File: clusterFunctions.R

package info (click to toggle)
r-cran-batchtools 0.9.15%2Bdfsg-1
  • links: PTS, VCS
  • area: main
  • in suites: bullseye
  • size: 1,416 kB
  • sloc: ansic: 172; sh: 156; makefile: 2
file content (328 lines) | stat: -rw-r--r-- 14,517 bytes parent folder | download | duplicates (4)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
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
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
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
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
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
#' @title ClusterFunctions Constructor
#'
#' @description
#' This is the constructor used to create \emph{custom} cluster functions.
#' Note that some standard implementations for TORQUE, Slurm, LSF, SGE, etc. ship
#' with the package.
#'
#' @param name [\code{character(1)}]\cr
#'   Name of cluster functions.
#' @param submitJob [\code{function(reg, jc, ...)}]\cr
#'   Function to submit new jobs. Must return a \code{\link{SubmitJobResult}} object.
#'   The arguments are \code{reg} (\code{\link{Registry}}) and \code{jobs} (\code{\link{JobCollection}}).
#' @param killJob [\code{function(reg, batch.id)}]\cr
#'   Function to kill a job on the batch system. Make sure that you definitely kill the job! Return
#'   value is currently ignored. Must have the arguments \code{reg} (\code{\link{Registry}}) and
#'   \code{batch.id} (\code{character(1)} as returned by \code{submitJob}).
#'   Note that there is a helper function \code{\link{cfKillJob}} to repeatedly try to kill jobs.
#'   Set \code{killJob} to \code{NULL} if killing jobs cannot be supported.
#' @param listJobsQueued [\code{function(reg)}]\cr
#'   List all queued jobs on the batch system for the current user.
#'   Must return an character vector of batch ids, same format as they
#'   are returned by \code{submitJob}.
#'   Set \code{listJobsQueued} to \code{NULL} if listing of queued jobs is not supported.
#' @param listJobsRunning [\code{function(reg)}]\cr
#'   List all running jobs on the batch system for the current user.
#'   Must return an character vector of batch ids, same format as they
#'   are returned by \code{submitJob}. It does not matter if you return a few job ids too many (e.g.
#'   all for the current user instead of all for the current registry), but you have to include all
#'   relevant ones. Must have the argument are \code{reg} (\code{\link{Registry}}).
#'   Set \code{listJobsRunning} to \code{NULL} if listing of running jobs is not supported.
#' @param array.var [\code{character(1)}]\cr
#'   Name of the environment variable set by the scheduler to identify IDs of job arrays.
#'   Default is \code{NA} for no array support.
#' @param store.job.collection [\code{logical(1)}]\cr
#'   Flag to indicate that the cluster function implementation of \code{submitJob} can not directly handle \code{\link{JobCollection}} objects.
#'   If set to \code{FALSE}, the \code{\link{JobCollection}} is serialized to the file system before submitting the job.
#' @param store.job.files [\code{logical(1)}]\cr
#'   Flag to indicate that job files need to be stored in the file directory.
#'   If set to \code{FALSE} (default), the job file is created in a temporary directory, otherwise (or if the debug mode is enabled) in
#'   the subdirectory \code{jobs} of the \code{file.dir}.
#' @param scheduler.latency [\code{numeric(1)}]\cr
#'   Time to sleep after important interactions with the scheduler to ensure a sane state.
#'   Currently only triggered after calling \code{\link{submitJobs}}.
#' @param fs.latency [\code{numeric(1)}]\cr
#'   Expected maximum latency of the file system, in seconds.
#'   Set to a positive number for network file systems like NFS which enables more robust (but also more expensive) mechanisms to
#'   access files and directories.
#'   Usually safe to set to \code{0} to disable the heuristic, e.g. if you are working on a local file system.
#' @param hooks [\code{list}]\cr
#'   Named list of functions which will we called on certain events like \dQuote{pre.submit} or \dQuote{post.sync}.
#'   See \link{Hooks}.
#' @export
#' @aliases ClusterFunctions
#' @family ClusterFunctions
#' @family ClusterFunctionsHelper
makeClusterFunctions = function(name, submitJob, killJob = NULL, listJobsQueued = NULL, listJobsRunning = NULL,
  array.var = NA_character_, store.job.collection = FALSE, store.job.files = FALSE, scheduler.latency = 0,
  fs.latency = 0, hooks = list()) {
  assertList(hooks, types = "function", names = "unique")
  assertSubset(names(hooks), unlist(batchtools$hooks, use.names = FALSE))

  setClasses(list(
      name = assertString(name, min.chars = 1L),
      submitJob = assertFunction(submitJob, c("reg", "jc"), null.ok = TRUE),
      killJob = assertFunction(killJob, c("reg", "batch.id"), null.ok = TRUE),
      listJobsQueued = assertFunction(listJobsQueued, "reg", null.ok = TRUE),
      listJobsRunning = assertFunction(listJobsRunning, "reg", null.ok = TRUE),
      array.var = assertString(array.var, na.ok = TRUE),
      store.job.collection = assertFlag(store.job.collection),
      store.job.files = assertFlag(store.job.files),
      scheduler.latency = assertNumber(scheduler.latency, lower = 0),
      fs.latency = assertNumber(fs.latency, lower = 0),
      hooks = hooks),
    "ClusterFunctions")
}

#' @export
print.ClusterFunctions = function(x, ...) {
  catf("ClusterFunctions for mode: %s", x$name)
  catf("  List queued Jobs : %s", !is.null(x$listJobsQueued))
  catf("  List running Jobs: %s", !is.null(x$listJobsRunning))
  catf("  Kill Jobs        : %s", !is.null(x$killJob))
  catf("  Hooks            : %s", if (length(x$hooks)) stri_flatten(names(x$hooks), ",") else "-")
}

#' @title Create a SubmitJobResult
#'
#' @description
#' This function is only intended for use in your own cluster functions implementation.
#'
#' Use this function in your implementation of \code{\link{makeClusterFunctions}} to create a return
#' value for the \code{submitJob} function.
#'
#' @param status [\code{integer(1)}]\cr
#'   Launch status of job. 0 means success, codes between 1 and 100 are temporary errors and any
#'   error greater than 100 is a permanent failure.
#' @param batch.id [\code{character()}]\cr
#'   Unique id of this job on batch system, as given by the batch system.
#'   Must be globally unique so that the job can be terminated using just this information.
#'   For array jobs, this may be a vector of length equal to the number of jobs in the array.
#' @param log.file [\code{character()}]\cr
#'   Log file. If \code{NA}, defaults to \code{[job.hash].log}.
#'   Some cluster functions set this for array jobs.
#' @param msg [\code{character(1)}]\cr
#'   Optional error message in case \code{status} is not equal to 0. Default is \dQuote{OK},
#'   \dQuote{TEMPERROR}, \dQuote{ERROR}, depending on \code{status}.
#' @return [\code{\link{SubmitJobResult}}]. A list, containing
#'   \code{status}, \code{batch.id} and \code{msg}.
#' @family ClusterFunctionsHelper
#' @aliases SubmitJobResult
#' @export
makeSubmitJobResult = function(status, batch.id, log.file = NA_character_, msg = NA_character_) {
  status = asInt(status)
  if (is.na(msg)) {
    msg = if (status == 0L)
      "OK"
    else if (status <= 100L)
      "TEMPERROR"
    else
      "ERROR"
  }
  "!DEBUG [makeSubmitJobResult]: Result for batch.id '`paste0(batch.id, sep = ',')`': `status` (`msg`)"

  setClasses(list(status = status, batch.id = batch.id, log.file = log.file, msg = msg), "SubmitJobResult")
}

#' @export
print.SubmitJobResult = function(x, ...) {
  cat("Job submission result\n")
  catf("  ID    : %s", stri_flatten(x$batch.id, ","))
  catf("  Status: %i", x$status)
  catf("  Msg   : %s", x$msg)
}

#' @title Cluster Functions Helper to Parse a Brew Template
#'
#' @description
#' This function is only intended for use in your own cluster functions implementation.
#'
#' This function is only intended for use in your own cluster functions implementation.
#' Simply reads your template file and returns it as a character vector.
#'
#' @param template [\code{character(1)}]\cr
#'   Path to template file which is then passed to \code{\link[brew]{brew}}.
#' @param comment.string [\code{character(1)}]\cr
#'   Ignore lines starting with this string.
#' @return [\code{character}].
#' @family ClusterFunctionsHelper
#' @export
cfReadBrewTemplate = function(template, comment.string = NA_character_) {
  "!DEBUG [cfReadBrewTemplate]: Parsing template file '`template`'"
  lines = stri_trim_both(readLines(template))

  lines = lines[!stri_isempty(lines)]
  if (!is.na(comment.string))
    lines = lines[!stri_startswith_fixed(lines, comment.string)]
  if (length(lines) == 0L)
    stopf("Error reading template '%s' or empty template", template)
  return(stri_flatten(lines, "\n"))
}

#' @title Cluster Functions Helper to Write Job Description Files
#'
#' @description
#' This function is only intended for use in your own cluster functions implementation.
#'
#' Calls brew silently on your template, any error will lead to an exception.
#' The file is stored at the same place as the corresponding job file in the \dQuote{jobs}-subdir
#' of your files directory.
#'
#' @template reg
#' @param text [\code{character(1)}]\cr
#'   String ready to be brewed. See \code{\link{cfReadBrewTemplate}} to read a template from the file system.
#' @param jc [\code{\link{JobCollection})}]\cr
#'   Will be used as environment to brew the template file in. See \code{\link{JobCollection}} for a list of all
#'   available variables.
#' @return [\code{character(1)}]. File path to brewed template file.
#' @family ClusterFunctionsHelper
#' @export
cfBrewTemplate = function(reg, text, jc) {
  assertString(text)
  outfile = fs::path(dir(reg, "jobs"), sprintf("%s.job", jc$job.hash))

  parent.env(jc) = asNamespace("batchtools")
  on.exit(parent.env(jc) <- emptyenv())
  "!DEBUG [cfBrewTemplate]: Brewing template to file '`outfile`'"

  z = try(brew(text = text, output = outfile, envir = jc), silent = TRUE)
  if (is.error(z))
    stopf("Error brewing template: %s", as.character(z))
  waitForFile(outfile, reg$cluster.functions$fs.latency)
  return(outfile)
}

#' @title Cluster Functions Helper to Handle Unknown Errors
#'
#' @description
#' This function is only intended for use in your own cluster functions implementation.
#'
#' Simply constructs a \code{\link{SubmitJobResult}} object with status code 101, NA as batch id and
#' an informative error message containing the output of the OS command in \code{output}.
#'
#' @param cmd [\code{character(1)}]\cr
#'   OS command used to submit the job, e.g. qsub.
#' @param exit.code [\code{integer(1)}]\cr
#'   Exit code of the OS command, should not be 0.
#' @param output [\code{character}]\cr
#'   Output of the OS command, hopefully an informative error message.
#'   If these are multiple lines in a vector, they are automatically joined.
#' @return [\code{\link{SubmitJobResult}}].
#' @family ClusterFunctionsHelper
#' @export
cfHandleUnknownSubmitError = function(cmd, exit.code, output) {
  assertString(cmd, min.chars = 1L)
  exit.code = asInt(exit.code)
  assertCharacter(output, any.missing = FALSE)
  msg = sprintf("Command '%s' produced exit code %i. Output: '%s'", cmd, exit.code, stri_flatten(output, "\n"))
  makeSubmitJobResult(status = 101L, batch.id = NA_character_, msg = msg)
}

#' @title Cluster Functions Helper to Kill Batch Jobs
#'
#' @description
#' This function is only intended for use in your own cluster functions implementation.
#'
#' Calls the OS command to kill a job via \code{\link[base]{system}} like this: \dQuote{cmd batch.job.id}. If the
#' command returns an exit code > 0, the command is repeated after a 1 second sleep
#' \code{max.tries-1} times. If the command failed in all tries, an error is generated.
#'
#' @template reg
#' @param cmd [\code{character(1)}]\cr
#'   OS command, e.g. \dQuote{qdel}.
#' @param args [\code{character}]\cr
#'   Arguments to \code{cmd}, including the batch id.
#' @param max.tries [\code{integer(1)}]\cr
#'   Number of total times to try execute the OS command in cases of failures.
#'   Default is \code{3}.
#' @inheritParams runOSCommand
#' @return \code{TRUE} on success. An exception is raised otherwise.
#' @family ClusterFunctionsHelper
#' @export
cfKillJob = function(reg, cmd, args = character(0L), max.tries = 3L, nodename = "localhost") {
  assertString(cmd, min.chars = 1L)
  assertCharacter(args, any.missing = FALSE)
  assertString(nodename)
  max.tries = asCount(max.tries)

  for (i in seq_len(max.tries)) {
    res = runOSCommand(cmd, args, nodename = nodename)
    if (res$exit.code == 0L)
      return(TRUE)
    Sys.sleep(1)
  }

  stopf("Really tried to kill job, but failed %i times with '%s'.\nMessage: %s",
    max.tries, stri_flatten(c(cmd, args), " "), stri_flatten(res$output, "\n"))
}

getBatchIds = function(reg, status = "all") {
  cf = reg$cluster.functions
  tab = data.table(batch.id = character(0L), status = character(0L))
  batch.id = NULL

  if (status %chin% c("all", "running") && !is.null(cf$listJobsRunning)) {
    "!DEBUG [getBatchIds]: Getting running Jobs"
    x = unique(cf$listJobsRunning(reg))
    if (length(x) > 0L)
      tab = rbind(tab, data.table(batch.id = x, status = "running"))
  }

  if (status %chin% c("all", "queued") && !is.null(cf$listJobsQueued)) {
    "!DEBUG [getBatchIds]: Getting queued Jobs"
    x = chsetdiff(cf$listJobsQueued(reg), tab$batch.id)
    if (length(x) > 0L)
      tab = rbind(tab, data.table(batch.id = unique(x), status = "queued"))
  }

  submitted = done = batch.id = NULL
  batch.ids = reg$status[!is.na(submitted) & is.na(done) & !is.na(batch.id), unique(batch.id)]
  tab[batch.id %in% batch.ids]
}


#' @title Find a batchtools Template File
#'
#' @description
#' This functions returns the path to a template file on the file system.
#' @template template
#' @return [\code{character}] Path to the file or \code{NA} if no template template file was found.
#' @keywords internal
#' @export
findTemplateFile = function(template) {
  assertString(template, min.chars = 1L)

  if (stri_endswith_fixed(template, ".tmpl")) {
    assertFileExists(template, access = "r")
    return(fs::path_abs(template))
  }

  x = Sys.getenv("R_BATCHTOOLS_SEARCH_PATH")
  if (nzchar(x)) {
    x = fs::path(x, sprintf("batchtools.%s.tmpl", template))
    if (fs::file_access(x, "read"))
      return(fs::path_abs(x))
  }

  x = sprintf("batchtools.%s.tmpl", template)
  if (fs::file_access(x, "read"))
    return(fs::path_abs(x))

  x = fs::path(user_config_dir("batchtools", expand = FALSE), sprintf("%s.tmpl", template))
  if (fs::file_access(x, "read"))
    return(x)

  x = fs::path("~", sprintf(".batchtools.%s.tmpl", template))
  if (fs::file_access(x, "read"))
    return(fs::path_abs(x))

  x = fs::path(site_config_dir("batchtools"), sprintf("%s.tmpl", template))
  if (fs::file_access(x, "read"))
    return(x)

  x = system.file("templates", sprintf("%s.tmpl", template), package = "batchtools")
  if (fs::file_access(x, "read"))
    return(x)

  return(NA_character_)
}