name in ",
"your request, adjust the parameter ",
"'dim_names_in_files' or fix the dimension name in ",
"the file.\n", file_path)
stop(paste0("Can not translate the provided selectors for '", inner_dim,
"' to numeric indices. Provide numeric indices and a ",
"'file_dim_reader' function, or a '", inner_dim,
"_var' in order to calculate the indices."))
}
# At this point, if no selector variable was provided, the variable
# data_dims has been populated. If a selector variable was provided,
# the variables var_dims, var_file_dims and var_full_dims have been
# populated instead.
fri <- first_round_indices <- NULL
sri <- second_round_indices <- NULL
# This variable will keep the indices needed to crop the transformed
# variable (the one that has been transformed without being subset
# with the first round indices).
tvi <- tranaformed_variable_indices <- NULL
ordered_fri <- NULL
ordered_sri <- NULL
if ((length(selector_array) == 1) && is.character(selector_array) &&
(selector_array %in% c('all', 'first', 'last')) &&
(chunks[[inner_dim]]['n_chunks'] == 1)) {
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
if (is.null(var_with_selectors_name)) {
fri <- vector('list', length = chunk_amount)
dim(fri) <- c(chunk_amount)
sri <- vector('list', length = chunk_amount)
dim(sri) <- c(chunk_amount)
if (selector_array == 'all') {
fri[] <- replicate(chunk_amount, list(1:(data_dims[inner_dim])))
taken_chunks <- rep(TRUE, chunk_amount)
#sri <- NULL
} else if (selector_array == 'first') {
fri[[1]] <- 1
taken_chunks[1] <- TRUE
#sri <- NULL
} else if (selector_array == 'last') {
fri[[chunk_amount]] <- data_dims[inner_dim]
taken_chunks[length(taken_chunks)] <- TRUE
#sri <- NULL
}
} else {
if ((!is.null(file_dim)) && !(file_dim %in% names(var_file_dims))) {
stop("The variable '", var_with_selectors_name, "' must also be ",
"requested for the file dimension '", file_dim, "' in ",
"this configuration.")
}
fri <- vector('list', length = prod(var_file_dims))
dim(fri) <- var_file_dims
ordered_fri <- fri
sri <- vector('list', length = prod(var_file_dims))
dim(sri) <- var_file_dims
ordered_sri <- sri
if (selector_array == 'all') {
# TODO: Populate ordered_fri
ordered_fri[] <- replicate(prod(var_file_dims), list(1:n))
fri[] <- replicate(prod(var_file_dims), list(var_unorder_indices[1:n]))
taken_chunks <- rep(TRUE, chunk_amount)
if (!with_transform) {
#fri[] <- replicate(prod(var_file_dims), list(1:n))
#taken_chunks <- rep(TRUE, chunk_amount)
#sri <- NULL
} else {
ordered_sri[] <- replicate(prod(var_file_dims), list(1:m))
sri[] <- replicate(prod(var_file_dims), list(1:m))
## var_file_dims instead??
#fri[] <- replicate(prod(var_file_dims), list(1:n))
#taken_chunks <- rep(TRUE, chunk_amount)
#sri[] <- replicate(prod(transformed_var_file_dims), list(1:m))
#} else {
#fri[] <- replicate(prod(var_file_dims), list(1:n))
#taken_chunks <- rep(TRUE, chunk_amount)
#sri[] <- replicate(prod(transformed_var_file_dims), list(1:m))
#}
tvi <- 1:m
}
} else if (selector_array == 'first') {
taken_chunks[1] <- TRUE
if (!with_transform) {
ordered_fri[[1]] <- 1
fri[[1]] <- var_unorder_indices[1]
#taken_chunks[1] <- TRUE
#sri <- NULL
} else {
ordered_fri[[1]] <- 1
fri[[1]] <- var_unorder_indices[1]
# TODO: TO BE IMPROVED
#taken_chunks[1] <- TRUE
ordered_sri[[1]] <- 1:ceiling(m / n)
sri[[1]] <- 1:ceiling(m / n)
tvi <- 1:ceiling(m / n)
} else {
ordered_fri[[1]] <- 1:ceiling(m / n)
fri[[1]] <- var_unorder_indices[1:ceiling(m / n)]
#taken_chunks[1] <- TRUE
ordered_sri[[1]] <- 1
sri[[1]] <- 1
tvi <- 1
}
}
} else if (selector_array == 'last') {
taken_chunks[length(taken_chunks)] <- TRUE
if (!with_transform) {
ordered_fri[[prod(var_file_dims)]] <- n
fri[[prod(var_file_dims)]] <- var_unorder_indices[n]
#taken_chunks[length(taken_chunks)] <- TRUE
#sri <- NULL
} else {
ordered_fri[[prod(var_file_dims)]] <- prod(var_dims)
fri[[prod(var_file_dims)]] <- var_unorder_indices[prod(var_dims)]
#taken_chunks[length(taken_chunks)] <- TRUE
ordered_sri[[prod(var_file_dims)]] <- 1:ceiling(m / n)
sri[[prod(var_file_dims)]] <- 1:ceiling(m / n)
# TODO: TO BE IMPROVED. THE TVI MAY BE WRONG IF THERE'S BEEN A REORDERING.
tvi <- 1:ceiling(m / n)
} else {
ordered_fri[[prod(var_file_dims)]] <- (n - ceiling(m / n) + 1):n
fri[[prod(var_file_dims)]] <- var_unorder_indices[(n - ceiling(m / n) + 1):n]
#taken_chunks[length(taken_chunks)] <- TRUE
ordered_sri[[prod(var_file_dims)]] <- 1
sri[[prod(var_file_dims)]] <- 1
tvi <- 1
}
}
}
}
# If the selectors are not 'all', 'first', 'last', ...
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
} else {
if (!is.null(var_with_selectors_name)) {
unmatching_file_dims <- which(!(names(var_file_dims) %in% names(selector_file_dims)))
if ((length(unmatching_file_dims) > 0)) {
raise_error <- FALSE
if (is.null(file_dim)) {
raise_error <- TRUE
} else {
if (!((length(unmatching_file_dims) == 1) &&
(names(var_file_dims)[unmatching_file_dims] == file_dim) &&
(inner_dim %in% names(selector_inner_dims)))) {
raise_error <- TRUE
}
}
if (raise_error) {
stop("Provided selectors for the dimension '", inner_dim, "' must have as many ",
"file dimensions as the variable the dimension is defined along, '",
var_with_selectors_name, "', with the exceptions of the file pattern dimension ('",
found_pattern_dim, "') and any depended file dimension (if specified as ",
"depended dimension in parameter 'inner_dims_across_files' and the ",
"depending file dimension is present in the provided selector array).")
}
}
if (any(dim(var_with_selectors)[names(selector_file_dims)] != selector_file_dims)) {
stop("Size of selector file dimensions must mach size of requested ",
"variable dimensions.")
}
}
## TODO: If var dimensions are not in the same order as selector dimensions, reorder
if (is.null(names(selector_file_dims))) {
if (is.null(file_dim)) {
fri_dims <- 1
} else {
fri_dims <- chunk_amount
names(fri_dims) <- file_dim
}
} else {
fri_dim_names <- names(selector_file_dims)
if (!is.null(file_dim)) {
fri_dim_names <- c(fri_dim_names, file_dim)
}
fri_dim_names <- found_file_dims[[i]][which(found_file_dims[[i]] %in% fri_dim_names)]
fri_dims <- rep(NA, length(fri_dim_names))
names(fri_dims) <- fri_dim_names
fri_dims[names(selector_file_dims)] <- selector_file_dims
if (!is.null(file_dim)) {
fri_dims[file_dim] <- chunk_amount
}
}
fri <- vector('list', length = prod(fri_dims))
dim(fri) <- fri_dims
sri <- vector('list', length = prod(fri_dims))
dim(sri) <- fri_dims
selector_file_dim_array <- array(1:prod(selector_file_dims), dim = selector_file_dims)
selector_store_position <- fri_dims
for (j in 1:prod(dim(selector_file_dim_array))) {
selector_indices_to_take <- which(selector_file_dim_array == j, arr.ind = TRUE)[1, ]
names(selector_indices_to_take) <- names(selector_file_dims)
selector_store_position[names(selector_indices_to_take)] <- selector_indices_to_take
sub_array_of_selectors <- Subset(selector_array, names(selector_indices_to_take),
as.list(selector_indices_to_take), drop = 'selected')
if (debug) {
if (inner_dim %in% dims_to_check) {
print("-> ITERATING OVER FILE DIMENSIONS OF THE SELECTORS.")
print("-> STRUCTURE OF A SUB ARRAY:")
print(str(sub_array_of_selectors))
print("-> STRUCTURE OF THE VARIABLE WITH SELECTORS:")
print(str(var_with_selectors))
print(dim(var_with_selectors))
}
}
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
sub_array_of_values <- NULL
#} else if (!is.null(var_ordered)) {
# sub_array_of_values <- var_ordered
} else {
if (length(var_file_dims) > 0) {
var_indices_to_take <- selector_indices_to_take[which(names(selector_indices_to_take) %in% names(var_file_dims))]
sub_array_of_values <- Subset(var_with_selectors, names(var_indices_to_take),
as.list(var_indices_to_take), drop = 'selected')
} else {
sub_array_of_values <- var_with_selectors
}
}
if (debug) {
if (inner_dim %in% dims_to_check) {
print("-> STRUCTURE OF THE SUB ARRAY FROM THE VARIABLE CORRESPONDING TO THE SUB ARRAY OF SELECTORS")
print(str(sub_array_of_values))
print(dim(sub_array_of_values))
print("-> NAME OF THE FILE DIMENSION THE CURRENT INNER DIMENSION EXTENDS ALONG:")
print(file_dim)
}
}
if ((!is.null(file_dim) && (file_dim %in% names(selector_file_dims))) || is.null(file_dim)) {
if (length(sub_array_of_selectors) > 0) {
if (debug) {
if (inner_dim %in% dims_to_check) {
print("-> THE INNER DIMENSION DOES NOT GO ACROSS ANY FILE DIMENSION OR IT DOES BUT IS IN THE PROVIDED SELECTOR ARRAY.")
}
}
if (!is.null(var_with_selectors_name)) {
max_allowed <- ifelse(aiat, m, n)
} else {
max_allowed <- data_dims[inner_dim]
}
if (any(na.omit(unlist(sub_array_of_selectors)) > max_allowed) ||
any(na.omit(unlist(sub_array_of_selectors)) < 1)) {
stop("Provided indices out of range for dimension '", inner_dim, "' ",
"for dataset '", dat[[i]][['name']], "' (accepted range: 1 to ",
# The selector_checker will return either a vector of indices or a list
# with the first and last desired indices.
goes_across_prime_meridian <- FALSE
if (!is.null(var_ordered) && !selectors_are_indices) {
if (!is.null(dim_reorder_params[[inner_dim]])) {
if (is.list(sub_array_of_selectors)) {
sub_array_reordered <- dim_reorder_params[[inner_dim]](unlist(sub_array_of_selectors))
sub_array_unorder <- sort(sub_array_reordered$ix, index.return = TRUE)$ix
sub_array_of_selectors <- as.list(sub_array_reordered$x[sub_array_unorder])
is_circular_dim <- attr(dim_reorder_params[[inner_dim]], 'circular')
if (!is.null(is_circular_dim)) {
if (is_circular_dim) {
goes_across_prime_meridian <- abs(sub_array_of_selectors[[1]]) > abs(sub_array_of_selectors[[2]])
## TODO: if (bounds[1] > bounds[2]) goes_across_prime_meridian <- !goes_across_prime_meridian
}
}
} else {
sub_array_of_selectors <- dim_reorder_params[[inner_dim]](sub_array_of_selectors)$x
}
}
sub_array_of_indices <- selector_checker(sub_array_of_selectors, var_ordered,
tolerance = if (aiat) {
NULL
} else {
tolerance_params[[inner_dim]]
})
sub_array_of_indices <- selector_checker(sub_array_of_selectors, sub_array_of_values,
tolerance = if (aiat) {
NULL
} else {
tolerance_params[[inner_dim]]
})
sub_array_of_indices <- sub_array_of_indices[chunk_indices(length(sub_array_of_indices),
chunks[[inner_dim]]['chunk'],
chunks[[inner_dim]]['n_chunks'],
inner_dim)]
# The sub_array_of_indices now contains numeric indices of the values to be taken.
if (debug) {
if (inner_dim %in% dims_to_check) {
print("-> TRANSFORMATION REQUESTED?")
print(with_transform)
print("-> BETA:")
print(beta)
}
}
if (with_transform) {
# If there is a transformation and selector values are provided, these
# selectors will be processed in the same way either if aiat = TRUE or
# aiat = FALSE.
## TODO: If sub_array_of_selectors was integer and aiat then... do what's commented 50 lines below.
## otherwise, do what's coded.
if (debug) {
if (inner_dim %in% dims_to_check) {
print("-> SELECTORS REQUESTED BEFORE TRANSFORM.")
}
}
if (goes_across_prime_meridian) {
sub_array_of_fri <- 1:n
#gap_width <- sub_array_of_indices[[1]] - sub_array_of_indices[[2]] - 1
#sub_array_of_fri <- c((1:(sub_array_of_indices[[2]] + min(gap_width, beta))),
# (sub_array_of_indices[[1]] - min(gap_width, beta)):n)
} else {
if (is.list(sub_array_of_indices)) {
sub_array_of_indices <- sub_array_of_indices[[1]]:sub_array_of_indices[[2]]
}
first_index <- min(unlist(sub_array_of_indices))
last_index <- max(unlist(sub_array_of_indices))
start_padding <- min(beta, first_index - 1)
end_padding <- min(beta, n - last_index)
sub_array_of_fri <- (first_index - start_padding):(last_index + end_padding)
}
subset_vars_to_transform <- vars_to_transform
if (!is.null(var_ordered)) {
subset_vars_to_transform[[var_with_selectors_name]] <- Subset(var_ordered, inner_dim, sub_array_of_fri)
} else {
subset_vars_to_transform[[var_with_selectors_name]] <- Subset(sub_array_of_values, inner_dim, sub_array_of_fri)
}
transformed_subset_var <- do.call(transform, c(list(data_array = NULL,
variables = subset_vars_to_transform,
file_selectors = selectors_of_first_files_with_data[[i]]),
transform_params))$variables[[var_with_selectors_name]]
# Sorting the transformed variable and working out the indices again after transform.
if (!is.null(dim_reorder_params[[inner_dim]])) {
transformed_subset_var_reorder <- dim_reorder_params[[inner_dim]](transformed_subset_var)
transformed_subset_var <- transformed_subset_var_reorder$x
transformed_subset_var_unorder <- sort(transformed_subset_var_reorder$ix, index.return = TRUE)$ix
} else {
transformed_subset_var_unorder <- 1:length(transformed_subset_var)
}
sub_array_of_sri <- selector_checker(sub_array_of_selectors, transformed_subset_var,
tolerance = if (aiat) {
tolerance_params[[inner_dim]]
} else {
NULL
})
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
if (goes_across_prime_meridian) {
sub_array_of_sri <- c(1:sub_array_of_sri[[2]], sub_array_of_sri[[1]]:length(transformed_subset_var))
#sub_array_of_sri <- c(sub_array_of_sri[[1]]:length(transformed_subset_var), 1:sub_array_of_sri[[2]])
} else if (is.list(sub_array_of_sri)) {
sub_array_of_sri <- sub_array_of_sri[[1]]:sub_array_of_sri[[2]]
}
ordered_sri <- sub_array_of_sri
sub_array_of_sri <- transformed_subset_var_unorder[sub_array_of_sri]
# In this case, the tvi are not defined and the 'transformed_subset_var'
# will be taken instead of the var transformed before in the code.
if (debug) {
if (inner_dim %in% dims_to_check) {
print("-> FIRST INDEX:")
print(first_index)
print("-> LAST INDEX:")
print(last_index)
print("-> STRUCTURE OF FIRST ROUND INDICES:")
print(str(sub_array_of_fri))
print("-> STRUCTURE OF SECOND ROUND INDICES:")
print(str(sub_array_of_sri))
print("-> STRUCTURE OF TRANSFORMED VARIABLE INDICES:")
print(str(tvi))
}
}
### # If the selectors are expressed after transformation
### } else {
###if (debug) {
###if (inner_dim %in% dims_to_check) {
###print("-> SELECTORS REQUESTED AFTER TRANSFORM.")
###}
###}
### if (goes_across_prime_meridian) {
### sub_array_of_indices <- c(sub_array_of_indices[[1]]:m,
### 1:sub_array_of_indices[[2]])
### }
### first_index <- min(unlist(sub_array_of_indices))
### last_index <- max(unlist(sub_array_of_indices))
### first_index_before_transform <- max(transform_indices(first_index, m, n) - beta, 1)
### last_index_before_transform <- min(transform_indices(last_index, m, n) + beta, n)
### sub_array_of_fri <- first_index_before_transform:last_index_before_transform
### n_of_extra_cells <- round(beta / n * m)
### if (is.list(sub_array_of_indices) && (length(sub_array_of_indices) > 1)) {
### sub_array_of_sri <- 1:(last_index - first_index + 1)
### if (is.null(tvi)) {
### tvi <- sub_array_of_sri + first_index - 1
### }
### } else {
### sub_array_of_sri <- sub_array_of_indices - first_index + 1
### if (is.null(tvi)) {
### tvi <- sub_array_of_indices
### }
### }
### sub_array_of_sri <- sub_array_of_sri + n_of_extra_cells
sri <- do.call('[[<-', c(list(x = sri), as.list(selector_store_position),
list(value = sub_array_of_sri)))
} else {
if (goes_across_prime_meridian) {
#sub_array_of_fri <- 1:n
sub_array_of_fri <- c(1:sub_array_of_indices[[2]], sub_array_of_indices[[1]]:n)
} else if (is.list(sub_array_of_indices)) {
sub_array_of_fri <- sub_array_of_indices[[1]]:sub_array_of_indices[[2]]
} else {
sub_array_of_fri <- sub_array_of_indices
}
}
if (!is.null(var_unorder_indices)) {
if (is.null(ordered_fri)) {
ordered_fri <- sub_array_of_fri
}
sub_array_of_fri <- var_unorder_indices[sub_array_of_fri]
}
fri <- do.call('[[<-', c(list(x = fri), as.list(selector_store_position),
list(value = sub_array_of_fri)))
if (!is.null(file_dim)) {
taken_chunks[selector_store_position[[file_dim]]] <- TRUE
} else {
taken_chunks <- TRUE
}
}
} else {
if (debug) {
if (inner_dim %in% dims_to_check) {
print("-> THE INNER DIMENSION GOES ACROSS A FILE DIMENSION.")
}
}
if (inner_dim %in% names(dim(sub_array_of_selectors))) {
if (is.null(var_with_selectors_name)) {
if (any(na.omit(unlist(sub_array_of_selectors)) < 1) ||
any(na.omit(unlist(sub_array_of_selectors)) > data_dims[inner_dim] * chunk_amount)) {
stop("Provided indices out of range for dimension '", inner_dim, "' ",
"for dataset '", dat[[i]][['name']], "' (accepted range: 1 to ",
data_dims[inner_dim] * chunk_amount, ").")
}
} else {
if (inner_dim %in% names(dim(sub_array_of_values))) {
inner_dim_pos_in_sub_array <- which(names(dim(sub_array_of_values)) == inner_dim)
if (inner_dim_pos_in_sub_array != 1) {
new_sub_array_order <- (1:length(dim(sub_array_of_values)))[-inner_dim_pos_in_sub_array]
new_sub_array_order <- c(inner_dim_pos_in_sub_array, new_sub_array_order)
sub_array_of_values <- .aperm2(sub_array_of_values, new_sub_array_order)
}
}
}
inner_dim_pos_in_sub_array <- which(names(dim(sub_array_of_selectors)) == inner_dim)
if (inner_dim_pos_in_sub_array != 1) {
new_sub_array_order <- (1:length(dim(sub_array_of_selectors)))[-inner_dim_pos_in_sub_array]
new_sub_array_order <- c(inner_dim_pos_in_sub_array, new_sub_array_order)
sub_array_of_selectors <- .aperm2(sub_array_of_selectors, new_sub_array_order)
}
sub_array_of_indices <- selector_checker(sub_array_of_selectors, sub_array_of_values,
tolerance = tolerance_params[[inner_dim]])
sub_array_of_indices <- sub_array_of_indices[chunk_indices(length(sub_array_of_indices),
chunks[[inner_dim]]['chunk'],
chunks[[inner_dim]]['n_chunks'],
inner_dim)]
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
sub_array_is_list <- FALSE
if (is.list(sub_array_of_indices)) {
sub_array_is_list <- TRUE
sub_array_of_indices <- unlist(sub_array_of_indices)
}
if (is.null(var_with_selectors_name)) {
indices_chunk <- floor((sub_array_of_indices - 1) / data_dims[inner_dim]) + 1
transformed_indices <- ((sub_array_of_indices - 1) %% data_dims[inner_dim]) + 1
} else {
indices_chunk <- floor((sub_array_of_indices - 1) / var_full_dims[inner_dim]) + 1
transformed_indices <- ((sub_array_of_indices - 1) %% var_full_dims[inner_dim]) + 1
}
if (sub_array_is_list) {
sub_array_of_indices <- as.list(sub_array_of_indices)
}
if (debug) {
if (inner_dim %in% dims_to_check) {
print("-> GOING TO ITERATE ALONG CHUNKS.")
}
}
for (chunk in 1:chunk_amount) {
if (!is.null(names(selector_store_position))) {
selector_store_position[file_dim] <- chunk
} else {
selector_store_position <- chunk
}
chunk_selectors <- transformed_indices[which(indices_chunk == chunk)]
sub_array_of_indices <- chunk_selectors
if (with_transform) {
# If the provided selectors are expressed in the world
# before transformation
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
first_index <- min(unlist(sub_array_of_indices))
last_index <- max(unlist(sub_array_of_indices))
sub_array_of_fri <- max(c(first_index - beta, 1)):min(c(last_index + beta, n))
sub_array_of_sri <- transform_indices(unlist(sub_array_of_indices) - first_index + 1, n, m)
if (is.list(sub_array_of_indices)) {
if (length(sub_array_of_sri) > 1) {
sub_array_of_sri <- sub_array_of_sri[[1]]:sub_array_of_sri[[2]]
}
}
##TODO: TRANSFORM SUBSET VARIABLE AS ABOVE, TO COMPUTE SRI
# If the selectors are expressed after transformation
} else {
first_index <- min(unlist(sub_array_of_indices))
last_index <- max(unlist(sub_array_of_indices))
first_index_before_transform <- max(transform_indices(first_index, m, n) - beta, 1)
last_index_before_transform <- min(transform_indices(last_index, m, n) + beta, n)
sub_array_of_fri <- first_index_before_transform:last_index_before_transform
if (is.list(sub_array_of_indices) && (length(sub_array_of_indices) > 1)) {
sub_array_of_sri <- 1:(last_index - first_index + 1) +
round(beta / n * m)
} else {
sub_array_of_sri <- sub_array_of_indices - first_index + 1 +
round(beta / n * m)
}
##TODO: FILL IN TVI
}
sri <- do.call('[[<-', c(list(x = sri), as.list(selector_store_position),
list(value = sub_array_of_sri)))
if (length(sub_array_of_sri) > 0) {
taken_chunks[chunk] <- TRUE
}
} else {
sub_array_of_fri <- sub_array_of_indices
if (length(sub_array_of_fri) > 0) {
taken_chunks[chunk] <- TRUE
}
}
if (!is.null(var_unorder_indices)) {
ordered_fri <- sub_array_of_fri
sub_array_of_fri <- var_unorder_indices[sub_array_of_fri]
}
fri <- do.call('[[<-', c(list(x = fri), as.list(selector_store_position),
list(value = sub_array_of_fri)))
}
if (debug) {
if (inner_dim %in% dims_to_check) {
print("-> FINISHED ITERATING ALONG CHUNKS")
}
}
} else {
stop("Provided array of indices for dimension '", inner_dim, "', ",
"which goes across the file dimension '", file_dim, "', but ",
"the provided array does not have the dimension '", inner_dim,
"', which is mandatory.")
}
}
}
}
if (debug) {
if (inner_dim %in% dims_to_check) {
print("-> PROCEEDING TO CROP VARIABLES")
}
}
#if ((length(selector_array) == 1) && (selector_array %in% c('all', 'first', 'last'))) {
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
#if (!is.null(var_with_selectors_name) || (is.null(var_with_selectors_name) && is.character(selector_array) &&
# (length(selector_array) == 1) && (selector_array %in% c('all', 'first', 'last')))) {
empty_chunks <- which(!taken_chunks)
if (length(empty_chunks) >= length(taken_chunks)) {
stop("Selectors do not match any of the possible values for the dimension '", inner_dim, "'.")
}
if (length(empty_chunks) > 0) {
# # Get the first group of chunks to remove, and remove them.
# # E.g., from c(1, 2, 4, 5, 6, 8, 9) remove only 1 and 2
# dist <- abs(rev(empty_chunks) - c(rev(empty_chunks)[1] - 1, head(rev(empty_chunks), length(rev(empty_chunks)) - 1)))
# if (all(dist == 1)) {
# start_chunks_to_remove <- NULL
# } else {
# first_chunk_to_remove <- tail(which(dist > 1), 1)
# start_chunks_to_remove <- rev(rev(empty_chunks)[first_chunk_to_remove:length(empty_chunks)])
# }
# # Get the last group of chunks to remove, and remove them.
# # E.g., from c(1, 2, 4, 5, 6, 8, 9) remove only 8 and 9
# dist <- abs(empty_chunks - c(empty_chunks[1] - 1, head(empty_chunks, length(empty_chunks) - 1)))
# if (all(dist == 1)) {
# first_chunk_to_remove <- 1
# } else {
# first_chunk_to_remove <- tail(which(dist > 1), 1)
# }
# end_chunks_to_remove <- empty_chunks[first_chunk_to_remove:length(empty_chunks)]
# chunks_to_keep <- which(!((1:length(taken_chunks)) %in% c(start_chunks_to_remove, end_chunks_to_remove)))
chunks_to_keep <- which(taken_chunks)
dims_to_crop[[file_dim]] <- c(dims_to_crop[[file_dim]], list(chunks_to_keep))
# found_indices <- Subset(found_indices, file_dim, chunks_to_keep)
# # Crop dataset variables file dims.
# for (picked_var in names(picked_vars[[i]])) {
# if (file_dim %in% names(dim(picked_vars[[i]][[picked_var]]))) {
# picked_vars[[i]][[picked_var]] <- Subset(picked_vars[[i]][[picked_var]], file_dim, chunks_to_keep)
# }
# }
}
#}
dat[[i]][['selectors']][[inner_dim]] <- list(fri = fri, sri = sri)
# Crop dataset variables inner dims.
# Crop common variables inner dims.
types_of_var_to_crop <- 'picked'
if (with_transform) {
types_of_var_to_crop <- c(types_of_var_to_crop, 'transformed')
}
if (!is.null(dim_reorder_params[[inner_dim]])) {
types_of_var_to_crop <- c(types_of_var_to_crop, 'reordered')
}
for (type_of_var_to_crop in types_of_var_to_crop) {
if (type_of_var_to_crop == 'transformed') {
if (is.null(tvi)) {
if (!is.null(dim_reorder_params[[inner_dim]])) {
crop_indices <- unique(unlist(ordered_sri))
} else {
crop_indices <- unique(unlist(sri))
}
} else {
crop_indices <- unique(unlist(tvi))
}
vars_to_crop <- transformed_vars[[i]]
common_vars_to_crop <- transformed_common_vars
} else if (type_of_var_to_crop == 'reordered') {
crop_indices <- unique(unlist(ordered_fri))
vars_to_crop <- picked_vars_ordered[[i]]
common_vars_to_crop <- picked_common_vars_ordered
} else {
crop_indices <- unique(unlist(fri))
vars_to_crop <- picked_vars[[i]]
common_vars_to_crop <- picked_common_vars
}
for (var_to_crop in names(vars_to_crop)) {
if (inner_dim %in% names(dim(vars_to_crop[[var_to_crop]]))) {
if (!is.null(crop_indices)) {
if (type_of_var_to_crop == 'transformed') {
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
vars_to_crop[[var_to_crop]] <- Subset(transformed_subset_var, inner_dim, crop_indices)
} else {
vars_to_crop[[var_to_crop]] <- Subset(vars_to_crop[[var_to_crop]], inner_dim, crop_indices)
}
} else {
vars_to_crop[[var_to_crop]] <- Subset(vars_to_crop[[var_to_crop]], inner_dim, crop_indices)
}
}
}
}
if (i == length(dat)) {
for (common_var_to_crop in names(common_vars_to_crop)) {
if (inner_dim %in% names(dim(common_vars_to_crop[[common_var_to_crop]]))) {
common_vars_to_crop[[common_var_to_crop]] <- Subset(common_vars_to_crop[[common_var_to_crop]], inner_dim, crop_indices)
}
}
}
if (type_of_var_to_crop == 'transformed') {
if (!is.null(vars_to_crop)) {
transformed_vars[[i]] <- vars_to_crop
}
if (i == length(dat)) {
transformed_common_vars <- common_vars_to_crop
}
} else if (type_of_var_to_crop == 'reordered') {
if (!is.null(vars_to_crop)) {
picked_vars_ordered[[i]] <- vars_to_crop
}
if (i == length(dat)) {
picked_common_vars_ordered <- common_vars_to_crop
}
} else {
if (!is.null(vars_to_crop)) {
picked_vars[[i]] <- vars_to_crop
}
if (i == length(dat)) {
picked_common_vars <- common_vars_to_crop
}
}
}
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
}
# After the selectors have been picked (using the original variables),
# the variables are transformed. At that point, the original selectors
# for the transformed variables are also kept in the variable original_selectors.
#print("L")
}
}
}
# if (!is.null(transformed_common_vars)) {
# picked_common_vars[names(transformed_common_vars)] <- transformed_common_vars
# }
# Remove the trailing chunks, if any.
for (file_dim in names(dims_to_crop)) {
# indices_to_keep <- min(sapply(dims_to_crop[[file_dim]], min)):max(sapply(dims_to_crop[[file_dim]], max))
## TODO: Merge indices in dims_to_crop with some advanced mechanism?
indices_to_keep <- unique(unlist(dims_to_crop[[file_dim]]))
array_of_files_to_load <- Subset(array_of_files_to_load, file_dim, indices_to_keep)
array_of_not_found_files <- Subset(array_of_not_found_files, file_dim, indices_to_keep)
for (i in 1:length(dat)) {
# Crop selectors
for (selector_dim in names(dat[[i]][['selectors']])) {
if (selector_dim == file_dim) {
for (j in 1:length(dat[[i]][['selectors']][[selector_dim]][['fri']])) {
dat[[i]][['selectors']][[selector_dim]][['fri']][[j]] <- dat[[i]][['selectors']][[selector_dim]][['fri']][[j]][indices_to_keep]
}
for (j in 1:length(dat[[i]][['selectors']][[selector_dim]][['sri']])) {
dat[[i]][['selectors']][[selector_dim]][['sri']][[j]] <- dat[[i]][['selectors']][[selector_dim]][['sri']][[j]][indices_to_keep]
}
}
if (file_dim %in% names(dim(dat[[i]][['selectors']][[selector_dim]][['fri']]))) {
dat[[i]][['selectors']][[selector_dim]][['fri']] <- Subset(dat[[i]][['selectors']][[selector_dim]][['fri']], file_dim, indices_to_keep)
dat[[i]][['selectors']][[selector_dim]][['sri']] <- Subset(dat[[i]][['selectors']][[selector_dim]][['sri']], file_dim, indices_to_keep)
}
}
# Crop dataset variables file dims.
for (picked_var in names(picked_vars[[i]])) {
if (file_dim %in% names(dim(picked_vars[[i]][[picked_var]]))) {
picked_vars[[i]][[picked_var]] <- Subset(picked_vars[[i]][[picked_var]], file_dim, indices_to_keep)
}
}
for (transformed_var in names(transformed_vars[[i]])) {
if (file_dim %in% names(dim(transformed_vars[[i]][[transformed_var]]))) {
transformed_vars[[i]][[transformed_var]] <- Subset(transformed_vars[[i]][[transformed_var]], file_dim, indices_to_keep)
}
}
}
# Crop common variables file dims.
for (picked_common_var in names(picked_common_vars)) {
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
if (file_dim %in% names(dim(picked_common_vars[[picked_common_var]]))) {
picked_common_vars[[picked_common_var]] <- Subset(picked_common_vars[[picked_common_var]], file_dim, indices_to_keep)
}
}
for (transformed_common_var in names(transformed_common_vars)) {
if (file_dim %in% names(dim(transformed_common_vars[[transformed_common_var]]))) {
transformed_common_vars[[transformed_common_var]] <- Subset(transformed_common_vars[[transformed_common_var]], file_dim, indices_to_keep)
}
}
}
# Calculate the size of the final array.
total_inner_dims <- NULL
for (i in 1:length(dat)) {
if (dataset_has_files[i]) {
inner_dims <- expected_inner_dims[[i]]
inner_dims <- sapply(inner_dims,
function(x) {
if (!all(sapply(dat[[i]][['selectors']][[x]][['sri']], is.null))) {
max(sapply(dat[[i]][['selectors']][[x]][['sri']], length))
} else {
if (length(var_params[[x]]) > 0) {
if (var_params[[x]] %in% names(transformed_vars[[i]])) {
length(transformed_vars[[i]][[var_params[[x]]]])
} else if (var_params[[x]] %in% names(transformed_common_vars)) {
length(transformed_common_vars[[var_params[[x]]]])
} else {
max(sapply(dat[[i]][['selectors']][[x]][['fri']], length))
}
} else {
max(sapply(dat[[i]][['selectors']][[x]][['fri']], length))
}
}
})
names(inner_dims) <- expected_inner_dims[[i]]
if (is.null(total_inner_dims)) {
total_inner_dims <- inner_dims
} else {
new_dims <- .MergeArrayDims(total_inner_dims, inner_dims)
total_inner_dims <- pmax(new_dims[[1]], new_dims[[2]])
}
}
}
new_dims <- .MergeArrayDims(dim(array_of_files_to_load), total_inner_dims)
final_dims <- pmax(new_dims[[1]], new_dims[[2]])[dim_names]
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
if (!silent) {
.message("Detected dimension sizes:")
longest_dim_len <- max(sapply(names(final_dims), nchar))
longest_size_len <- max(sapply(paste0(final_dims, ''), nchar))
sapply(names(final_dims),
function(x) {
message(paste0("* ", paste(rep(' ', longest_dim_len - nchar(x)), collapse = ''),
x, ": ", paste(rep(' ', longest_size_len - nchar(paste0(final_dims[x], ''))), collapse = ''),
final_dims[x]))
})
bytes <- prod(c(final_dims, 8))
dim_sizes <- paste(final_dims, collapse = ' x ')
if (retrieve) {
.message(paste("Total size of requested data:"))
} else {
.message(paste("Total size of involved data:"))
}
.message(paste(dim_sizes, " x 8 bytes =",
format(structure(bytes, class = "object_size"), units = "auto")),
indent = 2)
}
# The following several lines will only be run if retrieve = TRUE
if (retrieve) {
########## CREATING THE SHARED MATRIX AND DISPATCHING WORK PIECES ###########
# TODO: try performance of storing all in cols instead of rows
# Create the shared memory array, and a pointer to it, to be sent
# to the work pieces.
data_array <- big.matrix(nrow = prod(final_dims), ncol = 1)
shared_matrix_pointer <- describe(data_array)
if (is.null(num_procs)) {
# Creating a shared tmp folder to store metadata from each chunk
array_of_metadata_flags <- array(FALSE, dim = dim(array_of_files_to_load))
if (!is.null(metadata_dims)) {
metadata_indices_to_load <- as.list(rep(1, length(dim(array_of_files_to_load))))
names(metadata_indices_to_load) <- names(dim(array_of_files_to_load))
metadata_indices_to_load[metadata_dims] <- as.list(rep(TRUE, length(metadata_dims)))
array_of_metadata_flags <- do.call('[<-', c(list(array_of_metadata_flags), metadata_indices_to_load,
list(value = rep(TRUE, prod(dim(array_of_files_to_load)[metadata_dims])))))
}
metadata_file_counter <- 0
metadata_folder <- tempfile('metadata')
dir.create(metadata_folder)
# Build the work pieces, each with:
# - file path
# - total size (dims) of store array
# - start position in store array
# - file selectors (to provide extra info. useful e.g. to select variable)
# - indices to take from file
work_pieces <- list()
for (i in 1:length(dat)) {
if (dataset_has_files[i]) {
selectors <- dat[[i]][['selectors']]
file_dims <- found_file_dims[[i]]
inner_dims <- expected_inner_dims[[i]]
sub_array_dims <- final_dims[file_dims]
sub_array_dims[found_pattern_dim] <- 1
sub_array_of_files_to_load <- array(1:prod(sub_array_dims),
dim = sub_array_dims)
names(dim(sub_array_of_files_to_load)) <- names(sub_array_dims)
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
# Detect which of the dimensions of the dataset go across files.
file_dim_across_files <- lapply(inner_dims,
function(x) {
dim_across <- sapply(inner_dims_across_files, function(y) x %in% y)
if (any(dim_across)) {
names(inner_dims_across_files)[which(dim_across)[1]]
} else {
NULL
}
})
names(file_dim_across_files) <- inner_dims
j <- 1
while (j <= prod(sub_array_dims)) {
# Work out file path.
file_to_load_sub_indices <- which(sub_array_of_files_to_load == j, arr.ind = TRUE)[1, ]
names(file_to_load_sub_indices) <- names(sub_array_dims)
file_to_load_sub_indices[found_pattern_dim] <- i
big_dims <- rep(1, length(dim(array_of_files_to_load)))
names(big_dims) <- names(dim(array_of_files_to_load))
file_to_load_indices <- .MergeArrayDims(file_to_load_sub_indices, big_dims)[[1]]
file_to_load <- do.call('[[', c(list(array_of_files_to_load),
as.list(file_to_load_indices)))
not_found_file <- do.call('[[', c(list(array_of_not_found_files),
as.list(file_to_load_indices)))
load_file_metadata <- do.call('[', c(list(array_of_metadata_flags),
as.list(file_to_load_indices)))
if (load_file_metadata) {
metadata_file_counter <- metadata_file_counter + 1
}
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
if (!is.na(file_to_load) && !not_found_file) {
# Work out indices to take
first_round_indices <- lapply(inner_dims,
function (x) {
if (is.null(file_dim_across_files[[x]])) {
selectors[[x]][['fri']][[1]]
} else {
var_file_dims <- dim(selectors[inner_dims][[x]][['fri']])
which_indices <- file_to_load_sub_indices[which(names(sub_array_dims) %in% names(var_file_dims))]
do.call('[[', c(list(selectors[[x]][['fri']]), as.list(which_indices)))
}
})
names(first_round_indices) <- inner_dims
second_round_indices <- lapply(inner_dims,
function (x) {
if (is.null(file_dim_across_files[[x]])) {
selectors[[x]][['sri']][[1]]
} else {
var_file_dims <- dim(selectors[inner_dims][[x]][['sri']])
which_indices <- file_to_load_sub_indices[which(names(sub_array_dims) %in% names(var_file_dims))]
do.call('[[', c(list(selectors[[x]][['sri']]), as.list(which_indices)))
}
})
if (debug) {
print("-> BUILDING A WORK PIECE")
#print(str(selectors))
}
names(second_round_indices) <- inner_dims
if (!any(sapply(first_round_indices, length) == 0)) {
work_piece <- list()
work_piece[['first_round_indices']] <- first_round_indices
work_piece[['second_round_indices']] <- second_round_indices
work_piece[['file_indices_in_array_of_files']] <- file_to_load_indices
work_piece[['file_path']] <- file_to_load
work_piece[['store_dims']] <- final_dims
# Work out store position
store_position <- final_dims
store_position[names(file_to_load_indices)] <- file_to_load_indices
store_position[inner_dims] <- rep(1, length(inner_dims))
work_piece[['store_position']] <- store_position
# Work out file selectors
file_selectors <- sapply(file_dims,
function (x) {
vector_to_pick <- 1
if (x %in% names(depending_file_dims)) {
vector_to_pick <- file_to_load_indices[depending_file_dims[[x]]]
}
selectors[file_dims][[x]][[vector_to_pick]][file_to_load_indices[x]]
})
names(file_selectors) <- file_dims
work_piece[['file_selectors']] <- file_selectors
# Send variables for transformation
if (!is.null(transform) && (length(transform_vars) > 0)) {
vars_to_transform <- NULL
picked_vars_to_transform <- which(names(picked_vars[[i]]) %in% transform_vars)
if (length(picked_vars_to_transform) > 0) {
picked_vars_to_transform <- names(picked_vars[[i]])[picked_vars_to_transform]
vars_to_transform <- c(vars_to_transform, picked_vars[[i]][picked_vars_to_transform])
if (any(picked_vars_to_transform %in% names(picked_vars_ordered[[i]]))) {
picked_vars_ordered_to_transform <- picked_vars_to_transform[which(picked_vars_to_transform %in% names(picked_vars_ordered[[i]]))]
vars_to_transform[picked_vars_ordered_to_transform] <- picked_vars_ordered[[i]][picked_vars_ordered_to_transform]
}
}
picked_common_vars_to_transform <- which(names(picked_common_vars) %in% transform_vars)
if (length(picked_common_vars_to_transform) > 0) {
picked_common_vars_to_transform <- names(picked_common_vars)[picked_common_vars_to_transform]
vars_to_transform <- c(vars_to_transform, picked_common_vars[picked_common_vars_to_transform])
if (any(picked_common_vars_to_transform %in% names(picked_common_vars_ordered))) {
picked_common_vars_ordered_to_transform <- picked_common_vars_to_transform[which(picked_common_vars_to_transform %in% names(picked_common_vars_ordered))]
vars_to_transform[picked_common_vars_ordered_to_transform] <- picked_common_vars_ordered[picked_common_vars_ordered_to_transform]
}
}
work_piece[['vars_to_transform']] <- vars_to_transform
}
# Send flag to load metadata
if (load_file_metadata) {
work_piece[['save_metadata_in']] <- paste0(metadata_folder, '/', metadata_file_counter)
}
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
work_pieces <- c(work_pieces, list(work_piece))
}
}
j <- j + 1
}
}
}
#print("N")
if (debug) {
print("-> WORK PIECES BUILT")
}
# Calculate the progress %s that will be displayed and assign them to
# the appropriate work pieces.
if (length(work_pieces) / num_procs >= 2 && !silent) {
if (length(work_pieces) / num_procs < 10) {
amount <- 100 / ceiling(length(work_pieces) / num_procs)
reps <- ceiling(length(work_pieces) / num_procs)
} else {
amount <- 10
reps <- 10
}
progress_steps <- rep(amount, reps)
if (length(work_pieces) < (reps + 1)) {
selected_pieces <- length(work_pieces)
progress_steps <- c(sum(head(progress_steps, reps)),
tail(progress_steps, reps))
} else {
selected_pieces <- round(seq(1, length(work_pieces),
length.out = reps + 1))[-1]
}
progress_steps <- paste0(' + ', round(progress_steps, 2), '%')
progress_message <- 'Progress: 0%'
} else {
progress_message <- ''
selected_pieces <- NULL
}
piece_counter <- 1
step_counter <- 1
work_pieces <- lapply(work_pieces,
function (x) {
if (piece_counter %in% selected_pieces) {
wp <- c(x, list(progress_amount = progress_steps[step_counter]))
step_counter <<- step_counter + 1
} else {
wp <- x
}
piece_counter <<- piece_counter + 1
wp
})
if (!silent) {
.message("If the size of the requested data is close to or above the free shared RAM memory, R may crash.")
.message("If the size of the requested data is close to or above the half of the free RAM memory, R may crash.")
.message(paste0("Will now proceed to read and process ", length(work_pieces), " data files:"))
if (length(work_pieces) < 30) {
lapply(work_pieces, function (x) .message(x[['file_path']], indent = 2))
} else {
.message("The list of files is long. You can check it after Start() finishes in the output '$Files'.", indent = 2, exdent = 5)
}
}
# Build the cluster of processes that will do the work and dispatch work pieces.
# The function .LoadDataFile is applied to each work piece. This function will
# open the data file, regrid if needed, subset, apply the mask,
# compute and apply the weights if needed,
# disable extreme values and store in the shared memory matrix.
#print("O")
if (!silent) {
.message("Loading... This may take several minutes...")
if (progress_message != '') {
.message(progress_message, appendLF = FALSE)
}
}
if (num_procs == 1) {
found_files <- lapply(work_pieces, .LoadDataFile,
shared_matrix_pointer = shared_matrix_pointer,
file_data_reader = file_data_reader,
transform = transform,
transform_params = transform_params,
silent = silent, debug = debug)
} else {
cluster <- makeCluster(num_procs, outfile = "")
# Send the heavy work to the workers
work_errors <- try({
found_files <- clusterApplyLB(cluster, work_pieces, .LoadDataFile,
shared_matrix_pointer = shared_matrix_pointer,
file_data_reader = file_data_reader,
transform = transform,
transform_params = transform_params,
silent = silent, debug = debug)
})
stopCluster(cluster)
}
if (!silent) {
if (progress_message != '') {
.message("\n", tag = '')
}
}
#print("P")
data_array <- array(bigmemory::as.matrix(data_array), dim = final_dims)
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)
}
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
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
}
# 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.")
}
attributes(data_array) <- c(attributes(data_array),
list(Variables = c(list(common = picked_common_vars), picked_vars),
Files = array_of_files_to_load,
NotFoundFiles = array_of_not_found_files,
FileSelectors = file_selectors)
)
attr(data_array, 'class') <- c('startR_cube', 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]]) == 'name') {
start_call[[i]] <- eval(start_call[[i]])
}
}
start_call[['retrieve']] <- TRUE
attributes(start_call) <- c(attributes(start_call),
list(Dimensions = final_dims,
Variables = c(list(common = picked_common_vars), picked_vars),
ExpectedFiles = array_of_files_to_load,
FileSelectors = file_selectors)
)
attr(start_call, 'class') <- c('startR_header', 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,
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
transform, transform_params,
silent = FALSE, debug = FALSE) {
# suppressPackageStartupMessages({library(bigmemory)})
### TODO: Specify dependencies as parameter
# suppressPackageStartupMessages({library(ncdf4)})
# Auxiliar function to convert array indices to lineal indices.
# This function expects a numeric vector of single indices or a list of
# numeric vectors with multiple indices for each dimension, and a
# numeric vector of dimension sizes.
.arrayIndices2VectorIndices <- function(indices, dims) {
# Check indices
if (!is.list(indices)) {
if (is.numeric(indices)) {
indices <- as.list(indices)
} else {
stop("Parameter 'indices' must be a numeric vector or a list of ",
"numeric vectors.")
}
}
sapply(indices,
function(x) {
if (!is.numeric(x)) {
stop("Parameter 'indices' must be a numeric vector or a list ",
"of numeric vectors.")
} else if (any(is.na(x) | is.nan(x) | is.infinite(x))) {
stop("Parameter 'indices' must not contain NA/NaN/Inf values.")
} else if (length(x) < 1) {
stop("Parameter 'indices' must contain at least one index ",
"for each dimension.")
}
})
# Check dims
if (!is.numeric(dims)) {
stop("Parameter 'dims' must be a numeric vector.")
} else if (any(is.na(dims) | is.nan(dims) | is.infinite(dims))) {
stop("Parameter 'dims' must not contain NA/NaN/Inf values.")
} else if (any(sapply(dims, length) != 1)) {
stop("Parameter 'dims' must contain a single numeric element for ",
"each dimension.")
} else if (length(indices) != length(dims)) {
stop("There must be as many dimension selectors in 'indices' as ",
"dimensions in 'dims'.")
}
find_indices <- function(indices, dims) {
if (max(indices[[1]]) > dims[1] || min(indices[[1]]) < 1) {
stop("Provided indices out of range.")
}
if (length(dims) == 1) {
indices[[1]]
} else {
found_indices <- find_indices(indices[-1], dims[-1])
new_found_indices <- c()
for (i in 1:length(indices[[1]])) {
new_found_indices <- c(new_found_indices, (indices[[1]][i] - 1) * prod(dims[-1]) + found_indices)
}
new_found_indices
}
}
indices <- rev(indices)
dims <- rev(dims)
find_indices(indices, dims)
}
#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)
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
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']]),
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
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')
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
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]])
}
}
})
if (debug) {
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))
}
}
matrix_indices <- .arrayIndices2VectorIndices(store_indices, work_piece[['store_dims']])
data_array <- bigmemory::attach.big.matrix(shared_matrix_pointer)
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)
}