-
Notifications
You must be signed in to change notification settings - Fork 0
/
slurm_monitor.py
executable file
·1254 lines (899 loc) · 46.1 KB
/
slurm_monitor.py
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
329
330
331
332
333
334
335
336
337
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
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
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
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
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
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
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
613
614
615
616
617
618
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
655
656
657
658
659
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
708
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
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
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
940
941
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
#!/usr/bin/env python3
# AUTHOR ===================================================================================================================
# Manuel Lippert (GitHub: ManeLippert (https://github.com/ManeLippert))
# ==========================================================================================================================
# MODULES ==================================================================================================================
import datetime, time, os, sys, subprocess, math, argparse, pkg_resources
# PARSER ===================================================================================================================
description_text = """
===================================== DESCRIPTION =====================================
FEATURES:
o NO REQUIREMENTS: Default script runs with standard python3 library
o Creates jobscript file with defined content
(look into variable "jobscriptContent" to add more option)
o Start/Restarts job until criteria is suffused (default=10000)
o Makes backup after each run before Restart and Restore files after failed run
(uses rsync command line utility)
o Reset option after run fails and dump files were written
(rely on "h5py" & "pandas" & "numpy" which get installed by the script)
o Creates status file with current status and progress bars and updates it dynamically
Progress: Total progress of simulation
Run X : Progress of current run
o Sends mail at the beginning, end and restart (default=False) with status file
as attachment
(mailx has to be installed, look into "send_mail" function for more info)
START SCRIPT IN BACKGROUND:
o WITH NOHUP:
Command:
>>> nohup python3 -u slurm_monitor.py --job-name $JOBNAME &> /dev/null &
Kill Process:
>>> python3 -u slurm_monitor.py --job-name $JOBNAME --kill
o WITH SCREEN (has to be installed):
Create Screen:
>>> screen -S $SESSION
Command:
>>> python3 -u slurm_monitor.py --job-name $JOBNAME --verbose
Leave Screen:
>>> ((Strg + a) + d)
Enter Screen:
>>> screen -r
or
>>> screen -r $SESSION
Kill Process:
>>> ^C (Strg + c)
or
>>> (Strg + d) (kill Screen itself)
OUTPUT STATUS:
Just run the file will create an dynamic output (recommended with using screen) or
>>> cd $DATA && find . -name $STATUSFILENAME -exec tail -8 {} \;
====================================== ARGUMENTS ======================================
"""
parser = argparse.ArgumentParser(description=description_text, formatter_class=argparse.RawTextHelpFormatter)
#parser._action_groups.pop()
required = parser.add_argument_group("required arguments")
required.add_argument("-j", "--job-name", dest="jobname", nargs="?", type=str, required= True,
help="job name not longer than 8 character")
additional = parser.add_argument_group("additional arguments")
additional.add_argument("-n", dest="timesteps", nargs="?", type=int, default=10000,
help="required timesteps (default=10000)")
additional.add_argument("-r", "--reset", dest="reset", action="store_true",
help="Uses Dumpfiles to reset Simulation (default=False)")
additional.add_argument("-v", "--verbose", dest="verbose", action="store_true",
help="activate output of script (default=False)")
additional.add_argument("-b", "--backup", dest="backup", nargs="?", type=str,
help="backup location for files (default=None)\n"+
"- local (creates backup in simulation folder)\n"+
"- home (creates backup in home folder)")
additional.add_argument("-d", "--dir", dest="directory", nargs="?", type=str, default=os.getcwd(),
help="directory of simulation (default=cwd)\n")
additional.add_argument("--jobscript", dest="jobscriptFile", nargs="?", type=str, default="jobscript-create",
help="jobscript to run SLURM job (default=jobscript-create)")
additional.add_argument("--ntask-per-node", dest="tasks", nargs="?", type=str, default="32",
help="MPI task per node (default=32)")
additional.add_argument("--nodes", dest="nodes", nargs="?", type=str, default="3",
help="number of nodes (default=3)")
additional.add_argument("--walltime", dest="wallTime", nargs="?", type=str, default="0-24:00:00",
help="walltime of server (d-hh:mm:ss) (default=0-24:00:00)")
additional.add_argument("--mail", dest="mail", nargs="?", type=str,
help="mail address ([email protected]) (default=None)")
additional.add_argument("--restart-mail", dest="restartmail", action="store_true",
help="mail after every restart (default=False)")
additional.add_argument("--statusfile", dest="statusFile", nargs="?", type=str, default="status.txt",
help="file with output from nohup command (default=status.txt)")
additional.add_argument("--restartfile", dest="restartFile", nargs="?", type=str, default="FDS.dat",
help="restart file with data (default=FDS.dat)")
additional.add_argument("--format", dest="formattable", nargs="?", type=str,
help="format of output table (default=None)\n"+
"- fancy (round box)\n"+
"- universal (crossplattform)\n"+
"- None (No frame)")
additional.add_argument("--refresh-rate", dest="sleepTime", nargs="?", type=int, default=300,
help="time interval to check status in sec (default=300)")
additional.add_argument("--kill", dest="kill", action="store_true",
help="kills monitor process (default=False)")
args = parser.parse_args()
# PARSER VARIABLES =========================================================================================================
emailAddress = args.mail
RESTARTMAIL = args.restartmail
backupLocation = args.backup
folder = args.directory
jobName = args.jobname
tasks = args.tasks
nodes = args.nodes
wallTime = args.wallTime
nTimestepsRequired = args.timesteps
jobscriptFilename = args.jobscriptFile
restartFilename = args.restartFile
restartBin = restartFilename.replace(".dat", "")
statusFilename = args.statusFile
formatTable = args.formattable
VERBOSE = args.verbose
RESET = args.reset
# Kill process of monitoring
kill = args.kill
# Changing this value can cause problems in writing status file
sleepTime = args.sleepTime
# VARIABLES ================================================================================================================
outputCriteria = ["0", "Run successfully completed"]
slurmFiles = [f for f in os.listdir() if "slurm-" in f]
runCounter = len(slurmFiles)
nTimestepsCurrent = 0
currentTime = "00:00:00"
dataFilename = "gkwdata.h5"
check1Filename, check2Filename = "DM1.dat", "DM2.dat"
check1Bin, check2Bin = check1Filename.replace(".dat", ""), check2Filename.replace(".dat", "")
def outputFilename(info):
return "slurm-" + info + ".out"
## KILL JOB ================================================================================================================
commandMonitorKill = "ps ax | grep " + jobName + " | grep -v grep | grep -v kill | awk '{print $1}'"
PID = subprocess.getoutput(commandMonitorKill)
if kill:
subprocess.run(["kill", PID])
quit()
## JOBSCRIPT CONTENT =======================================================================================================
jobscriptContent = """#!/bin/bash -l
# jobname
#SBATCH --job-name=""" + jobName + """
# MPI tasks
#SBATCH --ntasks-per-node=""" + tasks + """
# number of nodes
#SBATCH --nodes=""" + nodes + """
# walltime
# d-hh:mm:ss
#SBATCH --time=""" + wallTime + """
# execute the job
time mpirun -np $SLURM_NTASKS """ + folder + "/" + """gkw.x > output.dat
# end
exit 0
"""
jobName = jobName[0:8]
## FLAGS ===================================================================================================================
startOutputFlag = "Submitted batch job "
restartFlag = "FILE_COUNT"
# SWITCHES =================================================================================================================
if emailAddress==None:
EMAIL = False
else:
EMAIL = True
if backupLocation==None:
BACKUP = False
else:
BACKUP = True
## PATHS ===================================================================================================================
user = os.getlogin()
if folder[-1] == "/":
folder = folder[:-1]
path = folder.split(user + "/")[1]
# Set backup location
if BACKUP:
# If local has been specified as backupLocation, then the backup is copied to the same directory as the simulation folder
# (simFolder) is located in. The backup is copied to a directory with name simFolder + "-backup".
if backupLocation == "local":
simFolder = path.split("/")[-1]
backupPath = folder + "/../" + simFolder + "-backup"
# Create backup in home folder
elif backupLocation == "home":
simFolder = path.split("/")[-1]
backupPath = "~/" + simFolder + "-backup"
# Otherwise, use the backupLocation parsed as argument.
else:
if backupLocation[-1] != "/":
backupLocation += "/"
backupPath = backupLocation + path
if not os.path.exists(backupPath):
os.makedirs(backupPath)
## COMMANDS ================================================================================================================
commandJobRunning = "squeue --states=running --name " + jobName
commandJobPending = "squeue --states=pending --name " + jobName
commandJobStarting = "sbatch"
if BACKUP:
commandBackup = ["rsync", "-a", folder + "/", backupPath]
commandRestore = ["rsync", "-a", "-I", "--exclude=" + statusFilename, backupPath + "/", folder]
# FUNCTIONS ================================================================================================================
## PROGRESSBAR =============================================================================================================
def progressbar(required_value, current_value, progressbar_width,
prefix="", prefix_width = 10, percent_width = 8, ratio_width = 13,
progress_fill="=", progress_fill_top="", progress_fill_bot="",
progress_unfill=".",
progress_bracket=["[","]"]):
prefix_format = "{:<" + str(prefix_width) + "}"
percentage = int(100*current_value/required_value)
percent = "(" + " "*(3-len(str(percentage))) + str(percentage) + "%)"
percent_format = "{:>" + str(percent_width) + "}"
ratio = str(current_value) + "/" + str(required_value)
ratio_format = "{:>" + str(ratio_width) + "}"
progressbar_bracket_width = (len(progress_bracket[0]) + len(progress_bracket[1])
+ len(progress_fill_top) + len(progress_fill_bot))
barsize = progressbar_width - prefix_width - percent_width - ratio_width - progressbar_bracket_width
if percentage <= 100:
x = int(barsize*current_value/required_value)
else:
x = barsize
progress_format = progress_bracket[0] + progress_fill_bot + "{}" + progress_fill_top + "{}" + progress_bracket[1]
bar_format = prefix_format +progress_format + percent_format + ratio_format
bar = bar_format.format(prefix, progress_fill*x, progress_unfill*(barsize-x), percent, ratio)
return bar
## OUTPUT TABLE ============================================================================================================
def message(string, add):
string = string + add + "\n"
return string
def print_table_row(content,
delete_line_index = -10, table_width = 80,
output_type = None,
TIMEINFO = True, WRITEFILE=True):
current_value, required_value = nTimestepsCurrent, nTimestepsRequired,
run_counter = runCounter
current_time, required_time = currentTime, wallTime
msg=""
table_inner_width = table_width - 4
table_content_width = table_inner_width - 47
if formatTable == "fancy":
table_outline = ["╭─", "─╮", "╰─", "─╯", "├─", "─┤", "│ ", " │", "─"]
if formatTable == "universal":
table_outline = ["+-", "-+", "+-", "-+", "+-", "-+", "| ", " |", "-"]
if formatTable == None:
table_outline = ["--", "--", "--", "--", "--", "--", " ", " ", "-"]
sep_top = table_outline[0] + table_inner_width*table_outline[8] + table_outline[1]
sep_mid = table_outline[4] + table_inner_width*table_outline[8] + table_outline[5]
sep_end = table_outline[2] + table_inner_width*table_outline[8] + table_outline[3]
row_cols = [2, 10, table_content_width, 13, 11, 13, 2]
row_format = "".join(["{:<" + str(col) + "}" for col in row_cols])
progress_cols = [2, table_inner_width, 2]
progress_format = "".join(["{:<" + str(col) + "}" for col in progress_cols])
progressbar_content = [progressbar(required_value, current_value, table_inner_width,
progress_fill_top=">", prefix="PROGRESS")]
progressbar_content.insert(0, table_outline[6])
progressbar_content.insert(len(progressbar_content), table_outline[7])
required_time = get_time_in_seconds(required_time)
current_time = get_time_in_seconds(current_time)
jobStatus = subprocess.getoutput("squeue --name " + jobName).strip().split("\n")
try:
jobStatusHeader = [" " + jobStatus[0]]
jobStatusInfo = [jobStatus[1][11:11+table_inner_width]]
except IndexError:
jobStatus_cols = [12, 12, 10, table_inner_width - 71, 13, 11, 13]
jobStatus_format = "".join(["{:<" + str(col) + "}" for col in jobStatus_cols])
jobStatusHeader = ["OUTPUT", "NAME", "USER", "" ,"DATE", "TIME", "W:DD:HH:MM:SS"]
jobStatusInfo = [content[0], jobName, user, "", time_date(), time_time(), time_duration(startTime, pastTime)]
jobStatusHeader = [jobStatus_format.format(*jobStatusHeader)]
jobStatusInfo = [jobStatus_format.format(*jobStatusInfo)]
jobStatusHeader.insert(0, table_outline[6])
jobStatusHeader.insert(len(jobStatusHeader), table_outline[7])
jobStatusInfo.insert(0, table_outline[6])
jobStatusInfo.insert(len(jobStatusInfo), table_outline[7])
progressbartime_content = [progressbar(required_time, current_time, table_inner_width,
progress_fill_top=">", prefix="RUN " + str(run_counter))]
progressbartime_content.insert(0, table_outline[6])
progressbartime_content.insert(len(progressbartime_content), table_outline[7])
content[1] = content[1][-(table_content_width-1):]
content.insert(0, table_outline[6])
if TIMEINFO:
if output_type == "header":
content.append("DATE")
content.append("TIME")
content.append("W:DD:HH:MM:SS")
else:
content.append(time_date())
content.append(time_time())
content.append(time_duration(startTime, pastTime))
content.insert(len(content), table_outline[7])
if output_type == "header":
delete_line_index = -1
msg += sep_top + "\n"
msg += row_format.format(*content) + "\n"
msg += sep_end + "\n"
elif output_type == "middle":
if VERBOSE:
sys.stdout.write("\x1b[1A"*(-delete_line_index + 1))
msg += sep_mid + "\n"
msg += row_format.format(*content) + "\n"
msg += sep_end + "\n"
elif output_type == "update":
if VERBOSE:
sys.stdout.write("\x1b[1A"*(-delete_line_index + 1))
msg += sep_end + "\n"
else:
if VERBOSE:
sys.stdout.write("\x1b[1A"*(-delete_line_index + 1))
msg += row_format.format(*content) + "\n"
msg += sep_end + "\n"
msg += " "*table_width + "\n" + " "*table_width + "\n"
msg += sep_top + "\n"
msg += progress_format.format(*jobStatusHeader) + "\n"
msg += progress_format.format(*jobStatusInfo) + "\n"
msg += sep_mid + "\n"
msg += progress_format.format(*progressbar_content) + "\n"
msg += progress_format.format(*progressbartime_content) + "\n"
msg += sep_end + "\n"
if VERBOSE:
print(msg, flush=True)
if WRITEFILE:
delete_write_line_to_file(statusFilename, msg, end=delete_line_index)
## PIP INSTALL =============================================================================================================
def pip_install(modules):
required = modules
installed = {pkg.key for pkg in pkg_resources.working_set}
missing = required - installed
if missing:
subprocess.check_call([sys.executable, "-m", "pip", "install", "--user", *missing],
stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
print_table_row(["INSTALL", "Required Modules installed"])
else:
print_table_row(["CHECK", "Modules already installed"])
## FILE ====================================================================================================================
def get_value_of_variable_from_file(filename, file_index, relative_index, string):
try:
content = [i.strip().split() for i in open(folder + "/" + filename).readlines()]
index = [idx for idx, s in enumerate(content) if string in s][file_index]
value = content[index][relative_index]
return value
except IndexError:
print_table_row(["ERROR", "String not found in file"])
quit()
def find_string_in_file(filename, string):
with open(folder + "/" + filename) as f:
if string in f.read():
return True
else:
return False
def delete_write_line_to_file(filename, add = "", start=None, end=None):
with open(folder + "/" + filename, "r") as file:
try:
lines = file.readlines()[start:end]
content = "".join(lines)
except IndexError:
pass
content += add
write_file(filename, content)
def write_file(filename, content):
with open(folder + "/" + filename, "w") as file:
file.write(content)
file.flush()
# Reset gkwdata.h5 with the use of dump files DM1, DM2
# AUTHOR: Florian Rath
# IMPORT: gkw_reset_checkpiont.py (https://bitbucket.org/gkw/gkw/src/develop/python/gkw_reset_checkpoint.py)
def reset_simulation(SIM_DIR, NTIME=None, use_ntime=False):
# Function that delets all data in interval [nt_reset:nt_broke].
# ncol considers, if data series is ordered by a multiple interger of
# ntime.
def reset_time_trace(indata, dim, ncol, nt_reset):
# Shift dimension dim to 0.
data_shifted = np.moveaxis(indata, dim, 0)
# Reset data.
data_shifted_trimmed = data_shifted[0:int(nt_reset*ncol),]
# Shift dimension back.
out = np.moveaxis(data_shifted_trimmed, 0, dim)
# free memory
del data_shifted
del data_shifted_trimmed
return out
# Checks if file has binary format.
def is_binary(filename):
try:
with open(filename, "tr") as check_file:
check_file.read()
return False
except:
return True
# Check for specific files that are no ordinary data files.
def is_file_exception(filename):
# substrings that have to be checked
check_list = ["geom.dat", "DM1.dat", "DM2.dat", "FDS.dat", ".o", "FDS",
"input.dat", "perform_first.dat", "perform.dat", "output.dat",
"gkwdata.meta", "gkw_hdf5_errors.txt", "kx_connect.dat",
"jobscript", "Poincare1.mat", "perfloop_first.dat", "par.dat",
"input_init.dat", "sgrid", "gkw", ".out", statusFilename]
for key in check_list:
if key in filename:
return True
return False
# Check if given file is a PBS or SLURM jobscript.
def is_jobscript(filename):
with open(filename,"r") as file:
for line in file:
if "#PBS -l" in line:
return True
if "#SBATCH" in line:
return True
return False
def get_timestep(FILE):
if(os.path.isfile(SIM_DIR+"/"+FILE+".dat")):
EXISTS = True
with open(SIM_DIR+"/"+FILE+".dat","r") as file:
for line in file:
if "NT_REMAIN" in line:
expr = line.replace(" ","")
expr = expr.replace(",","")
expr = expr.replace("\n","")
REMAIN = int(expr.split("=")[1])
if "NT_COMPLETE" in line:
expr = line.replace(" ","")
expr = expr.replace(",","")
expr = expr.replace("\n","")
COMPLETE = int(expr.split("=")[1])
else:
REMAIN, COMPLETE, EXISTS = None, None, False
return REMAIN, COMPLETE, EXISTS
HDF5_FILENAME = "gkwdata.h5"
RESTARTFILE,DUMPFILE1, DUMPFILE2 = "FDS","DM1", "DM2"
# Change to simulation directory.
if(not os.path.isdir(SIM_DIR)):
return
else:
os.chdir(SIM_DIR)
# Check if hdf5-file exists.
if(not os.path.isfile(SIM_DIR + "/" + HDF5_FILENAME)):
return
# First, read hdf5 file and determine the number of big time steps NTIME,
# requested in the input.dat file.
f = h5py.File(SIM_DIR + "/" + HDF5_FILENAME, "r+")
# Get requested big time steps from the /control group in the hdf5-file.
if(NTIME==None):
NTIME = int(f["input/control/ntime"][:])
# Get number of big time steps after which simulation broke.
# If time.dat exists read this file to obtain number of time steps after
# which simulation broke.
if(os.path.isfile(SIM_DIR + "/" + "time.dat")):
tim = pd.read_csv(SIM_DIR+"time.dat", header=None, sep="\s+").values
NT_BROKE = tim.shape[0]
# Else, get time from hdf5-file.
else:
NT_BROKE = f["diagnostic/diagnos_growth_freq/time"].shape[1]
# Set NT_BROKE for output files holding temporal derivates and therefore
# one timestep less.
NT_BROKE_DERIV = NT_BROKE-1
# Close the hdf5-file again.
f.close()
# Get the number of remaining big time steps NT_REMAIN from checkpoint
# files FDS.dat. This is used lateron to determine the most recent
# checkpoint file.
NT_REMAIN, NT_COMPLETE, FDS_EXISTS = get_timestep(RESTARTFILE)
# Get the number of remaining big time steps NT_REMAIN[1/2] from checkpoint
# files DM[1/2]. This is used lateron to determine the most recent
# checkpoint file.
NT_REMAIN1, NT_COMPLETE1, DM1_EXISTS = get_timestep(DUMPFILE1)
NT_REMAIN2, NT_COMPLETE2, DM2_EXISTS = get_timestep(DUMPFILE1)
# Check if FDS is the most recent checkpoint file. In this case
# resetting the simulation makes no sense.
if(FDS_EXISTS):
DM1_OLD, DM2_OLD = False, False
if(DM1_EXISTS):
if(NT_COMPLETE1 < NT_COMPLETE):
DM1_OLD = True
if(DM2_EXISTS):
if(NT_COMPLETE2 < NT_COMPLETE):
DM2_OLD = True
if(DM1_OLD and DM2_OLD):
return
# Now determine which checkpoint file is the recent one.
if(DM1_EXISTS and DM2_EXISTS):
if(NT_REMAIN1 > NT_REMAIN2):
NT_REMAIN, NT_COMPLETE, DUMPFILE = NT_REMAIN2, NT_COMPLETE2, DUMPFILE2
else:
NT_REMAIN, NT_COMPLETE, DUMPFILE = NT_REMAIN1, NT_COMPLETE1, DUMPFILE1
elif(DM1_EXISTS and not DM2_EXISTS):
NT_REMAIN, NT_COMPLETE, DUMPFILE = NT_REMAIN1, NT_COMPLETE1, DUMPFILE1
elif(DM2_EXISTS and not DM1_EXISTS):
NT_REMAIN, NT_COMPLETE, DUMPFILE = NT_REMAIN2, NT_COMPLETE2, DUMPFILE2
else:
return
if(not use_ntime):
# Find the total number ob big time steps the simulation time trace should have,
# when considering the big time steps already completed as well as the big time
# steps that remain. Can be different from NTIME, since the simulation could
# have been restarted several times such that NTIME > NT_COMPLETE.
NTOT = NT_COMPLETE + NT_REMAIN
N_REQUEST = NTOT
# Determine the time steps to which the time trace files have to be reset.
# Use NTOT here, since NTIME could have been changed at some point, or NT_COMPLETE
# could be larger than NTIME.
NT_RESET = NTOT-NT_REMAIN
else:
# Determine the time steps to which the time trace files have to be reset.
NT_RESET = NTIME-NT_REMAIN
N_REQUEST = NTIME
# Same for files holding time derivatives.
NT_RESET_DERIV = NT_RESET-1
# ----------------------------------------------------------------------
# Cycle over all nodes of hdf5-file and reset time trace datasets.
# Check if hdf5-file exists.
if(os.path.isfile(SIM_DIR + "/" + HDF5_FILENAME)):
# Find all possible keys items, i.e. both groups and datasets
f = h5py.File(SIM_DIR + "/" + HDF5_FILENAME, "a")
h5_keys = []
f.visit(h5_keys.append)
# Cycle over all keys items and check, if any dimension has size NT_BROKE,
# i.e., it is a time trace file.
for item in enumerate(h5_keys):
data = f.get(item)
# Consider datasets only.
if(isinstance(data, h5py.Dataset)):
#Check if any dimension is an integer multiple of NT_BROKE, by checking
# the residual of the division.
res = [None]*len(data.shape)
for i in range(len(data.shape)):
res[i] = data.shape[i]/NT_BROKE-np.floor(data.shape[i]/NT_BROKE)
if 0.0 in res:
# Check which dimension is integer multiple of NT_BROKE
# and save dimension as well as integer.
new_shape = data.shape
for i in range(len(data.shape)):
ncol = data.shape[i]/NT_BROKE
res = ncol - np.floor(ncol)
if(res == 0):
dim = i
ncol = int(ncol)
# adjust new shape to ncol*NT_RESET
y = list(new_shape)
y[dim] = int(ncol*NT_RESET)
new_shape = tuple(y)
break
# Reset dataset (.resize discards data with indices larger than
# ncol*NT_RESET along dimension dim).
dset = f[item]
dset.resize(int(ncol*NT_RESET),dim)
# After having repaired all datasets, close the hdf5-file again.
f.close()
# ----------------------------------------------------------------------
# Cycle over all csv-files and reset time trace.
for filename in os.listdir(SIM_DIR):
# First perform some checks on files; cycle if file is binary, an exception
# or a jobscript.
if(is_binary(filename)):
continue
if(is_file_exception(filename)):
continue
if(is_jobscript(filename)):
continue
# no file
if(not os.path.isfile(filename)):
continue
# Load csv file.
data = pd.read_csv(filename, header=None, sep="\s+").values
#Check if any dimension is an integer multiple of NT_BROKE, by checking
#the residual of the division.
res = [None]*len(data.shape)
# residul for output that holds time derivatives and therefore nt-1 datapoints
res_deriv = [None]*len(data.shape)
for i in range(len(data.shape)):
res[i] = data.shape[i]/NT_BROKE-np.floor(data.shape[i]/NT_BROKE)
res[i] = data.shape[i]/(NT_BROKE_DERIV)-np.floor(data.shape[i]/(NT_BROKE_DERIV))
# ordinary files
if 0.0 in res:
#Check which dimension is integer multiple of NT_BROKE
#and save dimension as well as integer.
for i in range(len(data.shape)):
ncol = data.shape[i]/NT_BROKE
res = ncol - np.floor(ncol)
if(res == 0):
dim = i
ncol = int(ncol)
break
# Load original dataset.
original_data = data
# print("\t Original shape: \t"+str(original_data.shape))
# Reset time trace.
reset_data = reset_time_trace(original_data,dim,ncol,NT_RESET)
# print("\t Reset shape: \t" +str(reset_data.shape))
# Save resetted data.
pd.DataFrame(reset_data).to_csv(filename, sep="\t", header=None, index=None)
# files holding time derivatives
if 0.0 in res_deriv:
# Check which dimension is integer multiple of NT_BROKE
# and save dimension as well as integer.
for i in range(len(data.shape)):
ncol = data.shape[i]/NT_BROKE_DERIV
res = ncol - np.floor(ncol)
if(res == 0):
dim = i
ncol = int(ncol)
break
# Load original dataset.
original_data = data
# Reset time trace.
reset_data = reset_time_trace(original_data,dim,ncol,NT_RESET_DERIV)
# Save resetted data.
pd.DataFrame(reset_data).to_csv(filename, sep="\t", header=None, index=None)
# ----------------------------------------------------------------------
# Finally, copy most recent dump file to FDS[/.dat].
# Copy the most recent dump file to FDS[/.dat].
copyfile(SIM_DIR+"/"+DUMPFILE, SIM_DIR+"/"+"FDS")
copyfile(SIM_DIR+"/"+DUMPFILE+".dat", SIM_DIR+"/"+"FDS.dat")
# First line of the so produced FDS.dat has to be modified.
old_text = "!Dump filename: "+DUMPFILE
new_text = "!Dump filename: "+"FDS"
# Replace first line in FDS.dat to set the correct file name.
with fileinput.input(SIM_DIR+"/"+"FDS.dat",inplace=True) as f:
for line in f:
line.replace(old_text, new_text)
def check_and_delete_file(filenames):
for filename in filenames:
if(os.path.isfile(folder + "/" + filename)):
os.remove(folder + "/" + filename)
def check_checkpoint_files():
DM1, DM2 = False, False
if(os.path.isfile(folder + "/" + check1Filename) and os.path.isfile(folder + "/" + check1Bin)):
DM1 = True
if(os.path.isfile(folder + "/" + check2Filename) and os.path.isfile(folder + "/" + check2Bin)):
DM2 = True
return DM1, DM2
def get_timestep_from_restartfile(filename, flag):
return int(get_value_of_variable_from_file(filename, 0, 2, flag).replace(",", ""))
def get_ntimestepCurrent(filenames):
ntimestep = 0
for f in filenames:
if(os.path.isfile(folder + "/" + f)):
ntimestepFile = get_timestep_from_restartfile(f, restartFlag)
if ntimestepFile > ntimestep:
ntimestep = ntimestepFile
return ntimestep
def get_time_from_statusfile(filename, line_index):
with open(folder + "/" + filename, "r") as file:
line = file.readlines()[line_index]
content = line.split(" ")
if formatTable == None:
time = content[-3]
else:
time = content[-2]
time_sec = get_time_in_seconds(time)
return time_sec
## TIME ====================================================================================================================
def format_num(time):
if time < 10:
return "0" + str(time)
else:
return str(time)
def get_time_as_string(sec):
mins, sec = sec // 60, sec % 60
hours, mins = mins // 60, mins % 60
days, hours = hours // 24, hours % 24
weeks, days = days // 7, days % 7
timeConvertedString = (str(int(weeks)) + ":" +
format_num(int(days)) + ":" + format_num(int(hours)) + ":" +
format_num(int(mins)) + ":" + format_num(int(sec)))
return timeConvertedString
def get_time_in_seconds(time):
# Format D-HH:MM:SS or HH:MM:SS or MM:SS
time = time.replace("-", ":")
time = time.replace("\n", "")
time_split = time.split(":")
seconds = [7*24*60*60,24*60*60, 60*60, 60, 1]
seconds = seconds[-len(time_split):]
time_sec = sum([a*b for a,b in zip(seconds, map(int,time_split))])
return time_sec
def time_date():
e = datetime.datetime.now()
return "%s-%s-%s" % (format_num(e.year), format_num(e.month), format_num(e.day))
def time_time():
e = datetime.datetime.now()
return "%s:%s:%s" % (format_num(e.hour), format_num(e.minute), format_num(e.second))
def time_duration(startTime, pastTime):
stop = time.time()
return get_time_as_string(stop - startTime + pastTime)
## STATUS ==================================================================================================================
def get_job_status():
jobStatusRunning = subprocess.getoutput(commandJobRunning).strip().split()
jobStatusPending = subprocess.getoutput(commandJobPending).strip().split()
return jobStatusRunning, jobStatusPending
def set_output_type():
jobStatusRunning, jobStatusPending = get_job_status()
if jobName in jobStatusRunning:
outputType = "running"
elif jobName in jobStatusPending:
outputType = "pending"
else:
outputType ="no Output"
return outputType
def get_error_type(filename):
slurm_errors = {"executable":["error on file ./gkw.x (No such file or directory)", "No executable found"],
"walltime":["process killed (SIGTERM)", "Exceeded wall time"],
"timeout":["DUE TO TIME LIMIT", "Exceeded time limit"],
"config":["couldn't open config directory", "Config not loading"],
"hdf5":["HDF5-DIAG", "Writing h5 file failed"]}
for key in slurm_errors:
if find_string_in_file(filename, slurm_errors[key][0]):
return slurm_errors[key][1]
return "Unknown error occurred"
## MAIL ====================================================================================================================
def send_mail(recipient, subject, body = None):
recipient = recipient.encode("utf_8")
subject = subject.replace(" ", "_")
subject = subject.encode("utf_8")
if body == None:
body = "For futher information open attachment"
body = body.encode("utf_8")
attachmentPath = folder + "/" + statusFilename
attachment = attachmentPath.encode("utf_8")
process = subprocess.Popen(["ssh", "master", "/usr/bin/mailx", "-s", subject, "-a", attachment, recipient],
stdin=subprocess.PIPE)
process.communicate(body)
# START/RESTART JOB ========================================================================================================
startTime = time.time()
outputType = set_output_type()
jobStatusRunning, jobStatusPending = get_job_status()
# Set current Time for progress bar