initexit.py 29.5 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

Valid options are defined in 

15
.. autofunction:: da.tools.initexit.parse_options
16
17
18
19
20
21

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
   :members: initialize, finalize,  collect_restart_data, move_restart_data, 
44
             submit_next_cycle, setup_file_structure, recover_run, random_seed
45
46
47
48
49
50
51

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:

52
53
.. autofunction:: da.tools.initexit.start_logger 
.. autofunction:: da.tools.initexit.validate_opts_args 
54
55


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
        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``. 
87
88
        This rc-file is loaded by method :meth:`~da.tools.initexit.CycleControl.load_rc` and validated
        by :meth:`~da.tools.initexit.CycleControl.validate_rc`
89

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

93
94
        """

95
96
        self.load_rc(args['rc'])
        self.validate_rc()
97
98
99
100
        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


110
    def load_rc(self, RcFileName):
111
112
113
        """ 
        This method loads a DA Cycle rc-file with settings for this simulation 
        """
114
        import da.tools.rc as rc
115
116

        rcdata = rc.read(RcFileName)
karolina's avatar
karolina committed
117
        for k, v in rcdata.iteritems():
118
            self[k] = v
karolina's avatar
karolina committed
119
120
        self.RcFileName = RcFileName
        self.DaRcLoaded = True
121

karolina's avatar
karolina committed
122
        logging.info('DA Cycle rc-file (%s) loaded successfully' % self.RcFileName)
123
124


125
    def validate_rc(self):
126
        """ 
Peters, Wouter's avatar
Peters, Wouter committed
127
128
        Validate the contents of the rc-file given a dictionary of required keys. 
        Currently required keys are :attr:`~da.tools.initexit.needed_da_items`
129
        """
130
        from da.tools.general import to_datetime
131

karolina's avatar
karolina committed
132
133
134
135
136
        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
137
            if 'date' in k : 
138
                self[k] = to_datetime(v)
139
            if 'time.start' in k : 
140
                self[k] = to_datetime(v)
141
            if 'time.end' in k : 
142
                self[k] = to_datetime(v)
143
            if 'time.finish' in k : 
144
                self[k] = to_datetime(v)
145
146
        for key in needed_da_items:
            if not self.has_key(key):
karolina's avatar
karolina committed
147
148
149
150
151
152
153
154
                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')
155

156
    def parse_times(self):
157
158
159
        """ 
        Parse time related parameters into datetime objects for later use 
        """
160
        from da.tools.general import advance_time
161
162
163
164
165

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

        if finaldate <= startdate:
karolina's avatar
karolina committed
166
            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'))) 
167
168
169
170
171
172
173
174
            raise ValueError
        cyclelength = self['time.cycle']                 # get time step

# Determine end date

        if cyclelength == 'infinite':
            enddate = finaldate
        else:
175
            enddate = advance_time(startdate, cyclelength)
176

karolina's avatar
karolina committed
177
        dt = enddate - startdate
178

karolina's avatar
karolina committed
179
        if enddate > finaldate:  # do not run beyond finaldate
180
181
            enddate = finaldate

karolina's avatar
karolina committed
182
183
184
185
        self['time.start'] = startdate
        self['time.end'] = enddate
        self['time.finish'] = finaldate
        self['cyclelength'] = dt
186

karolina's avatar
karolina committed
187
188
        logging.info("===============================================================")
        logging.info("DA Cycle start date is %s" % startdate.strftime('%Y-%m-%d %H:%M'))
189
        logging.info("DA Cycle end date is %s" % enddate.strftime('%Y-%m-%d %H:%M'))
karolina's avatar
karolina committed
190
191
192
193
        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("===============================================================")
194
195


196
    def set_sample_times(self, lag):
197
198
199
200
201
202
203
        """
        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
204
205
        self['time.sample.start'] = copy.deepcopy(self['time.start'])
        self['time.sample.end'] = copy.deepcopy(self['time.end'])
206
207
208
209

        # Now advance depending on lag

        for l in range(lag):
210
            self.advance_sample_times()
211
212


213
    def advance_sample_times(self):
214
215
216
        """ 
        Advance sampling start and end time by one cycle interval
        """
217
        from da.tools.general import advance_time
218

219
        days = self['cyclelength'].days                
220

221
222
        self['time.sample.start'] = advance_time(self['time.sample.start'], days)
        self['time.sample.end'] = advance_time(self['time.sample.end'], days)
223
224
    

225
    def advance_cycle_times(self):
226
227
228
        """ 
        Advance cycle start and end time by one cycle interval
        """
229
230
231
        from da.tools.general import advance_time
              
        days = self['cyclelength'].days                  
232

233
234
        startdate = advance_time(self['time.start'], days)
        enddate = advance_time(self['time.end'], days)
235

karolina's avatar
karolina committed
236
237
        filtertime = startdate.strftime('%Y%m%d')
        self['dir.output'] = os.path.join(self['dir.da_run'], 'output', filtertime)
238

karolina's avatar
karolina committed
239
240
        self['time.start'] = startdate
        self['time.end'] = enddate
241
242


243
    def random_seed(self, action='read'):
Peters, Wouter's avatar
Peters, Wouter committed
244
245
        """ 
        Get the randomseed and save it, or read the random seed and set it. The seed is currently stored
246
        in a python :mod:`pickle` file, residing in the ``exec`` directory
Peters, Wouter's avatar
Peters, Wouter committed
247
248

        """
249
250
251
        import cPickle
        import numpy as np

karolina's avatar
karolina committed
252
        filename = os.path.join(self['dir.exec'], 'randomseed.pickle')
253
254

        if action == 'write':
karolina's avatar
karolina committed
255
256
            f = open(filename, 'wb')
            seed = np.random.get_state()
karolina's avatar
karolina committed
257
258
            cPickle.dump(seed, f, -1)
            f.close()
259

karolina's avatar
karolina committed
260
            msg = "Saved the random seed generator values to file"
261
262

        if action == 'read':
karolina's avatar
karolina committed
263
264
            f = open(filename, 'rb')
            seed = cPickle.load(f)
karolina's avatar
karolina committed
265
266
            np.random.set_state(seed)
            f.close()
267

karolina's avatar
karolina committed
268
            msg = "Retrieved the random seed generator values from file"
269
270

        logging.info(msg)
karolina's avatar
karolina committed
271
        self.RestartFileList.append(filename)
karolina's avatar
karolina committed
272
        logging.debug("Added the randomseed.pickle file to the RestartFileList")
273
274


275
    def initialize(self):
276
277
278
        """ 
        This method determines how to proceed with the cycle. Three options are implemented:

Peters, Wouter's avatar
Peters, Wouter committed
279
280
281
            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
282
283
284

        The choice that gets executed depends on the presence of 

Peters, Wouter's avatar
Peters, Wouter committed
285
286
            # the ``-r`` option on the command line, this triggers a recover
            # the ``time.restart : True`` option in the da.rc file
287
288

        The latter is automatically set if the filter submits the next cycle at the end of the current one, 
289
        through method :meth:`~da.tools.initexit.CycleControl.submit_next_cycle`.
Peters, Wouter's avatar
Peters, Wouter committed
290

Peters, Wouter's avatar
Peters, Wouter committed
291
        The specific call tree under each scenario is: 
Peters, Wouter's avatar
Peters, Wouter committed
292

Peters, Wouter's avatar
Peters, Wouter committed
293
            1. *Fresh Start*
294
                *  :meth:`~da.tools.initexit.CycleControl.setup_file_structure()`  <- Create directory tree
Peters, Wouter's avatar
Peters, Wouter committed
295
            2. *Restart*
296
297
                *  :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
298
            3. *Recover*
299
300
301
                *  :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
302
303
304

        And is always followed by a call to

305
            * parse_times()
Peters, Wouter's avatar
Peters, Wouter committed
306
            * WriteRc('jobfilename')
307
308
309
310
311
        """        
#
# case 1: A recover from a previous crash, this is signaled by flag "-r"
#
        if self['da.crash.recover']:
karolina's avatar
karolina committed
312
            logging.info("Recovering simulation from data in: %s" % self['dir.da_run'])
313
314
315
            self.setup_file_structure()
            self.recover_run()
            self.random_seed('read')
316
317
318
319
#
# case 2: A continuation, this is signaled by rc-item time.restart = True
#
        elif self['time.restart']:
karolina's avatar
karolina committed
320
            logging.info("Restarting filter from previous step")
321
            self.setup_file_structure()
322
            strippedname = os.path.split(self['jobrcfilename'])[-1]
karolina's avatar
karolina committed
323
            self['jobrcfilename'] = os.path.join(self['dir.exec'], strippedname)
324
            self.random_seed('read')
325
326
327
328
#
# case 3: A fresh start, this is signaled by rc-item time.restart = False
#
        elif not self['time.restart']:
karolina's avatar
karolina committed
329
            logging.info("First time step in filter sequence")
330
            self.setup_file_structure()
331
332

            # expand jobrcfilename to include exec dir from now on.
333
334
335
            # First strip current leading path from filename

            strippedname = os.path.split(self['jobrcfilename'])[-1]
karolina's avatar
karolina committed
336
            self['jobrcfilename'] = os.path.join(self['dir.exec'], strippedname)
337

338
339
        self.parse_times()
        self.write_rc(self['jobrcfilename'])
340

341
    def setup_file_structure(self):
342
343
        """ 
        Create file structure needed for data assimilation system.
Peters, Wouter's avatar
Peters, Wouter committed
344
345
346
347
348
349
350
351
352
        In principle this looks like:

            * ``${da_rundir}``
            * ``${da_rundir}/input``
            * ``${da_rundir}/output``
            * ``${da_rundir}/exec``
            * ``${da_rundir}/diagnostics``
            * ``${da_rundir}/analysis``
            * ``${da_rundir}/jobs``
353
354
            * ``${da_rundir}/restart/current``
            * ``${da_rundir}/restart/one-ago``
Peters, Wouter's avatar
Peters, Wouter committed
355

356
        .. note:: The exec dir will actually be a simlink to the directory where
357
358
359
                 the observation operator executable lives. This directory is passed through
                 the ``da.rc`` file. 

360
        .. note:: The observation input files will be placed in the exec dir,
361
                 and the resulting simulated values will be retrieved from there as well.
Peters, Wouter's avatar
Peters, Wouter committed
362

363
        """
364
        from da.tools.general import create_dirs
365
366
367

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

karolina's avatar
karolina committed
368
        filtertime = self['time.start'].strftime('%Y%m%d')
369

karolina's avatar
karolina committed
370
371
372
373
374
375
376
377
378
        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')
379

380
381
382
383
384
385
386
387
388
389
        create_dirs(self['dir.da_run'])
        create_dirs(os.path.join(self['dir.exec']))
        create_dirs(os.path.join(self['dir.input']))
        create_dirs(os.path.join(self['dir.output']))
        create_dirs(os.path.join(self['dir.diagnostics']))
        create_dirs(os.path.join(self['dir.analysis']))
        create_dirs(os.path.join(self['dir.jobs']))
        create_dirs(os.path.join(self['dir.restart']))
        create_dirs(os.path.join(self['dir.restart.current']))
        create_dirs(os.path.join(self['dir.restart.oneago']))
390

karolina's avatar
karolina committed
391
        logging.info('Succesfully created the file structure for the assimilation job')
392
#LU tutaj chyba brakuje move restart data
393
    def recover_run(self):
Peters, Wouter's avatar
Peters, Wouter committed
394
395
        """
        Prepare a recovery from a crashed run. This consists of: 
396
        
397
            - copying all data from the restart/one-ago folder (:meth:`~da.tools.initexit.CycleControl.move_restart_data`),
398
            - replacing all ``rc-file`` items with those from the ``da_runtime.rc`` in the restart/current dir 
399
            - resetting the seed of the random number generator to the value it had before the crash (:meth:`~da.tools.initexit.CycleControl.random_seed`)
400
            - replacing the output dir name, since it has the sample time in it...
Peters, Wouter's avatar
Peters, Wouter committed
401
            
402
        """
403
        import da.tools.rc as rc
404
        from da.tools.general import create_dirs
405

406
        # Replace rc-items with those from the crashed run's last rc-file (now in restart.current dir)
407

karolina's avatar
karolina committed
408
409
        file_rc_rec = os.path.join(self['dir.restart.current'], 'da_runtime.rc')
        rc_rec = rc.read(file_rc_rec)
410

karolina's avatar
karolina committed
411
        for k, v in rc_rec.iteritems():
412
413
            self[k] = v

414
        self.validate_rc()
415

karolina's avatar
karolina committed
416
417
        logging.debug("Replaced rc-items.... ")
        logging.debug("Next cycle start date is %s" % self['time.start'])
418

419
        # Copy randomseed.pickle file to exec dir
420
        source = os.path.join(self['dir.restart.current'], 'randomseed.pickle') #LU wydaje mi sie ze tutaj nie trzeba podawac nazwy pliku w folderze docelowym, jesli sie obczai ze to folder to sie kopiuje.
karolina's avatar
karolina committed
421
422
        dest = os.path.join(self['dir.exec'], 'randomseed.pickle')
        shutil.copy(source, dest)
423

karolina's avatar
karolina committed
424
        logging.debug("Replaced randomseed file with previous cycles' last values")
425
426
        # Re-create the output dir for this time step, if needed

karolina's avatar
karolina committed
427
428
        filtertime = self['time.start'].strftime('%Y%m%d')
        self['dir.output'] = os.path.join(self['dir.da_run'], 'output', filtertime)
429
        create_dirs(os.path.join(self['dir.output']))
430

431
    def finalize(self):
432
        """
433
        finalize the da cycle, this means writing the save data and rc-files for the next run. 
434
435
436
        The following sequence of actions occur:

            * Write the randomseed to file for reuse in next cycle
Peters, Wouter's avatar
Peters, Wouter committed
437
            * Write a new ``rc-file`` with ``time.restart : True``, and new ``time.start`` and ``time.end``
438
439
440
441
            * 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

442
        """
443
444
445
446
447
448
449
450
        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):
451
        """ Collect files that are part of the requested output for this cycle. This function allows users to add files 
452
453
454
455
456
457
458
            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.


        """
459
        from da.tools.general import create_dirs
460

karolina's avatar
karolina committed
461
        targetdir = os.path.join(self['dir.output'])
462
        create_dirs(os.path.join(targetdir))
463

karolina's avatar
karolina committed
464
465
        logging.info("Collecting the required output data") 
        logging.debug("           to   directory: %s " % targetdir)
466
467
468
469

        for file in set(self.OutputFileList):
            if os.path.isdir(file): # skip dirs
                continue
470
            if not os.path.exists(file): # skip dirs
karolina's avatar
karolina committed
471
                logging.warning("           [not found] .... %s " % file)
472
                continue
473

karolina's avatar
karolina committed
474
475
            logging.debug("           [copy] .... %s " % file)
            shutil.copy(file, file.replace(os.path.split(file)[0], targetdir))
476
477
478



479
    def collect_restart_data(self):
480
        """ Collect files needed for the restart of this cycle in case of a crash, or for the continuation of the next cycle. 
481
            All files needed are written to the restart/current directory. The list of files included is read from the 
482
483
484
            attribute "RestartFileList" which is a simple list of files that can be appended by other objects/methods that
            require restart data to be saved.

485
            .. note:: Before collecting the files in the ``RestartFileList``, the restart/current directory will be emptied and
486
487
488
489
490
                     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:
491

492
                * The ``da_runtime.rc`` file
493
                * The ``randomseed.pickle`` file
494
495
                * The savestate.nc file
                * The files in the ``ObservationOperator.RestartFileList``, i.e., restart data for the transport model
496

497

Peters, Wouter's avatar
Peters, Wouter committed
498
            .. note:: We assume that the restart files for the :ref:`ObservationOperator` 
499
                      reside in a separate folder, i.e, the ObservationOperator does *not* write directly to the CTDAS restart dir!
500
501

        """
502
        from da.tools.general import create_dirs
503

karolina's avatar
karolina committed
504
        targetdir = os.path.join(self['dir.restart.current'])
505

karolina's avatar
karolina committed
506
        logging.info("Purging the current restart directory before collecting new data")
507

508
        create_dirs(os.path.join(targetdir), forceclean=True)
509

karolina's avatar
karolina committed
510
511
        logging.info("Collecting the required restart data")
        logging.debug("           to   directory: %s " % targetdir)
512

513
        for file in set(self.RestartFileList):
514
515
            if os.path.isdir(file): # skip dirs
                continue
516
            if not os.path.exists(file): # skip dirs
karolina's avatar
karolina committed
517
                logging.warning("           [not found] .... %s " % file)
518
            else:
karolina's avatar
karolina committed
519
520
                logging.debug("           [copy] .... %s " % file)
                shutil.copy(file, file.replace(os.path.split(file)[0], targetdir))
521
522


523
    def move_restart_data(self, io_option='restore'):
Peters, Wouter's avatar
Peters, Wouter committed
524
        """ 
525
        Store or restore model state to/from a restart directory. 
526

527
            Two IO options are available:
528

529
530
            (1) io_option = restore    : Get data from restart.oneago directory
            (2) io_option = store      : Save data to restart.oneago directory
531

532
            In case of a 'store' command the restart.oneago folder is re-created so that the contents are empty to begin with.
533
534

        """
535
        from da.tools.general import create_dirs
536

karolina's avatar
karolina committed
537
538
        if io_option not in ['store', 'restore']:
            raise ValueError, 'Invalid option specified for io_option (%s)' % io_option
539

540
541
542
543
544
545
        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']
546
547
548
549

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

        if io_option == 'store':
550
            create_dirs(os.path.join(targetdir), forceclean=True)
551

552
        logging.debug("Performing a %s of data" % io_option)
karolina's avatar
karolina committed
553
554
        logging.debug("           from directory: %s " % sourcedir)
        logging.debug("           to   directory: %s " % targetdir)
555
556
557


        for file in os.listdir(sourcedir):
karolina's avatar
karolina committed
558
            file = os.path.join(sourcedir, file)
559
            if not os.path.exists(file):
karolina's avatar
karolina committed
560
                logging.debug("Cannot find requested file to move: %s " % file)  
561
562
                sys.exit(2)

563
            if os.path.isdir(file): # skip dirs
karolina's avatar
karolina committed
564
                logging.debug("           [skip] .... %s " % file)
565
                continue    
566
            else:
karolina's avatar
karolina committed
567
568
                logging.debug("           [copy] .... %s " % file)
                shutil.copy(file, file.replace(sourcedir, targetdir))
569
570

#
571
    def write_new_rc_file(self):
572
573
574
575
576
577
        """ 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
578
579
            
        """
580
        import da.tools.rc as rc
581
        import copy
582

583
        # We make a copy of the current DaCycle object, and modify the start + end dates and restart value
584

karolina's avatar
karolina committed
585
        newDaCycle = copy.deepcopy(self)
586
        newDaCycle.advance_cycle_times()
karolina's avatar
karolina committed
587
        newDaCycle['time.restart'] = True
588

589
        # Create the name of the rc-file that will hold this new input, and write it
590

karolina's avatar
karolina committed
591
592
593
        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)
594

595
        # The rest is info needed for a system restart, so it modifies the current DaCycle object (self)
596

karolina's avatar
karolina committed
597
        self['da.restart.fname'] = fname    # needed for next job template
598
        self.RestartFileList.append(fname)  # current restart list holds next rc file name
599

karolina's avatar
karolina committed
600
        logging.debug('Added da_runtime.rc to the RestartFileList for later collection')
601
602


603
    def write_rc(self, fname):
604
        """ Write RC file after each process to reflect updated info """
605
        import da.tools.rc as rc
606

karolina's avatar
karolina committed
607
608
609
        rc.write(fname, self)
        logging.debug('Wrote expanded rc-file (%s)' % (fname))
        
610

611
    def submit_next_cycle(self):
612
        """ 
Peters, Wouter's avatar
Peters, Wouter committed
613
        Submit the next job of a DA cycle, this consists of 
614
615
616
            * 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
617
618
619

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

620
        """
621
        from string import join
622

623
        if self['time.end'] < self['time.finish']:
624

Peters, Wouter's avatar
test    
Peters, Wouter committed
625
            # file ID and names
karolina's avatar
karolina committed
626
627
628
            jobid = self['time.end'].strftime('%Y%m%d') 
            targetdir = os.path.join(self['dir.exec'])
            jobfile = os.path.join(targetdir, 'jb.%s.jb' % jobid)
629
630
            logfile = os.path.join(targetdir, 'jb.%s.log' % jobid)
#LU tutaj sa parametry ktore ida na gore do pliku job. nie zawsze koniecznie potrzebne.
Peters, Wouter's avatar
test    
Peters, Wouter committed
631
            # Template and commands for job
632
633
            jobparams = {'jobname':"j.%s" % jobid, 'jobtime':'06:00:00', 'logfile': logfile, 'errfile': logfile}
            template = self.DaPlatForm.get_job_template(jobparams)
karolina's avatar
karolina committed
634
            execcommand = os.path.join(self['dir.da_submit'], sys.argv[0]) 
635
            template += 'python %s rc=%s %s >&%s' % (execcommand, self['da.restart.fname'], join(self.opts, ''),logfile) 
Peters, Wouter's avatar
test    
Peters, Wouter committed
636
637

            # write and submit 
638
639
            self.DaPlatForm.write_job(jobfile, template, jobid)
            jobid = self.DaPlatForm.submit_job(jobfile, joblog=logfile) 
640
641
642
643
        else:
            logging.info('Final date reached, no new cycle started')


644
def start_logger(level=logging.INFO):
645
646
647
648
    """ start the logging of messages to screen"""

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

karolina's avatar
karolina committed
649
650
651
    logging.basicConfig(level=level,
                        format=' [%(levelname)-7s] (%(asctime)s) py-%(module)-20s : %(message)s',
                        datefmt='%Y-%m-%d %H:%M:%S')
652

653
def parse_options():
654
655
656
657
658
659
660
661
662
663
664
665
666
    """ 
    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
    ========  =======

    """
667
668
669
670
    import getopt

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

karolina's avatar
karolina committed
671
672
    opts = []
    args = []
673
    try:                                
674
        opts, args = getopt.gnu_getopt(sys.argv[1:], "-rv")
675
    except getopt.GetoptError, msg:           
karolina's avatar
karolina committed
676
        logging.error('%s' % msg)
677
678
679
        sys.exit(2)      

    for options in opts:
karolina's avatar
karolina committed
680
        options = options[0].lower()
681
682
683
684
        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
685
            logging.root.setLevel(logging.DEBUG)
686

687
    if opts: 
688
        optslist = [item[0] for item in opts] #LU ze co same minusy zwroci?
689
    else:
karolina's avatar
karolina committed
690
        optslist = []
691

692
693
# Parse arguments and return as dictionary

karolina's avatar
karolina committed
694
    arguments = {}
695
696
697
698
699
700
701
702
    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
703
704
            logging.error('%s' % 'Argument passed without description (%s)' % item)
            raise getopt.GetoptError, arg
705

karolina's avatar
karolina committed
706
        arguments[key] = arg
707
708


709
    return optslist, arguments
710

711
def validate_opts_args(opts, args):
712
713
714
715
716
    """ 
 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.  
 
    """
717
    if not args.has_key("rc"):
karolina's avatar
karolina committed
718
719
720
        msg = "There is no rc-file specified on the command line. Please use rc=yourfile.rc"
        logging.error(msg)
        raise IOError, msg
721
    elif not os.path.exists(args['rc']):
karolina's avatar
karolina committed
722
723
724
        msg = "The specified rc-file (%s) does not exist " % args['rc'] 
        logging.error(msg)
        raise IOError, msg
725
726
727
728
729
730
731
732
733

    # 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
734
    return opts, args
735
736
737


if __name__ == "__main__":
738
739

    sys.path.append('../../')
740
    opts, args = parse_options()
741
742
    print opts
    print args
743