initexit.py 28.9 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

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
karolina's avatar
karolina committed
61
import copy
62
63
64
import getopt
import cPickle
import numpy as np
karolina's avatar
karolina committed
65
66
67
from string import join

import da.tools.rc as rc
68
from da.tools.general import create_dirs, to_datetime, advance_time
69

karolina's avatar
karolina committed
70
needed_da_items = [
71
72
73
74
75
    'time.start',
    'time.finish',
    'time.nlag',
    'time.cycle',
    'dir.da_run',
76
77
78
79
80
    'da.system',
    'da.system.rc',
    'da.obsoperator',
    'da.obsoperator.rc',
    'da.optimizer.nmembers']
81
82
83
84
85
86
87

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


class CycleControl(dict):
    """
88
    This object controls the CTDAS system flow and functionality.
89
    """
90
        
karolina's avatar
karolina committed
91
    def __init__(self, opts=[], args={}):
92
        """
Peters, Wouter's avatar
Peters, Wouter committed
93
94
        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``. 
95
96
        This rc-file is loaded by method :meth:`~da.tools.initexit.CycleControl.load_rc` and validated
        by :meth:`~da.tools.initexit.CycleControl.validate_rc`
97

Peters, Wouter's avatar
Peters, Wouter committed
98
        Options for the CycleControl consist of accepted command line flags or arguments 
99
        in :func:`~da.tools.initexit.CycleControl.parse_options`
100

101
102
        """

103
104
        self.load_rc(args['rc'])
        self.validate_rc()
105
106
107
108
        self.opts = opts

        # Add some useful variables to the rc-file dictionary

karolina's avatar
karolina committed
109
110
        self['jobrcfilename'] = self.RcFileName
        self['dir.da_submit'] = os.getcwd()
111
        self['da.crash.recover'] = '-r' in opts
karolina's avatar
karolina committed
112
113
114
115
        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
116
117


118
    def load_rc(self, RcFileName):
119
120
121
122
123
        """ 
        This method loads a DA Cycle rc-file with settings for this simulation 
        """

        rcdata = rc.read(RcFileName)
karolina's avatar
karolina committed
124
        for k, v in rcdata.iteritems():
125
            self[k] = v
karolina's avatar
karolina committed
126
127
        self.RcFileName = RcFileName
        self.DaRcLoaded = True
128

karolina's avatar
karolina committed
129
        logging.info('DA Cycle rc-file (%s) loaded successfully' % self.RcFileName)
130
131


132
    def validate_rc(self):
133
        """ 
Peters, Wouter's avatar
Peters, Wouter committed
134
135
        Validate the contents of the rc-file given a dictionary of required keys. 
        Currently required keys are :attr:`~da.tools.initexit.needed_da_items`
136
137
        """

karolina's avatar
karolina committed
138
139
140
141
142
        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
143
            if 'date' in k : 
144
                self[k] = to_datetime(v)
145
            if 'time.start' in k : 
146
                self[k] = to_datetime(v)
147
            if 'time.end' in k : 
148
                self[k] = to_datetime(v)
149
            if 'time.finish' in k : 
150
                self[k] = to_datetime(v)
151
152
        for key in needed_da_items:
            if not self.has_key(key):
karolina's avatar
karolina committed
153
154
155
156
157
158
159
160
                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')
161

162
    def parse_times(self):
163
164
165
166
167
168
169
170
        """ 
        Parse time related parameters into datetime objects for later use 
        """

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

        if finaldate <= startdate:
karolina's avatar
karolina committed
171
            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'))) 
172
173
174
175
176
177
178
179
            raise ValueError
        cyclelength = self['time.cycle']                 # get time step

# Determine end date

        if cyclelength == 'infinite':
            enddate = finaldate
        else:
180
            enddate = advance_time(startdate, cyclelength)
181

karolina's avatar
karolina committed
182
        dt = enddate - startdate
183

karolina's avatar
karolina committed
184
        if enddate > finaldate:  # do not run beyond finaldate
185
186
            enddate = finaldate

karolina's avatar
karolina committed
187
188
189
190
        self['time.start'] = startdate
        self['time.end'] = enddate
        self['time.finish'] = finaldate
        self['cyclelength'] = dt
191

karolina's avatar
karolina committed
192
193
        logging.info("===============================================================")
        logging.info("DA Cycle start date is %s" % startdate.strftime('%Y-%m-%d %H:%M'))
194
        logging.info("DA Cycle end date is %s" % enddate.strftime('%Y-%m-%d %H:%M'))
karolina's avatar
karolina committed
195
196
197
198
        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("===============================================================")
199
200


201
    def set_sample_times(self, lag):
202
203
204
205
206
207
        """
        Set the times over which a sampling interval will loop, depending on 
        the lag. Note that lag falls in the interval [0,nlag-1]
        """

        # Start from cycle times 
karolina's avatar
karolina committed
208
209
        self['time.sample.start'] = copy.deepcopy(self['time.start'])
        self['time.sample.end'] = copy.deepcopy(self['time.end'])
210
211
212
213

        # Now advance depending on lag

        for l in range(lag):
214
            self.advance_sample_times()
215
216


217
    def advance_sample_times(self):
218
219
220
221
        """ 
        Advance sampling start and end time by one cycle interval
        """

222
        days = self['cyclelength'].days                
223

224
225
        self['time.sample.start'] = advance_time(self['time.sample.start'], days)
        self['time.sample.end'] = advance_time(self['time.sample.end'], days)
226
227
    

228
    def advance_cycle_times(self):
229
230
231
        """ 
        Advance cycle start and end time by one cycle interval
        """
232
233
              
        days = self['cyclelength'].days                  
234

235
236
        startdate = advance_time(self['time.start'], days)
        enddate = advance_time(self['time.end'], days)
237

karolina's avatar
karolina committed
238
239
        filtertime = startdate.strftime('%Y%m%d')
        self['dir.output'] = os.path.join(self['dir.da_run'], 'output', filtertime)
240

karolina's avatar
karolina committed
241
242
        self['time.start'] = startdate
        self['time.end'] = enddate
243
244


245
    def random_seed(self, action='read'):
Peters, Wouter's avatar
Peters, Wouter committed
246
247
        """ 
        Get the randomseed and save it, or read the random seed and set it. The seed is currently stored
248
        in a python :mod:`pickle` file, residing in the ``exec`` directory
Peters, Wouter's avatar
Peters, Wouter committed
249
250

        """
251

252

karolina's avatar
karolina committed
253
        filename = os.path.join(self['dir.exec'], 'randomseed.pickle')
254
255

        if action == 'write':
karolina's avatar
karolina committed
256
257
            f = open(filename, 'wb')
            seed = np.random.get_state()
karolina's avatar
karolina committed
258
259
            cPickle.dump(seed, f, -1)
            f.close()
260

karolina's avatar
karolina committed
261
            msg = "Saved the random seed generator values to file"
262
263

        if action == 'read':
karolina's avatar
karolina committed
264
265
            f = open(filename, 'rb')
            seed = cPickle.load(f)
karolina's avatar
karolina committed
266
267
            np.random.set_state(seed)
            f.close()
268

karolina's avatar
karolina committed
269
            msg = "Retrieved the random seed generator values from file"
270
271

        logging.info(msg)
karolina's avatar
karolina committed
272
        self.RestartFileList.append(filename)
karolina's avatar
karolina committed
273
        logging.debug("Added the randomseed.pickle file to the RestartFileList")
274
275


276
    def initialize(self):
277
278
279
        """ 
        This method determines how to proceed with the cycle. Three options are implemented:

Peters, Wouter's avatar
Peters, Wouter committed
280
281
282
            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
283
284
285

        The choice that gets executed depends on the presence of 

Peters, Wouter's avatar
Peters, Wouter committed
286
287
            # the ``-r`` option on the command line, this triggers a recover
            # the ``time.restart : True`` option in the da.rc file
288
289

        The latter is automatically set if the filter submits the next cycle at the end of the current one, 
290
        through method :meth:`~da.tools.initexit.CycleControl.submit_next_cycle`.
Peters, Wouter's avatar
Peters, Wouter committed
291

Peters, Wouter's avatar
Peters, Wouter committed
292
        The specific call tree under each scenario is: 
Peters, Wouter's avatar
Peters, Wouter committed
293

Peters, Wouter's avatar
Peters, Wouter committed
294
            1. *Fresh Start*
295
                *  :meth:`~da.tools.initexit.CycleControl.setup_file_structure()`  <- Create directory tree
Peters, Wouter's avatar
Peters, Wouter committed
296
            2. *Restart*
297
298
                *  :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
299
            3. *Recover*
300
301
302
                *  :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
303
304
305

        And is always followed by a call to

306
            * parse_times()
Peters, Wouter's avatar
Peters, Wouter committed
307
            * WriteRc('jobfilename')
308
309
310
311
312
        """        
#
# case 1: A recover from a previous crash, this is signaled by flag "-r"
#
        if self['da.crash.recover']:
karolina's avatar
karolina committed
313
            logging.info("Recovering simulation from data in: %s" % self['dir.da_run'])
314
315
316
            self.setup_file_structure()
            self.recover_run()
            self.random_seed('read')
317
318
319
320
#
# case 2: A continuation, this is signaled by rc-item time.restart = True
#
        elif self['time.restart']:
karolina's avatar
karolina committed
321
            logging.info("Restarting filter from previous step")
322
            self.setup_file_structure()
323
            strippedname = os.path.split(self['jobrcfilename'])[-1]
karolina's avatar
karolina committed
324
            self['jobrcfilename'] = os.path.join(self['dir.exec'], strippedname)
325
            self.random_seed('read')
326
327
328
329
#
# case 3: A fresh start, this is signaled by rc-item time.restart = False
#
        elif not self['time.restart']:
karolina's avatar
karolina committed
330
            logging.info("First time step in filter sequence")
331
            self.setup_file_structure()
332
333

            # expand jobrcfilename to include exec dir from now on.
334
335
336
            # First strip current leading path from filename

            strippedname = os.path.split(self['jobrcfilename'])[-1]
karolina's avatar
karolina committed
337
            self['jobrcfilename'] = os.path.join(self['dir.exec'], strippedname)
338

339
340
        self.parse_times()
        self.write_rc(self['jobrcfilename'])
341

342
    def setup_file_structure(self):
343
344
        """ 
        Create file structure needed for data assimilation system.
Peters, Wouter's avatar
Peters, Wouter committed
345
346
347
348
349
350
351
352
353
        In principle this looks like:

            * ``${da_rundir}``
            * ``${da_rundir}/input``
            * ``${da_rundir}/output``
            * ``${da_rundir}/exec``
            * ``${da_rundir}/diagnostics``
            * ``${da_rundir}/analysis``
            * ``${da_rundir}/jobs``
354
355
            * ``${da_rundir}/restart/current``
            * ``${da_rundir}/restart/one-ago``
Peters, Wouter's avatar
Peters, Wouter committed
356

357
        .. note:: The exec dir will actually be a simlink to the directory where
358
359
360
                 the observation operator executable lives. This directory is passed through
                 the ``da.rc`` file. 

361
        .. note:: The observation input files will be placed in the exec dir,
362
                 and the resulting simulated values will be retrieved from there as well.
Peters, Wouter's avatar
Peters, Wouter committed
363

364
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
        """
karolina's avatar
karolina committed
403
        
404

405
        # Replace rc-items with those from the crashed run's last rc-file (now in restart.current dir)
406

karolina's avatar
karolina committed
407
408
        file_rc_rec = os.path.join(self['dir.restart.current'], 'da_runtime.rc')
        rc_rec = rc.read(file_rc_rec)
409

karolina's avatar
karolina committed
410
        for k, v in rc_rec.iteritems():
411
412
            self[k] = v

413
        self.validate_rc()
414

karolina's avatar
karolina committed
415
416
        logging.debug("Replaced rc-items.... ")
        logging.debug("Next cycle start date is %s" % self['time.start'])
417

418
        # Copy randomseed.pickle file to exec dir
419
        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
420
421
        dest = os.path.join(self['dir.exec'], 'randomseed.pickle')
        shutil.copy(source, dest)
422

karolina's avatar
karolina committed
423
        logging.debug("Replaced randomseed file with previous cycles' last values")
424
425
        # Re-create the output dir for this time step, if needed

karolina's avatar
karolina committed
426
427
        filtertime = self['time.start'].strftime('%Y%m%d')
        self['dir.output'] = os.path.join(self['dir.da_run'], 'output', filtertime)
428
        create_dirs(os.path.join(self['dir.output']))
429

430
    def finalize(self):
431
        """
432
        finalize the da cycle, this means writing the save data and rc-files for the next run. 
433
434
435
        The following sequence of actions occur:

            * Write the randomseed to file for reuse in next cycle
Peters, Wouter's avatar
Peters, Wouter committed
436
            * Write a new ``rc-file`` with ``time.restart : True``, and new ``time.start`` and ``time.end``
437
438
439
440
            * 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

441
        """
442
443
444
445
446
447
448
449
        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):
450
        """ Collect files that are part of the requested output for this cycle. This function allows users to add files 
451
452
453
454
455
456
457
            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.


        """
karolina's avatar
karolina committed
458
        targetdir = os.path.join(self['dir.output'])
459
        create_dirs(targetdir)
460

karolina's avatar
karolina committed
461
462
        logging.info("Collecting the required output data") 
        logging.debug("           to   directory: %s " % targetdir)
463
464
465
466

        for file in set(self.OutputFileList):
            if os.path.isdir(file): # skip dirs
                continue
467
            if not os.path.exists(file): # skip dirs
karolina's avatar
karolina committed
468
                logging.warning("           [not found] .... %s " % file)
469
                continue
470

karolina's avatar
karolina committed
471
472
            logging.debug("           [copy] .... %s " % file)
            shutil.copy(file, file.replace(os.path.split(file)[0], targetdir))
473
474
475



476
    def collect_restart_data(self):
477
        """ Collect files needed for the restart of this cycle in case of a crash, or for the continuation of the next cycle. 
478
            All files needed are written to the restart/current directory. The list of files included is read from the 
479
480
481
            attribute "RestartFileList" which is a simple list of files that can be appended by other objects/methods that
            require restart data to be saved.

482
            .. note:: Before collecting the files in the ``RestartFileList``, the restart/current directory will be emptied and
483
484
485
486
487
                     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:
488

489
                * The ``da_runtime.rc`` file
490
                * The ``randomseed.pickle`` file
491
492
                * The savestate.nc file
                * The files in the ``ObservationOperator.RestartFileList``, i.e., restart data for the transport model
493

494

Peters, Wouter's avatar
Peters, Wouter committed
495
            .. note:: We assume that the restart files for the :ref:`ObservationOperator` 
496
                      reside in a separate folder, i.e, the ObservationOperator does *not* write directly to the CTDAS restart dir!
497
498
499

        """

karolina's avatar
karolina committed
500
        targetdir = os.path.join(self['dir.restart.current'])
501

karolina's avatar
karolina committed
502
        logging.info("Purging the current restart directory before collecting new data")
503

504
        create_dirs(targetdir, forceclean=True)
505

karolina's avatar
karolina committed
506
507
        logging.info("Collecting the required restart data")
        logging.debug("           to   directory: %s " % targetdir)
508

509
        for file in set(self.RestartFileList):
510
511
            if os.path.isdir(file): # skip dirs
                continue
512
            if not os.path.exists(file): # skip dirs
karolina's avatar
karolina committed
513
                logging.warning("           [not found] .... %s " % file)
514
            else:
karolina's avatar
karolina committed
515
516
                logging.debug("           [copy] .... %s " % file)
                shutil.copy(file, file.replace(os.path.split(file)[0], targetdir))
517
518


519
    def move_restart_data(self, io_option='restore'):
Peters, Wouter's avatar
Peters, Wouter committed
520
        """ 
521
        Store or restore model state to/from a restart directory. 
522

523
            Two IO options are available:
524

525
526
            (1) io_option = restore    : Get data from restart.oneago directory
            (2) io_option = store      : Save data to restart.oneago directory
527

528
            In case of a 'store' command the restart.oneago folder is re-created so that the contents are empty to begin with.
529
530

        """
karolina's avatar
karolina committed
531
532
        if io_option not in ['store', 'restore']:
            raise ValueError, 'Invalid option specified for io_option (%s)' % io_option
533

534
535
536
537
538
539
        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']
540
541
542
543

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

        if io_option == 'store':
544
            create_dirs(os.path.join(targetdir), forceclean=True)
545

546
        logging.debug("Performing a %s of data" % io_option)
karolina's avatar
karolina committed
547
548
        logging.debug("           from directory: %s " % sourcedir)
        logging.debug("           to   directory: %s " % targetdir)
549
550
551


        for file in os.listdir(sourcedir):
karolina's avatar
karolina committed
552
            file = os.path.join(sourcedir, file)
553
            if not os.path.exists(file):
karolina's avatar
karolina committed
554
                logging.debug("Cannot find requested file to move: %s " % file)  
555
556
                sys.exit(2)

557
            if os.path.isdir(file): # skip dirs
karolina's avatar
karolina committed
558
                logging.debug("           [skip] .... %s " % file)
559
                continue    
560
            else:
karolina's avatar
karolina committed
561
562
                logging.debug("           [copy] .... %s " % file)
                shutil.copy(file, file.replace(sourcedir, targetdir))
563
564

#
565
    def write_new_rc_file(self):
566
567
568
569
570
571
        """ 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
572
573
            
        """
karolina's avatar
karolina committed
574
575
        
        
576

577
        # We make a copy of the current DaCycle object, and modify the start + end dates and restart value
578

karolina's avatar
karolina committed
579
        newDaCycle = copy.deepcopy(self)
580
        newDaCycle.advance_cycle_times()
karolina's avatar
karolina committed
581
        newDaCycle['time.restart'] = True
582

583
        # Create the name of the rc-file that will hold this new input, and write it
584

karolina's avatar
karolina committed
585
586
587
        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)
588

589
        # The rest is info needed for a system restart, so it modifies the current DaCycle object (self)
590

karolina's avatar
karolina committed
591
        self['da.restart.fname'] = fname    # needed for next job template
592
        self.RestartFileList.append(fname)  # current restart list holds next rc file name
593

karolina's avatar
karolina committed
594
        logging.debug('Added da_runtime.rc to the RestartFileList for later collection')
595
596


597
    def write_rc(self, fname):
598
599
        """ Write RC file after each process to reflect updated info """

karolina's avatar
karolina committed
600
        rc.write(fname, self)
karolina's avatar
karolina committed
601
        logging.debug('Wrote expanded rc-file (%s)' % fname)
karolina's avatar
karolina committed
602
        
603

604
    def submit_next_cycle(self):
605
        """ 
Peters, Wouter's avatar
Peters, Wouter committed
606
        Submit the next job of a DA cycle, this consists of 
607
608
609
            * 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
610
611
612

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

613
        """
karolina's avatar
karolina committed
614
        
615

616
        if self['time.end'] < self['time.finish']:
617

Peters, Wouter's avatar
test    
Peters, Wouter committed
618
            # file ID and names
karolina's avatar
karolina committed
619
620
621
            jobid = self['time.end'].strftime('%Y%m%d') 
            targetdir = os.path.join(self['dir.exec'])
            jobfile = os.path.join(targetdir, 'jb.%s.jb' % jobid)
622
623
            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
624
            # Template and commands for job
625
626
            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
627
            execcommand = os.path.join(self['dir.da_submit'], sys.argv[0]) 
karolina's avatar
karolina committed
628
            template += 'python %s rc=%s %s >&%s' % (execcommand, self['da.restart.fname'], join(self.opts, ''), logfile) 
Peters, Wouter's avatar
test    
Peters, Wouter committed
629
630

            # write and submit 
631
632
            self.DaPlatForm.write_job(jobfile, template, jobid)
            jobid = self.DaPlatForm.submit_job(jobfile, joblog=logfile) 
633
634
635
636
        else:
            logging.info('Final date reached, no new cycle started')


637
def start_logger(level=logging.INFO):
638
639
640
641
    """ start the logging of messages to screen"""

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

karolina's avatar
karolina committed
642
643
644
    logging.basicConfig(level=level,
                        format=' [%(levelname)-7s] (%(asctime)s) py-%(module)-20s : %(message)s',
                        datefmt='%Y-%m-%d %H:%M:%S')
645

646
def parse_options():
647
648
649
650
651
652
653
654
655
656
657
658
659
    """ 
    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
    ========  =======

    """
660
661
662

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

karolina's avatar
karolina committed
663
664
    opts = []
    args = []
665
    try:                                
666
        opts, args = getopt.gnu_getopt(sys.argv[1:], "-rv")
667
    except getopt.GetoptError, msg:           
karolina's avatar
karolina committed
668
        logging.error('%s' % msg)
669
670
671
        sys.exit(2)      

    for options in opts:
karolina's avatar
karolina committed
672
        options = options[0].lower()
673
674
675
676
        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
677
            logging.root.setLevel(logging.DEBUG)
678

679
    if opts: 
680
        optslist = [item[0] for item in opts] #LU ze co same minusy zwroci?
681
    else:
karolina's avatar
karolina committed
682
        optslist = []
683

684
685
# Parse arguments and return as dictionary

karolina's avatar
karolina committed
686
    arguments = {}
687
688
689
690
691
692
693
694
    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
695
696
            logging.error('%s' % 'Argument passed without description (%s)' % item)
            raise getopt.GetoptError, arg
697

karolina's avatar
karolina committed
698
        arguments[key] = arg
699
700


701
    return optslist, arguments
702

703
def validate_opts_args(opts, args):
704
705
706
707
708
    """ 
 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.  
 
    """
709
    if not args.has_key("rc"):
karolina's avatar
karolina committed
710
711
712
        msg = "There is no rc-file specified on the command line. Please use rc=yourfile.rc"
        logging.error(msg)
        raise IOError, msg
713
    elif not os.path.exists(args['rc']):
karolina's avatar
karolina committed
714
715
716
        msg = "The specified rc-file (%s) does not exist " % args['rc'] 
        logging.error(msg)
        raise IOError, msg
717
718
719
720
721
722
723
724
725

    # 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
726
    return opts, args
727
728
729


if __name__ == "__main__":
730
731

    sys.path.append('../../')
732
    opts, args = parse_options()
733
734
    print opts
    print args
735