initexit.py 30.9 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
272
273
            f = open(filename, 'wb')
            seed = np.random.get_state()
            dummy = cPickle.dump (seed, f, -1)
            dummy = 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
280
281
            f = open(filename, 'rb')
            seed = cPickle.load(f)
            dummy = np.random.set_state(seed)
            dummy = 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
287
        self.RestartFileList.extend([filename])
        logging.debug("Added the randomseed.pickle file to the RestartFileList")
288
289


290
291
292
        return None

    def Initialize(self):
293
294
295
        """ 
        This method determines how to proceed with the cycle. Three options are implemented:

Peters, Wouter's avatar
Peters, Wouter committed
296
297
298
            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
299
300
301

        The choice that gets executed depends on the presence of 

Peters, Wouter's avatar
Peters, Wouter committed
302
303
            # the ``-r`` option on the command line, this triggers a recover
            # the ``time.restart : True`` option in the da.rc file
304
305

        The latter is automatically set if the filter submits the next cycle at the end of the current one, 
306
        through method :meth:`~da.tools.initexit.CycleControl.submit_next_cycle`.
Peters, Wouter's avatar
Peters, Wouter committed
307

Peters, Wouter's avatar
Peters, Wouter committed
308
        The specific call tree under each scenario is: 
Peters, Wouter's avatar
Peters, Wouter committed
309

Peters, Wouter's avatar
Peters, Wouter committed
310
            1. *Fresh Start*
311
                *  :meth:`~da.tools.initexit.CycleControl.setup_file_structure()`  <- Create directory tree
Peters, Wouter's avatar
Peters, Wouter committed
312
            2. *Restart*
313
314
                *  :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
315
            3. *Recover*
316
317
318
                *  :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
319
320
321

        And is always followed by a call to

Peters, Wouter's avatar
Peters, Wouter committed
322
323
            * ParseTimes()
            * WriteRc('jobfilename')
324
325
326
327
328
        """        
#
# case 1: A recover from a previous crash, this is signaled by flag "-r"
#
        if self['da.crash.recover']:
karolina's avatar
karolina committed
329
            logging.info("Recovering simulation from data in: %s" % self['dir.da_run'])
330
331
332
            self.setup_file_structure()
            self.recover_run()
            self.random_seed('read')
333
334
335
336
#
# case 2: A continuation, this is signaled by rc-item time.restart = True
#
        elif self['time.restart']:
karolina's avatar
karolina committed
337
            logging.info("Restarting filter from previous step")
338
            self.setup_file_structure()
339
            strippedname = os.path.split(self['jobrcfilename'])[-1]
karolina's avatar
karolina committed
340
            self['jobrcfilename'] = os.path.join(self['dir.exec'], strippedname)
341
            self.random_seed('read')
342
343
344
345
#
# case 3: A fresh start, this is signaled by rc-item time.restart = False
#
        elif not self['time.restart']:
karolina's avatar
karolina committed
346
            logging.info("First time step in filter sequence")
347
            self.setup_file_structure()
348
349

            # expand jobrcfilename to include exec dir from now on.
350
351
352
            # First strip current leading path from filename

            strippedname = os.path.split(self['jobrcfilename'])[-1]
karolina's avatar
karolina committed
353
            self['jobrcfilename'] = os.path.join(self['dir.exec'], strippedname)
354
355
356
357

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

358
    def setup_file_structure(self):
359
360
        """ 
        Create file structure needed for data assimilation system.
Peters, Wouter's avatar
Peters, Wouter committed
361
362
363
364
365
366
367
368
369
        In principle this looks like:

            * ``${da_rundir}``
            * ``${da_rundir}/input``
            * ``${da_rundir}/output``
            * ``${da_rundir}/exec``
            * ``${da_rundir}/diagnostics``
            * ``${da_rundir}/analysis``
            * ``${da_rundir}/jobs``
370
371
            * ``${da_rundir}/restart/current``
            * ``${da_rundir}/restart/one-ago``
Peters, Wouter's avatar
Peters, Wouter committed
372

373
        .. note:: The exec dir will actually be a simlink to the directory where
374
375
376
                 the observation operator executable lives. This directory is passed through
                 the ``da.rc`` file. 

377
        .. note:: The observation input files will be placed in the exec dir,
378
                 and the resulting simulated values will be retrieved from there as well.
Peters, Wouter's avatar
Peters, Wouter committed
379

380
381
382
383
384
        """
        from da.tools.general import CreateDirs

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

karolina's avatar
karolina committed
385
        filtertime = self['time.start'].strftime('%Y%m%d')
386

karolina's avatar
karolina committed
387
388
389
390
391
392
393
394
395
        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')
396
397
398
399
400
401
402
403

        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']))
404
405
406
        CreateDirs(os.path.join(self['dir.restart']))
        CreateDirs(os.path.join(self['dir.restart.current']))
        CreateDirs(os.path.join(self['dir.restart.oneago']))
407

karolina's avatar
karolina committed
408
        logging.info('Succesfully created the file structure for the assimilation job')
409

410
    def recover_run(self):
Peters, Wouter's avatar
Peters, Wouter committed
411
412
        """
        Prepare a recovery from a crashed run. This consists of: 
413
        
414
            - copying all data from the restart/one-ago folder (:meth:`~da.tools.initexit.CycleControl.move_restart_data`),
415
            - replacing all ``rc-file`` items with those from the ``da_runtime.rc`` in the restart/current dir 
416
            - resetting the seed of the random number generator to the value it had before the crash (:meth:`~da.tools.initexit.CycleControl.random_seed`)
417
            - replacing the output dir name, since it has the sample time in it...
Peters, Wouter's avatar
Peters, Wouter committed
418
            
419
        """
420
        import da.tools.rc as rc
421
        from da.tools.general import CreateDirs
422

423
        # Replace rc-items with those from the crashed run's last rc-file (now in restart.current dir)
424

karolina's avatar
karolina committed
425
426
        file_rc_rec = os.path.join(self['dir.restart.current'], 'da_runtime.rc')
        rc_rec = rc.read(file_rc_rec)
427

karolina's avatar
karolina committed
428
        for k, v in rc_rec.iteritems():
429
430
431
432
            self[k] = v

        self.ValidateRC()

karolina's avatar
karolina committed
433
434
        logging.debug("Replaced rc-items.... ")
        logging.debug("Next cycle start date is %s" % self['time.start'])
435

436
        # Copy randomseed.pickle file to exec dir
karolina's avatar
karolina committed
437
438
439
        source = os.path.join(self['dir.restart.current'], 'randomseed.pickle')
        dest = os.path.join(self['dir.exec'], 'randomseed.pickle')
        shutil.copy(source, dest)
440

karolina's avatar
karolina committed
441
        logging.debug("Replaced randomseed file with previous cycles' last values")
442
443
        # Re-create the output dir for this time step, if needed

karolina's avatar
karolina committed
444
445
446
        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']))
447
448
449

    def Finalize(self):
        """
450
451
452
453
        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
454
            * Write a new ``rc-file`` with ``time.restart : True``, and new ``time.start`` and ``time.end``
455
456
457
458
            * 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

459
        """
460
461
462
463
464
465
466
467
        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):
468
        """ Collect files that are part of the requested output for this cycle. This function allows users to add files 
469
470
471
472
473
474
475
476
477
            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
478
479
        targetdir = os.path.join(self['dir.output'])
        CreateDirs(os.path.join(targetdir))
480

karolina's avatar
karolina committed
481
482
        logging.info("Collecting the required output data") 
        logging.debug("           to   directory: %s " % targetdir)
483
484
485
486

        for file in set(self.OutputFileList):
            if os.path.isdir(file): # skip dirs
                continue
487
            if not os.path.exists(file): # skip dirs
karolina's avatar
karolina committed
488
                logging.warning("           [not found] .... %s " % file)
489
                continue
490

karolina's avatar
karolina committed
491
492
            logging.debug("           [copy] .... %s " % file)
            shutil.copy(file, file.replace(os.path.split(file)[0], targetdir))
493
494
495



496
    def collect_restart_data(self):
497
        """ Collect files needed for the restart of this cycle in case of a crash, or for the continuation of the next cycle. 
498
            All files needed are written to the restart/current directory. The list of files included is read from the 
499
500
501
            attribute "RestartFileList" which is a simple list of files that can be appended by other objects/methods that
            require restart data to be saved.

502
            .. note:: Before collecting the files in the ``RestartFileList``, the restart/current directory will be emptied and
503
504
505
506
507
                     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:
508

509
                * The ``da_runtime.rc`` file
510
                * The ``randomseed.pickle`` file
511
512
                * The savestate.nc file
                * The files in the ``ObservationOperator.RestartFileList``, i.e., restart data for the transport model
513

514

Peters, Wouter's avatar
Peters, Wouter committed
515
            .. note:: We assume that the restart files for the :ref:`ObservationOperator` 
516
                      reside in a separate folder, i.e, the ObservationOperator does *not* write directly to the CTDAS restart dir!
517
518

        """
519
        from da.tools.general import CreateDirs
520

karolina's avatar
karolina committed
521
        targetdir = os.path.join(self['dir.restart.current'])
522

karolina's avatar
karolina committed
523
        logging.info("Purging the current restart directory before collecting new data")
524

karolina's avatar
karolina committed
525
        CreateDirs(os.path.join(targetdir), forceclean=True)
526

karolina's avatar
karolina committed
527
528
        logging.info("Collecting the required restart data")
        logging.debug("           to   directory: %s " % targetdir)
529

530
        for file in set(self.RestartFileList):
531
532
            if os.path.isdir(file): # skip dirs
                continue
533
            if not os.path.exists(file): # skip dirs
karolina's avatar
karolina committed
534
                logging.warning("           [not found] .... %s " % file)
535
            else:
karolina's avatar
karolina committed
536
537
                logging.debug("           [copy] .... %s " % file)
                shutil.copy(file, file.replace(os.path.split(file)[0], targetdir))
538
539


540
    def move_restart_data(self, io_option='restore'):
Peters, Wouter's avatar
Peters, Wouter committed
541
        """ 
542
        Store or restore model state to/from a restart directory. 
543

544
            Two IO options are available:
545

546
547
            (1) io_option = restore    : Get data from restart.oneago directory
            (2) io_option = store      : Save data to restart.oneago directory
548

549
            In case of a 'store' command the restart.oneago folder is re-created so that the contents are empty to begin with.
550
551
552
553

        """
        from da.tools.general import CreateDirs

karolina's avatar
karolina committed
554
555
        if io_option not in ['store', 'restore']:
            raise ValueError, 'Invalid option specified for io_option (%s)' % io_option
556

557
558
559
560
561
562
        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']
563
564
565
566

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

        if io_option == 'store':
karolina's avatar
karolina committed
567
            CreateDirs(os.path.join(targetdir), forceclean=True)
568

karolina's avatar
karolina committed
569
570
571
        logging.debug("Performing a %s of data" % (io_option))
        logging.debug("           from directory: %s " % sourcedir)
        logging.debug("           to   directory: %s " % targetdir)
572
573
574


        for file in os.listdir(sourcedir):
karolina's avatar
karolina committed
575
            file = os.path.join(sourcedir, file)
576
            if not os.path.exists(file):
karolina's avatar
karolina committed
577
                logging.debug("Cannot find requested file to move: %s " % file)  
578
579
                sys.exit(2)

580
            if os.path.isdir(file): # skip dirs
karolina's avatar
karolina committed
581
                logging.debug("           [skip] .... %s " % file)
582
                continue    
583
            else:
karolina's avatar
karolina committed
584
585
                logging.debug("           [copy] .... %s " % file)
                shutil.copy(file, file.replace(sourcedir, targetdir))
586
587

#
588
    def write_new_rc_file(self):
589
590
591
592
593
594
        """ 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
595
596
            
        """
597
        import da.tools.rc as rc
598
        import copy
599

600
        # We make a copy of the current DaCycle object, and modify the start + end dates and restart value
601

karolina's avatar
karolina committed
602
603
604
        newDaCycle = copy.deepcopy(self)
        newDaCycle.AdvanceCycleTimes()
        newDaCycle['time.restart'] = True
605

606
        # Create the name of the rc-file that will hold this new input, and write it
607

karolina's avatar
karolina committed
608
609
610
        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)
611

612
        # The rest is info needed for a system restart, so it modifies the current DaCycle object (self)
613

karolina's avatar
karolina committed
614
615
        self['da.restart.fname'] = fname    # needed for next job template
        self.RestartFileList.extend([fname])  # current restart list holds next rc file name
616

karolina's avatar
karolina committed
617
        logging.debug('Added da_runtime.rc to the RestartFileList for later collection')
618
619


karolina's avatar
karolina committed
620
    def WriteRC(self, fname):
621
        """ Write RC file after each process to reflect updated info """
622
        import da.tools.rc as rc
623

karolina's avatar
karolina committed
624
625
626
        rc.write(fname, self)
        logging.debug('Wrote expanded rc-file (%s)' % (fname))
        
627

628
    def submit_next_cycle(self):
629
        """ 
Peters, Wouter's avatar
Peters, Wouter committed
630
        Submit the next job of a DA cycle, this consists of 
631
632
633
            * 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
634
635
636

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

637
        """
638
        from string import join
639
640
641

        DaPlatForm = self.DaPlatForm

642
        if self['time.end'] < self['time.finish']:
643

Peters, Wouter's avatar
test    
Peters, Wouter committed
644
            # file ID and names
karolina's avatar
karolina committed
645
646
647
648
            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
649
650

            # Template and commands for job
karolina's avatar
karolina committed
651
            jobparams = {'jobname':"j.%s" % jobid, 'jobtime':'06:00:00', 'logfile':logfile, 'errfile':logfile}
652
            template = DaPlatForm.get_job_template(jobparams)
karolina's avatar
karolina committed
653
654
            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
655
656

            # write and submit 
657
658
            DaPlatForm.write_job(jobfile, template, jobid)
            jobid = DaPlatForm.submit_job(jobfile, joblog=logfile) 
659
660
661
662
        else:
            logging.info('Final date reached, no new cycle started')


karolina's avatar
karolina committed
663
    def SubmitSubStep(self, stepname):
Peters, Wouter's avatar
Peters, Wouter committed
664
665
        """ 
        Submit the next substep of a DA cycle, this consists of 
666
            * getting a job template as returned by :meth:`~da.tools.baseclasses.platform.get_job_template`
Peters, Wouter's avatar
Peters, Wouter committed
667
            * adding the lines needed to start a next run with a newly created rc-file
668
669
            * Writing the jobfile as done by :meth:`~da.tools.baseclasses.platform.write_job`
            * Submitting the jobfile as done by :meth:`~da.tools.baseclasses.platform.write_job`
Peters, Wouter's avatar
Peters, Wouter committed
670
671

        """
672
        from string import join
karolina's avatar
karolina committed
673
        
674
675
        DaPlatForm = self.DaPlatForm

karolina's avatar
karolina committed
676
        jobparams = {'jobname':'das.%s' % stepname}
677
        template = DaPlatForm.get_job_template(jobparams)
karolina's avatar
karolina committed
678
679
        template += 'cd %s\n' % os.getcwd()
        template += '%s rc=%s process=%s %s' % (sys.argv[0], self['jobrcfilename'], stepname, join(self.opts, ''),) 
680
681
682
        jobfile = DaPlatForm.write_job(self, template, stepname)
        DaPlatForm.submit_job(jobfile) 

683

684
685


686

687
def StartLogger(level=logging.INFO):
688
689
690
691
    """ start the logging of messages to screen"""

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

karolina's avatar
karolina committed
692
693
694
    logging.basicConfig(level=level,
                        format=' [%(levelname)-7s] (%(asctime)s) py-%(module)-20s : %(message)s',
                        datefmt='%Y-%m-%d %H:%M:%S')
695
696

def ParseOptions():
697
698
699
700
701
702
703
704
705
706
707
708
709
    """ 
    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
    ========  =======

    """
710
711
712
713
    import getopt

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

karolina's avatar
karolina committed
714
715
    opts = []
    args = []
716
717
718
    try:                                
        opts, args = getopt.gnu_getopt(sys.argv[1:], "-hrv")
    except getopt.GetoptError, msg:           
karolina's avatar
karolina committed
719
        logging.error('%s' % msg)
720
721
722
        sys.exit(2)      

    for options in opts:
karolina's avatar
karolina committed
723
        options = options[0].lower()
724
725
726
727
728
729
730
731
        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
732
            logging.root.setLevel(logging.DEBUG)
733

734
    if opts: 
karolina's avatar
karolina committed
735
        optslist = [item[0] for item in opts]
736
    else:
karolina's avatar
karolina committed
737
        optslist = []
738

739
740
# Parse arguments and return as dictionary

karolina's avatar
karolina committed
741
    arguments = {}
742
743
744
745
746
747
748
749
    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
750
751
            logging.error('%s' % 'Argument passed without description (%s)' % item)
            raise getopt.GetoptError, arg
752

karolina's avatar
karolina committed
753
        arguments[key] = arg
754
755


756
    return optslist, arguments
757

karolina's avatar
karolina committed
758
def ValidateOptsArgs(opts, args):
759
760
761
762
763
    """ 
 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.  
 
    """
764
    if not args.has_key("rc"):
karolina's avatar
karolina committed
765
766
767
        msg = "There is no rc-file specified on the command line. Please use rc=yourfile.rc"
        logging.error(msg)
        raise IOError, msg
768
    elif not os.path.exists(args['rc']):
karolina's avatar
karolina committed
769
770
771
        msg = "The specified rc-file (%s) does not exist " % args['rc'] 
        logging.error(msg)
        raise IOError, msg
772
773
774
775
776
777
778
779
780

    # 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
781
    return opts, args
782
783
784


if __name__ == "__main__":
785
786

    sys.path.append('../../')
karolina's avatar
karolina committed
787
    opts, args = ParseOptions()
788
789
    print opts
    print args
790