initexit.py 29.9 KB
Newer Older
1
2
3
4
#!/usr/bin/env python
# da_initexit.py

"""
Peters, Wouter's avatar
Peters, Wouter committed
5
6
.. module:: initexit
.. moduleauthor:: Wouter Peters 
7
8
9
10

Revision History:
File created on 13 May 2009.

Peters, Wouter's avatar
Peters, Wouter committed
11
The CycleControl class is found in the module :mod:`initexit`. It is derived from the standard python :class:`dictionary` object. It is the only core object of CTDAS that is automatically created in the pipeline, the user (normally) does not need to modify or extend it. The class is created based on options and arguments passes on the command line when submitting your main CTDAS job. 
12
13
14
15
16
17
18
19
20
21

Valid options are defined in 

.. autofunction:: da.tools.initexit.ParseOptions

With the name of a valid ``rc-file``, the CycleControl object is instantiated and validated. An example rc-file looks
like this:::

    ! Info on the data assimilation cycle

22
23
24
25
26
27
    time.restart        : False                     ! Restart from an existing run T/F
    time.start          : 2000-01-01 00:00:00       ! Start time of first cycle
    time.finish         : 2000-01-08 00:00:00       ! End time of last cycle
    time.cycle          : 7                         ! length of each cycle, 7 means one week
    time.nlag           : 5                         ! number of cycles in one smoother window
    dir.da_run          : ${HOME}/tmp/test_da       ! the run directory for you project
28
29
30

    ! Info on the DA system used

31
32
    da.system           : CarbonTracker             ! an identifier for your inversion system
    da.system.rc        : da/rc/carbontracker.rc    ! the settings needed in your inversion system
33
34
35

    ! Info on the forward model to be used

36
37
38
    da.obsoperator         : TM5                                ! an identifier for your observation operator
    da.obsoperator.rc      : ${HOME}/Modeling/TM5/tm5-ctdas.rc  ! the rc-file needed to run youobservation operator
    da.optimizer.nmembers  : 30                                 ! the number of ensemble members desired in the optimization
39
40
41
42

The most important method of the CycleControl object are listed below:

.. autoclass:: da.tools.initexit.CycleControl 
43
44
   :members: Initialize, Finalize,  collect_restart_data, move_restart_data, 
             submit_next_cycle, setup_file_structure, recover_run, random_seed
45
46
47
48
49
50
51
52
53
54
55

Two important attributes of the CycleControl object are:
    (1) DaSystem, an instance of a :ref:`dasystem`
    (2) DaPlatForm, an instance of a :ref:`platform`

Other functions in the module initexit that are related to the control of a DA cycle are:

.. autofunction:: da.tools.initexit.StartLogger 
.. autofunction:: da.tools.initexit.ValidateOptsArgs 


56
"""
karolina's avatar
karolina committed
57
58
59
60
import logging
import os
import sys
import shutil
61

karolina's avatar
karolina committed
62
needed_da_items = [
63
64
65
66
67
    'time.start',
    'time.finish',
    'time.nlag',
    'time.cycle',
    'dir.da_run',
68
69
70
71
72
    'da.system',
    'da.system.rc',
    'da.obsoperator',
    'da.obsoperator.rc',
    'da.optimizer.nmembers']
73
74
75
76
77
78
79

# only needed in an earlier implemented where each substep was a separate job
# validprocesses = ['start','done','samplestate','advance','invert']


class CycleControl(dict):
    """
80
    This object controls the CTDAS system flow and functionality.
81
    """
82
        
karolina's avatar
karolina committed
83
    def __init__(self, opts=[], args={}):
84
        """
Peters, Wouter's avatar
Peters, Wouter committed
85
86
87
88
        The CycleControl object is instantiated with a set of options and arguments.
        The list of arguments must contain the name of an existing ``rc-file``. 
        This rc-file is loaded by method :meth:`~da.tools.initexit.CycleControl.LoadRc` and validated
        by :meth:`~da.tools.initexit.CycleControl.ValidateRC`
89

Peters, Wouter's avatar
Peters, Wouter committed
90
91
        Options for the CycleControl consist of accepted command line flags or arguments 
        in :func:`~da.tools.initexit.CycleControl.ParseOptions`
92

93
94
95
96
97
98
99
100
        """

        self.LoadRc(args['rc'])
        self.ValidateRC()
        self.opts = opts

        # Add some useful variables to the rc-file dictionary

karolina's avatar
karolina committed
101
102
        self['jobrcfilename'] = self.RcFileName
        self['dir.da_submit'] = os.getcwd()
103
        self['da.crash.recover'] = '-r' in opts
karolina's avatar
karolina committed
104
105
106
107
        self['verbose'] = '-v' in opts
        self.DaSystem = None # to be filled later
        self.RestartFileList = [] # List of files needed for restart, to be extended later
        self.OutputFileList = [] # List of files needed for output, to be extended later
108
109

    def __str__(self):
karolina's avatar
karolina committed
110
111
112
113
114
115
116
        msg = "===============================================================" 
        msg += "DA Cycle rc-file is %s" % self.RcFileName            
        msg += "DA Cycle run directory is %s" % self['dir.da_run']   
        msg += "DA Cycle inverse system is %s" % self['da.system']   
        msg += "DA Cycle obs operator is %s" % self['da.obsoperator'] 
        msg += "===============================================================" 
        return msg
117
118


karolina's avatar
karolina committed
119
    def LoadRc(self, RcFileName):
120
121
122
        """ 
        This method loads a DA Cycle rc-file with settings for this simulation 
        """
123
        import da.tools.rc as rc
124
125

        rcdata = rc.read(RcFileName)
karolina's avatar
karolina committed
126
        for k, v in rcdata.iteritems():
127
            self[k] = v
karolina's avatar
karolina committed
128
129
        self.RcFileName = RcFileName
        self.DaRcLoaded = True
130

karolina's avatar
karolina committed
131
        logging.info('DA Cycle rc-file (%s) loaded successfully' % self.RcFileName)
132
133
134
135


    def ValidateRC(self):
        """ 
Peters, Wouter's avatar
Peters, Wouter committed
136
137
        Validate the contents of the rc-file given a dictionary of required keys. 
        Currently required keys are :attr:`~da.tools.initexit.needed_da_items`
138
139
140
        """
        from da.tools.general import ToDatetime

karolina's avatar
karolina committed
141
142
143
144
145
        for k, v in self.iteritems():
            if v in ['True', 'true', 't', 'T', 'y', 'yes']:
                self[k] = True
            if v in ['False', 'false', 'f', 'F', 'n', 'no']:
                self[k] = False
146
147
148
149
150
151
152
153
154
            if 'date' in k : self[k] = ToDatetime(v)
            if 'time.start' in k : 
                self[k] = ToDatetime(v)
            if 'time.end' in k : 
                self[k] = ToDatetime(v)
            if 'time.finish' in k : 
                self[k] = ToDatetime(v)
        for key in needed_da_items:
            if not self.has_key(key):
karolina's avatar
karolina committed
155
156
157
158
159
160
161
162
                msg = 'Missing a required value in rc-file : %s' % key
                logging.error(msg)
                logging.error('!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ')
                logging.error('Please note the update on Dec 02 2011 where rc-file names for DaSystem and ')
                logging.error('are from now on specified in the main rc-file (see da/rc/da.rc for example)')
                logging.error('!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ')
                raise IOError, msg
        logging.debug('DA Cycle settings have been validated succesfully')
163
164
165
166
167
168
169
170
171
172
173

    def ParseTimes(self):
        """ 
        Parse time related parameters into datetime objects for later use 
        """
        from da.tools.general import AdvanceTime

        startdate = self['time.start']
        finaldate = self['time.finish']                  

        if finaldate <= startdate:
karolina's avatar
karolina committed
174
            logging.error('The start date (%s) is not greater than the end date (%s), please revise' % (startdate.strftime('%Y%m%d'), finaldate.strftime('%Y%m%d'))) 
175
176
177
178
179
180
181
182
            raise ValueError
        cyclelength = self['time.cycle']                 # get time step

# Determine end date

        if cyclelength == 'infinite':
            enddate = finaldate
        else:
karolina's avatar
karolina committed
183
            enddate = AdvanceTime(startdate, cyclelength)
184

karolina's avatar
karolina committed
185
        dt = enddate - startdate
186

karolina's avatar
karolina committed
187
        if enddate > finaldate:  # do not run beyond finaldate
188
189
            enddate = finaldate

karolina's avatar
karolina committed
190
191
192
193
        self['time.start'] = startdate
        self['time.end'] = enddate
        self['time.finish'] = finaldate
        self['cyclelength'] = dt
194

karolina's avatar
karolina committed
195
196
197
198
199
200
201
        logging.info("===============================================================")
        logging.info("DA Cycle start date is %s" % startdate.strftime('%Y-%m-%d %H:%M'))
        logging.info("DA Cycle end   date is %s" % enddate.strftime('%Y-%m-%d %H:%M'))
        logging.info("DA Cycle final date is %s" % finaldate.strftime('%Y-%m-%d %H:%M'))  
        logging.info("DA Cycle cycle length is %s" % cyclelength)
        logging.info("DA Cycle restart is %s" % str(self['time.restart']))
        logging.info("===============================================================")
202
203


204
    def set_sample_times(self, lag):
205
206
207
208
209
210
211
        """
        Set the times over which a sampling interval will loop, depending on 
        the lag. Note that lag falls in the interval [0,nlag-1]
        """
        import copy

        # Start from cycle times 
karolina's avatar
karolina committed
212
213
        self['time.sample.start'] = copy.deepcopy(self['time.start'])
        self['time.sample.end'] = copy.deepcopy(self['time.end'])
214
215
216
217

        # Now advance depending on lag

        for l in range(lag):
218
            self.advance_sample_times()
219
220


221
    def advance_sample_times(self):
222
223
224
225
226
227
        """ 
        Advance sampling start and end time by one cycle interval
        """
        from da.tools.general import AdvanceTime

        startdate = self['time.sample.start']
karolina's avatar
karolina committed
228
229
        enddate = self['time.sample.end']                  
        cyclelength = self['cyclelength']                  
230

karolina's avatar
karolina committed
231
232
        startdate = AdvanceTime(startdate, cyclelength.days)
        enddate = AdvanceTime(enddate, cyclelength.days)
233

karolina's avatar
karolina committed
234
235
        self['time.sample.start'] = startdate
        self['time.sample.end'] = enddate
236
237
238
239
240
241
242
243
    

    def AdvanceCycleTimes(self):
        """ 
        Advance cycle start and end time by one cycle interval
        """
        from da.tools.general import AdvanceTime

karolina's avatar
karolina committed
244
245
246
        startdate = self['time.start']
        enddate = self['time.end']                  
        cyclelength = self['cyclelength']                  
247

karolina's avatar
karolina committed
248
249
        startdate = AdvanceTime(startdate, cyclelength.days)
        enddate = AdvanceTime(enddate, cyclelength.days)
250

karolina's avatar
karolina committed
251
252
        filtertime = startdate.strftime('%Y%m%d')
        self['dir.output'] = os.path.join(self['dir.da_run'], 'output', filtertime)
253

karolina's avatar
karolina committed
254
255
        self['time.start'] = startdate
        self['time.end'] = enddate
256
257


258
    def random_seed(self, action='read'):
Peters, Wouter's avatar
Peters, Wouter committed
259
260
        """ 
        Get the randomseed and save it, or read the random seed and set it. The seed is currently stored
261
        in a python :mod:`pickle` file, residing in the ``exec`` directory
Peters, Wouter's avatar
Peters, Wouter committed
262
263

        """
264
265
266
        import cPickle
        import numpy as np

karolina's avatar
karolina committed
267
        filename = os.path.join(self['dir.exec'], 'randomseed.pickle')
268
269

        if action == 'write':
karolina's avatar
karolina committed
270
271
            f = open(filename, 'wb')
            seed = np.random.get_state()
karolina's avatar
karolina committed
272
273
            cPickle.dump(seed, f, -1)
            f.close()
274

karolina's avatar
karolina committed
275
            msg = "Saved the random seed generator values to file"
276
277

        if action == 'read':
karolina's avatar
karolina committed
278
279
            f = open(filename, 'rb')
            seed = cPickle.load(f)
karolina's avatar
karolina committed
280
281
            np.random.set_state(seed)
            f.close()
282

karolina's avatar
karolina committed
283
            msg = "Retrieved the random seed generator values from file"
284
285

        logging.info(msg)
karolina's avatar
karolina committed
286
        self.RestartFileList.append(filename)
karolina's avatar
karolina committed
287
        logging.debug("Added the randomseed.pickle file to the RestartFileList")
288
289


290
    def Initialize(self):
291
292
293
        """ 
        This method determines how to proceed with the cycle. Three options are implemented:

Peters, Wouter's avatar
Peters, Wouter committed
294
295
296
            1. *Fresh start*  : set up the required file structure for this simulation and start
            2. *Restart*      : use latest da_runtime variables from the exec dir and restart
            3. *Recover*      : restart after crash by getting data from restart/one-ago folder
297
298
299

        The choice that gets executed depends on the presence of 

Peters, Wouter's avatar
Peters, Wouter committed
300
301
            # the ``-r`` option on the command line, this triggers a recover
            # the ``time.restart : True`` option in the da.rc file
302
303

        The latter is automatically set if the filter submits the next cycle at the end of the current one, 
304
        through method :meth:`~da.tools.initexit.CycleControl.submit_next_cycle`.
Peters, Wouter's avatar
Peters, Wouter committed
305

Peters, Wouter's avatar
Peters, Wouter committed
306
        The specific call tree under each scenario is: 
Peters, Wouter's avatar
Peters, Wouter committed
307

Peters, Wouter's avatar
Peters, Wouter committed
308
            1. *Fresh Start*
309
                *  :meth:`~da.tools.initexit.CycleControl.setup_file_structure()`  <- Create directory tree
Peters, Wouter's avatar
Peters, Wouter committed
310
            2. *Restart*
311
312
                *  :meth:`~da.tools.initexit.CycleControl.setup_file_structure()`
                *  :meth:`~da.tools.initexit.CycleControl.random_seed`    <- Read the random seed from file
Peters, Wouter's avatar
Peters, Wouter committed
313
            3. *Recover*
314
315
316
                *  :meth:`~da.tools.initexit.CycleControl.setup_file_structure()`
                *  :meth:`~da.tools.initexit.CycleControl.recover_run()`          <- Recover files from restart/one-ago dir, reset ``time.start``
                *  :meth:`~da.tools.initexit.CycleControl.random_seed` 
Peters, Wouter's avatar
Peters, Wouter committed
317
318
319

        And is always followed by a call to

Peters, Wouter's avatar
Peters, Wouter committed
320
321
            * ParseTimes()
            * WriteRc('jobfilename')
322
323
324
325
326
        """        
#
# case 1: A recover from a previous crash, this is signaled by flag "-r"
#
        if self['da.crash.recover']:
karolina's avatar
karolina committed
327
            logging.info("Recovering simulation from data in: %s" % self['dir.da_run'])
328
329
330
            self.setup_file_structure()
            self.recover_run()
            self.random_seed('read')
331
332
333
334
#
# case 2: A continuation, this is signaled by rc-item time.restart = True
#
        elif self['time.restart']:
karolina's avatar
karolina committed
335
            logging.info("Restarting filter from previous step")
336
            self.setup_file_structure()
337
            strippedname = os.path.split(self['jobrcfilename'])[-1]
karolina's avatar
karolina committed
338
            self['jobrcfilename'] = os.path.join(self['dir.exec'], strippedname)
339
            self.random_seed('read')
340
341
342
343
#
# case 3: A fresh start, this is signaled by rc-item time.restart = False
#
        elif not self['time.restart']:
karolina's avatar
karolina committed
344
            logging.info("First time step in filter sequence")
345
            self.setup_file_structure()
346
347

            # expand jobrcfilename to include exec dir from now on.
348
349
350
            # First strip current leading path from filename

            strippedname = os.path.split(self['jobrcfilename'])[-1]
karolina's avatar
karolina committed
351
            self['jobrcfilename'] = os.path.join(self['dir.exec'], strippedname)
352
353
354
355

        self.ParseTimes()
        self.WriteRC(self['jobrcfilename'])

356
    def setup_file_structure(self):
357
358
        """ 
        Create file structure needed for data assimilation system.
Peters, Wouter's avatar
Peters, Wouter committed
359
360
361
362
363
364
365
366
367
        In principle this looks like:

            * ``${da_rundir}``
            * ``${da_rundir}/input``
            * ``${da_rundir}/output``
            * ``${da_rundir}/exec``
            * ``${da_rundir}/diagnostics``
            * ``${da_rundir}/analysis``
            * ``${da_rundir}/jobs``
368
369
            * ``${da_rundir}/restart/current``
            * ``${da_rundir}/restart/one-ago``
Peters, Wouter's avatar
Peters, Wouter committed
370

371
        .. note:: The exec dir will actually be a simlink to the directory where
372
373
374
                 the observation operator executable lives. This directory is passed through
                 the ``da.rc`` file. 

375
        .. note:: The observation input files will be placed in the exec dir,
376
                 and the resulting simulated values will be retrieved from there as well.
Peters, Wouter's avatar
Peters, Wouter committed
377

378
379
380
381
382
        """
        from da.tools.general import CreateDirs

# Create the run directory for this DA job, including I/O structure

karolina's avatar
karolina committed
383
        filtertime = self['time.start'].strftime('%Y%m%d')
384

karolina's avatar
karolina committed
385
386
387
388
389
390
391
392
393
        self['dir.exec'] = os.path.join(self['dir.da_run'], 'exec')
        self['dir.input'] = os.path.join(self['dir.da_run'], 'input')
        self['dir.output'] = os.path.join(self['dir.da_run'], 'output', filtertime)
        self['dir.diagnostics'] = os.path.join(self['dir.da_run'], 'diagnostics')
        self['dir.analysis'] = os.path.join(self['dir.da_run'], 'analysis')
        self['dir.jobs'] = os.path.join(self['dir.da_run'], 'jobs')
        self['dir.restart'] = os.path.join(self['dir.da_run'], 'restart')
        self['dir.restart.current'] = os.path.join(self['dir.restart'], 'current')
        self['dir.restart.oneago'] = os.path.join(self['dir.restart'], 'one-ago')
394
395
396
397
398
399
400
401

        CreateDirs(self['dir.da_run'])
        CreateDirs(os.path.join(self['dir.exec']))
        CreateDirs(os.path.join(self['dir.input']))
        CreateDirs(os.path.join(self['dir.output']))
        CreateDirs(os.path.join(self['dir.diagnostics']))
        CreateDirs(os.path.join(self['dir.analysis']))
        CreateDirs(os.path.join(self['dir.jobs']))
402
403
404
        CreateDirs(os.path.join(self['dir.restart']))
        CreateDirs(os.path.join(self['dir.restart.current']))
        CreateDirs(os.path.join(self['dir.restart.oneago']))
405

karolina's avatar
karolina committed
406
        logging.info('Succesfully created the file structure for the assimilation job')
407

408
    def recover_run(self):
Peters, Wouter's avatar
Peters, Wouter committed
409
410
        """
        Prepare a recovery from a crashed run. This consists of: 
411
        
412
            - copying all data from the restart/one-ago folder (:meth:`~da.tools.initexit.CycleControl.move_restart_data`),
413
            - replacing all ``rc-file`` items with those from the ``da_runtime.rc`` in the restart/current dir 
414
            - resetting the seed of the random number generator to the value it had before the crash (:meth:`~da.tools.initexit.CycleControl.random_seed`)
415
            - replacing the output dir name, since it has the sample time in it...
Peters, Wouter's avatar
Peters, Wouter committed
416
            
417
        """
418
        import da.tools.rc as rc
419
        from da.tools.general import CreateDirs
420

421
        # Replace rc-items with those from the crashed run's last rc-file (now in restart.current dir)
422

karolina's avatar
karolina committed
423
424
        file_rc_rec = os.path.join(self['dir.restart.current'], 'da_runtime.rc')
        rc_rec = rc.read(file_rc_rec)
425

karolina's avatar
karolina committed
426
        for k, v in rc_rec.iteritems():
427
428
429
430
            self[k] = v

        self.ValidateRC()

karolina's avatar
karolina committed
431
432
        logging.debug("Replaced rc-items.... ")
        logging.debug("Next cycle start date is %s" % self['time.start'])
433

434
        # Copy randomseed.pickle file to exec dir
karolina's avatar
karolina committed
435
436
437
        source = os.path.join(self['dir.restart.current'], 'randomseed.pickle')
        dest = os.path.join(self['dir.exec'], 'randomseed.pickle')
        shutil.copy(source, dest)
438

karolina's avatar
karolina committed
439
        logging.debug("Replaced randomseed file with previous cycles' last values")
440
441
        # Re-create the output dir for this time step, if needed

karolina's avatar
karolina committed
442
443
444
        filtertime = self['time.start'].strftime('%Y%m%d')
        self['dir.output'] = os.path.join(self['dir.da_run'], 'output', filtertime)
        CreateDirs(os.path.join(self['dir.output']))
445
446
447

    def Finalize(self):
        """
448
449
450
451
        Finalize the da cycle, this means writing the save data and rc-files for the next run. 
        The following sequence of actions occur:

            * Write the randomseed to file for reuse in next cycle
Peters, Wouter's avatar
Peters, Wouter committed
452
            * Write a new ``rc-file`` with ``time.restart : True``, and new ``time.start`` and ``time.end``
453
454
455
456
            * Collect all needed data needed for check-pointing (restart from current system state)
            * Move the previous check pointing data out of the way, and replace with current
            * Submit the next cycle

457
        """
458
459
460
461
462
463
464
465
        self.random_seed('write')
        self.write_new_rc_file()
        self.move_restart_data(io_option='store')  # Move restart data from current to one-ago
        self.collect_restart_data()  # Collect restart data for next cycle into a clean restart/current folder
        self.collect_output()  # Collect restart data for next cycle into a clean restart/current folder
        self.submit_next_cycle()

    def collect_output(self):
466
        """ Collect files that are part of the requested output for this cycle. This function allows users to add files 
467
468
469
470
471
472
473
474
475
            to a list, and then the system will copy these to the current cycle's output directory.
            The list of files included is read from the 
            attribute "OutputFileList" which is a simple list of files that can be appended by other objects/methods that
            require output data to be saved.


        """
        from da.tools.general import CreateDirs

karolina's avatar
karolina committed
476
477
        targetdir = os.path.join(self['dir.output'])
        CreateDirs(os.path.join(targetdir))
478

karolina's avatar
karolina committed
479
480
        logging.info("Collecting the required output data") 
        logging.debug("           to   directory: %s " % targetdir)
481
482
483
484

        for file in set(self.OutputFileList):
            if os.path.isdir(file): # skip dirs
                continue
485
            if not os.path.exists(file): # skip dirs
karolina's avatar
karolina committed
486
                logging.warning("           [not found] .... %s " % file)
487
                continue
488

karolina's avatar
karolina committed
489
490
            logging.debug("           [copy] .... %s " % file)
            shutil.copy(file, file.replace(os.path.split(file)[0], targetdir))
491
492
493



494
    def collect_restart_data(self):
495
        """ Collect files needed for the restart of this cycle in case of a crash, or for the continuation of the next cycle. 
496
            All files needed are written to the restart/current directory. The list of files included is read from the 
497
498
499
            attribute "RestartFileList" which is a simple list of files that can be appended by other objects/methods that
            require restart data to be saved.

500
            .. note:: Before collecting the files in the ``RestartFileList``, the restart/current directory will be emptied and
501
502
503
504
505
                     recreated. This prevents files from accumulating in the restart/current and restart/one-ago folders. It 
                     also means that if a file is missing from the ``RestartFileList``, it will not be available for check-pointing
                     if your run crashes or dies!

            Currently, the following files are included:
506

507
                * The ``da_runtime.rc`` file
508
                * The ``randomseed.pickle`` file
509
510
                * The savestate.nc file
                * The files in the ``ObservationOperator.RestartFileList``, i.e., restart data for the transport model
511

512

Peters, Wouter's avatar
Peters, Wouter committed
513
            .. note:: We assume that the restart files for the :ref:`ObservationOperator` 
514
                      reside in a separate folder, i.e, the ObservationOperator does *not* write directly to the CTDAS restart dir!
515
516

        """
517
        from da.tools.general import CreateDirs
518

karolina's avatar
karolina committed
519
        targetdir = os.path.join(self['dir.restart.current'])
520

karolina's avatar
karolina committed
521
        logging.info("Purging the current restart directory before collecting new data")
522

karolina's avatar
karolina committed
523
        CreateDirs(os.path.join(targetdir), forceclean=True)
524

karolina's avatar
karolina committed
525
526
        logging.info("Collecting the required restart data")
        logging.debug("           to   directory: %s " % targetdir)
527

528
        for file in set(self.RestartFileList):
529
530
            if os.path.isdir(file): # skip dirs
                continue
531
            if not os.path.exists(file): # skip dirs
karolina's avatar
karolina committed
532
                logging.warning("           [not found] .... %s " % file)
533
            else:
karolina's avatar
karolina committed
534
535
                logging.debug("           [copy] .... %s " % file)
                shutil.copy(file, file.replace(os.path.split(file)[0], targetdir))
536
537


538
    def move_restart_data(self, io_option='restore'):
Peters, Wouter's avatar
Peters, Wouter committed
539
        """ 
540
        Store or restore model state to/from a restart directory. 
541

542
            Two IO options are available:
543

544
545
            (1) io_option = restore    : Get data from restart.oneago directory
            (2) io_option = store      : Save data to restart.oneago directory
546

547
            In case of a 'store' command the restart.oneago folder is re-created so that the contents are empty to begin with.
548
549
550
551

        """
        from da.tools.general import CreateDirs

karolina's avatar
karolina committed
552
553
        if io_option not in ['store', 'restore']:
            raise ValueError, 'Invalid option specified for io_option (%s)' % io_option
554

555
556
557
558
559
560
        if io_option == 'store':
            targetdir = self['dir.restart.oneago']
            sourcedir = self['dir.restart.current']
        elif io_option == 'restore':
            sourcedir = self['dir.restart.oneago']
            targetdir = self['dir.restart.current']
561
562
563
564

# If "store" is requested, recreate target dir, cleaning the contents 

        if io_option == 'store':
karolina's avatar
karolina committed
565
            CreateDirs(os.path.join(targetdir), forceclean=True)
566

karolina's avatar
karolina committed
567
568
569
        logging.debug("Performing a %s of data" % (io_option))
        logging.debug("           from directory: %s " % sourcedir)
        logging.debug("           to   directory: %s " % targetdir)
570
571
572


        for file in os.listdir(sourcedir):
karolina's avatar
karolina committed
573
            file = os.path.join(sourcedir, file)
574
            if not os.path.exists(file):
karolina's avatar
karolina committed
575
                logging.debug("Cannot find requested file to move: %s " % file)  
576
577
                sys.exit(2)

578
            if os.path.isdir(file): # skip dirs
karolina's avatar
karolina committed
579
                logging.debug("           [skip] .... %s " % file)
580
                continue    
581
            else:
karolina's avatar
karolina committed
582
583
                logging.debug("           [copy] .... %s " % file)
                shutil.copy(file, file.replace(sourcedir, targetdir))
584
585

#
586
    def write_new_rc_file(self):
587
588
589
590
591
592
        """ Write the rc-file for the next DA cycle. 

            .. note:: The start time for the next cycle is the end time of this one, while 
                      the end time for the next cycle is the current end time + one cycle length. 
                      
            The resulting rc-file is written to the ``dir.exec`` so that it can be used when resubmitting the next cycle
593
594
            
        """
595
        import da.tools.rc as rc
596
        import copy
597

598
        # We make a copy of the current DaCycle object, and modify the start + end dates and restart value
599

karolina's avatar
karolina committed
600
601
602
        newDaCycle = copy.deepcopy(self)
        newDaCycle.AdvanceCycleTimes()
        newDaCycle['time.restart'] = True
603

604
        # Create the name of the rc-file that will hold this new input, and write it
605

karolina's avatar
karolina committed
606
607
608
        fname = os.path.join(self['dir.exec'], 'da_runtime.rc')  # current exec dir holds next rc file
        rc.write(fname, newDaCycle)
        logging.debug('Wrote new da_runtime.rc (%s) to exec dir' % fname)
609

610
        # The rest is info needed for a system restart, so it modifies the current DaCycle object (self)
611

karolina's avatar
karolina committed
612
613
        self['da.restart.fname'] = fname    # needed for next job template
        self.RestartFileList.extend([fname])  # current restart list holds next rc file name
614

karolina's avatar
karolina committed
615
        logging.debug('Added da_runtime.rc to the RestartFileList for later collection')
616
617


karolina's avatar
karolina committed
618
    def WriteRC(self, fname):
619
        """ Write RC file after each process to reflect updated info """
620
        import da.tools.rc as rc
621

karolina's avatar
karolina committed
622
623
624
        rc.write(fname, self)
        logging.debug('Wrote expanded rc-file (%s)' % (fname))
        
625

626
    def submit_next_cycle(self):
627
        """ 
Peters, Wouter's avatar
Peters, Wouter committed
628
        Submit the next job of a DA cycle, this consists of 
629
630
631
            * Changing to the working directory from which the job was started initially
            * create a line to start the master script again with a newly created rc-file
            * Submitting the jobfile 
Peters, Wouter's avatar
Peters, Wouter committed
632
633
634

        If the end of the cycle series is reached, no new job is submitted.

635
        """
636
        from string import join
637
638
639

        DaPlatForm = self.DaPlatForm

640
        if self['time.end'] < self['time.finish']:
641

Peters, Wouter's avatar
test    
Peters, Wouter committed
642
            # file ID and names
karolina's avatar
karolina committed
643
644
645
646
            jobid = self['time.end'].strftime('%Y%m%d') 
            targetdir = os.path.join(self['dir.exec'])
            jobfile = os.path.join(targetdir, 'jb.%s.jb' % jobid)
            logfile = jobfile.replace('.jb', '.log')
Peters, Wouter's avatar
test    
Peters, Wouter committed
647
648

            # Template and commands for job
649
            jobparams = {'jobname':"j.%s" % jobid, 'jobtime':'24:00:00', 'logfile':logfile, 'errfile':logfile}
650
            template = DaPlatForm.get_job_template(jobparams)
karolina's avatar
karolina committed
651
652
            execcommand = os.path.join(self['dir.da_submit'], sys.argv[0]) 
            template += 'python %s rc=%s %s' % (execcommand, self['da.restart.fname'], join(self.opts, ''),) 
Peters, Wouter's avatar
test    
Peters, Wouter committed
653
654

            # write and submit 
655
656
            DaPlatForm.write_job(jobfile, template, jobid)
            jobid = DaPlatForm.submit_job(jobfile, joblog=logfile) 
657
658
659
660
        else:
            logging.info('Final date reached, no new cycle started')


661
def StartLogger(level=logging.INFO):
662
663
664
665
    """ start the logging of messages to screen"""

# start the logging basic configuration by setting up a log file

karolina's avatar
karolina committed
666
667
668
    logging.basicConfig(level=level,
                        format=' [%(levelname)-7s] (%(asctime)s) py-%(module)-20s : %(message)s',
                        datefmt='%Y-%m-%d %H:%M:%S')
669
670

def ParseOptions():
671
672
673
674
675
676
677
678
679
680
681
682
683
    """ 
    Function parses options from the command line and returns the arguments as a dictionary.
    Accepted command line arguments are:

    ========  =======
    Argument  Meaning
    ========  =======
    -v        verbose output in log files
    -h        display help
    -r        start a simulation by recovering from a previous crash
    ========  =======

    """
684
685
686
687
    import getopt

# Parse keywords, the only option accepted so far is the "-h" flag for help

karolina's avatar
karolina committed
688
689
    opts = []
    args = []
690
691
692
    try:                                
        opts, args = getopt.gnu_getopt(sys.argv[1:], "-hrv")
    except getopt.GetoptError, msg:           
karolina's avatar
karolina committed
693
        logging.error('%s' % msg)
694
695
696
        sys.exit(2)      

    for options in opts:
karolina's avatar
karolina committed
697
        options = options[0].lower()
698
699
700
701
702
703
704
705
        if options == '-h':
            print ""
            print helptext
            sys.exit(2)      
        if options == '-r':
            logging.info('-r flag specified on command line: recovering from crash')
        if options == '-v':
            logging.info('-v flag specified on command line: extra verbose output')
karolina's avatar
karolina committed
706
            logging.root.setLevel(logging.DEBUG)
707

708
    if opts: 
karolina's avatar
karolina committed
709
        optslist = [item[0] for item in opts]
710
    else:
karolina's avatar
karolina committed
711
        optslist = []
712

713
714
# Parse arguments and return as dictionary

karolina's avatar
karolina committed
715
    arguments = {}
716
717
718
719
720
721
722
723
    for item in args:
        #item=item.lower()

# Catch arguments that are passed not in "key=value" format

        if '=' in item:
            key, arg = item.split('=')
        else:
karolina's avatar
karolina committed
724
725
            logging.error('%s' % 'Argument passed without description (%s)' % item)
            raise getopt.GetoptError, arg
726

karolina's avatar
karolina committed
727
        arguments[key] = arg
728
729


730
    return optslist, arguments
731

karolina's avatar
karolina committed
732
def ValidateOptsArgs(opts, args):
733
734
735
736
737
    """ 
 Validate the options and arguments passed from the command line before starting the cycle. The validation consists of checking for the presence of an argument "rc", and the existence of
 the specified rc-file.  
 
    """
738
    if not args.has_key("rc"):
karolina's avatar
karolina committed
739
740
741
        msg = "There is no rc-file specified on the command line. Please use rc=yourfile.rc"
        logging.error(msg)
        raise IOError, msg
742
    elif not os.path.exists(args['rc']):
karolina's avatar
karolina committed
743
744
745
        msg = "The specified rc-file (%s) does not exist " % args['rc'] 
        logging.error(msg)
        raise IOError, msg
746
747
748
749
750
751
752
753
754

    # WP not needed anymore
    #if not args.has_key('process'):
    #    msg = "There is no process specified on the command line, assuming process=Start"   ; logging.info(msg)
    #    args['process'] = 'start'
    #if args['process'].lower() not in validprocesses:
    #    msg = "The specified process (%s) is not valid"%args['process']   ; logging.error(msg)
    #    raise IOError,msg

karolina's avatar
karolina committed
755
    return opts, args
756
757
758


if __name__ == "__main__":
759
760

    sys.path.append('../../')
karolina's avatar
karolina committed
761
    opts, args = ParseOptions()
762
763
    print opts
    print args
764