databases.py 19.3 KB
Newer Older
André Anjos's avatar
André Anjos committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
#!/usr/bin/env python
# vim: set fileencoding=utf-8 :

###############################################################################
#                                                                             #
# Copyright (c) 2016 Idiap Research Institute, http://www.idiap.ch/           #
# Contact: beat.support@idiap.ch                                              #
#                                                                             #
# This file is part of the beat.cmdline module of the BEAT platform.          #
#                                                                             #
# Commercial License Usage                                                    #
# Licensees holding valid commercial BEAT licenses may use this file in       #
# accordance with the terms contained in a written agreement between you      #
# and Idiap. For further information contact tto@idiap.ch                     #
#                                                                             #
# Alternatively, this file may be used under the terms of the GNU Affero      #
# Public License version 3 as published by the Free Software and appearing    #
# in the file LICENSE.AGPL included in the packaging of this file.            #
# The BEAT platform is distributed in the hope that it will be useful, but    #
# WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY  #
# or FITNESS FOR A PARTICULAR PURPOSE.                                        #
#                                                                             #
# You should have received a copy of the GNU Affero Public License along      #
# with the BEAT platform. If not, see http://www.gnu.org/licenses/.           #
#                                                                             #
###############################################################################


"""Usage:
  %(prog)s databases list [--remote]
  %(prog)s databases check [<name>]...
  %(prog)s databases pull [--force] [<name>]...
  %(prog)s databases push [--force] [--dry-run] [<name>]...
  %(prog)s databases diff <name>
  %(prog)s databases status
  %(prog)s databases version <name>
  %(prog)s databases index [--list | --delete | --checksum] [<name>]...
  %(prog)s databases view [--exclude=<output>] <set_name>
  %(prog)s databases --help


Arguments:
  <name>           Database name formated as "<database>/<version>"
  <set_name>       Set formatted as "<database>/<version>/<protocol>/<set>"


Commands:
  list      Lists all the databases available on the platform
  check     Checks a local database for validity
  pull      Downloads the specified databases from the server
  push      Uploads databases to the server (must provide a valid admin token)
  diff      Shows changes between the local database and the remote version
  status    Shows (editing) status for all available databases
  version   Creates a new version of an existing database
  index     Indexes all outputs (of all sets) of a database.
  view      View the data of the specified dataset.


Options:
  --help              Display this screen
  --remote            Only acts on the remote copy of the database
  --exclude=<output>  When viewing, excludes this output
  --list              List index files matching output if they exist
  --delete            Delete index files matching output if they exist (also,
                      recursively deletes empty directories)
  --checksum          Checksums index files


Examples:

  To list all existing databases on your local prefix:

    $ %(prog)s db list

  To view the contents of a specific set

    $ %(prog)s db view simple/1/protocol/set

  To index the contents of a database

    $ %(prog)s db index simple/1

  To index the contents of a protocol on a database

    $ %(prog)s db index simple/1/double

  To index the contents of a set in a protocol on a database

    $ %(prog)s db index simple/1/double/double
"""

import os
import glob
94
95
import zmq.green as zmq

André Anjos's avatar
André Anjos committed
96
97
98
99
100
101
102
103
104
105
import logging
logger = logging.getLogger()

import simplejson

from beat.core.hash import hash, toPath, hashFileContents
from beat.core.utils import NumpyJSONEncoder
from beat.core.database import Database
from beat.core.dataformat import DataFormat
from beat.core.data import load_data_index
106
107
108
from beat.core import dock
from beat.core import inputs
from beat.core import utils
André Anjos's avatar
André Anjos committed
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246

from . import common


def pull(webapi, prefix, names, force, indentation, format_cache):
  """Copies databases (and required dataformats) from the server.

  Parameters:

    webapi (object): An instance of our WebAPI class, prepared to access the
      BEAT server of interest

    prefix (str): A string representing the root of the path in which the user
      objects are stored

    names (list): A list of strings, each representing the unique relative path
      of the objects to retrieve or a list of usernames from which to retrieve
      objects. If the list is empty, then we pull all available objects of a
      given type. If no user is set, then pull all public objects of a given
      type.

    force (bool): If set to ``True``, then overwrites local changes with the
      remotely retrieved copies.

    indentation (int): The indentation level, useful if this function is called
      recursively while downloading different object types. This is normally
      set to ``0`` (zero).

    format_cache (dict): A dictionary containing all dataformats already
      downloaded.


  Returns:

    int: Indicating the exit status of the command, to be reported back to the
      calling process. This value should be zero if everything works OK,
      otherwise, different than zero (POSIX compliance).

  """

  from .dataformats import pull as dataformats_pull

  status, names = common.pull(webapi, prefix, 'database', names,
          ['declaration', 'code', 'description'], force, indentation)

  # see what dataformats one needs to pull
  indent = indentation * ' '

  dataformats = []
  for name in names:
    obj = Database(prefix, name)
    dataformats.extend(obj.dataformats.keys())

  # downloads any formats to which we depend on
  df_status = dataformats_pull(webapi, prefix, dataformats, force,
          indentation + 2, format_cache)

  return status + df_status


def index_output(configuration, names, ls, delete, checksum):

  names = common.make_up_local_list(configuration.path, 'database', names)

  retcode = 0

  for database_name in names:

    parts = database_name.split('/')

    if len(parts) == 2:
      db_name         = os.path.join(*parts[:2])
      protocol_filter = None
      set_filter      = None
      if ls:
        logger.info("Listing database %s indexes...", database_name)
      elif delete:
        logger.info("Deleting database %s indexes...", database_name)
      elif checksum:
        logger.info("Checksumming database %s indexes...", database_name)
      else:
        logger.info("Indexing database %s...", database_name)

    elif len(parts) == 3:
      db_name         = os.path.join(*parts[:2])
      protocol_filter = parts[2]
      set_filter      = None
      if ls:
        logger.info("Listing database protocol %s indexes...", database_name)
      elif delete:
        logger.info("Deleting database protocol %s indexes...", database_name)
      elif checksum:
        logger.info("Checksumming database protocol %s indexes...",
            database_name)
      else:
        logger.info("Indexing database protocol %s...", database_name)

    elif len(parts) == 4:
      db_name         = os.path.join(*parts[:2])
      protocol_filter = parts[2]
      set_filter      = parts[3]
      if ls:
        logger.info("Listing database set %s indexes...", database_name)
      elif delete:
        logger.info("Deleting database set %s indexes...", database_name)
      elif checksum:
        logger.info("Checksumming database set %s indexes...", database_name)
      else:
        logger.info("Indexing database set %s...", database_name)

    else:
      logger.error("Database specification should have the format " \
              "`<database>/<version>/[<protocol>/[<set>]]', the value you " \
              "passed (%s) is not valid", (dataset_name))
      retcode += 1
      continue

    dataformat_cache = {}
    database = Database(configuration.path, db_name, dataformat_cache)
    if not database.valid:
      logger.error("Failed to load the database `%s':", db_name)
      for e in database.errors: logger.error('  * %s', e)
      retcode += 1
      continue

    protocols = database.protocol_names

    if protocol_filter is not None:

      if protocol_filter not in protocols:
        logger.error("The database `%s' does not have the protocol `%s' - " \
                "choose one of `%s'", db_name, protocol_filter,
                ', '.join(protocols))
        retcode += 1
        continue

      protocols = [protocol_filter]

247
248
249
    host = dock.Host()
    host.setup(raise_on_errors=True)

André Anjos's avatar
André Anjos committed
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
    for protocol_name in protocols:

      sets = database.set_names(protocol_name)

      if set_filter is not None:
        if set_filter not in sets:
          logger.error("The database/protocol `%s/%s' does not have the " \
                  "set `%s' - choose one of `%s'", db_name, protocol_name,
                  set_filter, ', '.join(sets))
          retcode += 1
          continue

        sets = [z for z in sets if z == set_filter]

      for set_name in sets:

266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
        zmq_context = zmq.Context()
        db_socket = zmq_context.socket(zmq.PAIR)
        db_address = 'tcp://' + host.ip
        port = db_socket.bind_to_random_port(db_address)
        db_address += ':%d' % port

        input_list = inputs.InputList()

        input_group = inputs.RemoteInputGroup(set_name, restricted_access=True, socket=db_socket)
        input_list.add(input_group)

        db_set = database.set(protocol_name, set_name)

        db_configuration = {
          'inputs': {},
          'channel': set_name,
        }

        for output_name, dataformat_name in db_set['outputs'].items():
          input = inputs.RemoteInput(output_name, database.dataformats[dataformat_name], db_socket)
          input_group.add(input)

          db_configuration['inputs'][output_name] = dict(
            database=db_name,
            protocol=protocol_name,
            set=set_name,
            output=output_name,
            channel=set_name
          )
André Anjos's avatar
André Anjos committed
295

296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
        db_tempdir = utils.temporary_directory()

        with open(os.path.join(db_tempdir, 'configuration.json'), 'wb') as f:
          simplejson.dump(db_configuration, f, indent=4)

        tmp_prefix = os.path.join(db_tempdir, 'prefix')
        if not os.path.exists(tmp_prefix):
          os.makedirs(tmp_prefix)

        database.export(tmp_prefix)

        json_path = os.path.join(tmp_prefix, 'databases', db_name + '.json')

        with open(json_path, 'r') as f:
          db_data = simplejson.load(f)

        database_path = db_data['root_folder']
        db_data['root_folder'] = os.path.join('/databases', db_name)

        with open(json_path, 'w') as f:
          simplejson.dump(db_data, f, indent=4)

        try:
          db_envkey = host.db2docker([db_name])
        except:
          raise RuntimeError("No environment found for the database `%s' " \
              "- available environments are %s" % (
                db_name,
                ", ".join(host.db_environments.keys())))

        tmp_dir = os.path.join('/tmp', os.path.basename(db_tempdir))
        db_cmd = ['databases_provider', db_address, tmp_dir]

        volumes = {}
        volumes[database_path] = {
          'bind': os.path.join('/databases', db_name),
          'mode': 'ro',
        }

        # Note: we only support one databases image loaded at the same time
        db_process = dock.Popen(
          host,
          db_envkey,
          command=db_cmd,
          tmp_archive=db_tempdir,
          volumes=volumes
        )
André Anjos's avatar
André Anjos committed
343
344
345
346

        index_filenames = []
        previous_data_indices = []

347
348
        for output_name in db_set['outputs'].keys():
          index_hash = database.hash_output(protocol_name, set_name, output_name)
André Anjos's avatar
André Anjos committed
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
          index_filename = os.path.join(configuration.cache,
              toPath(index_hash, '.index'))

          if ls:
            basename = os.path.splitext(index_filename)[0]
            for g in glob.glob(basename + '.*'): logger.info(g)

          elif delete:

            basename = os.path.splitext(index_filename)[0]
            for g in glob.glob(basename + '.*'):
              logger.info("removing `%s'...", g)
              os.unlink(g)
            common.recursive_rmdir_if_empty(os.path.dirname(basename),
                configuration.cache)

          elif checksum:
            assert load_data_index(index_filename, index_filename)
            logger.info("index for `%s' can be loaded and checksums",
                index_filename)

          else:

            index_filenames.append(index_filename)
            previous_data_indices.append(None)

            logger.info("Indexing database `%s', protocol `%s', set `%s', " \
                    "output `%s' -> `%s'", db_name, protocol_name, set_name,
377
                    output_name, index_filename)
André Anjos's avatar
André Anjos committed
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425

            if os.path.exists(index_filename):
              logger.extra("Overwriting existing index file `%s'",
                  index_filename)
              os.remove(index_filename)

            index_dir = os.path.dirname(index_filename)
            if not os.path.exists(index_dir):
              logger.extra("Creating directory `%s'", index_dir)
              os.makedirs(index_dir)

            # Create empty lock file to indicate that this cache file *must*
            # not be cleaned
            lock_filename = os.path.join(configuration.cache,
                toPath(index_hash, '.lock'))
            with open(lock_filename, 'a'): os.utime(lock_filename, None)

        if not (ls or delete or checksum): #we're indexing

          try:

            while input_group.hasMoreData():

              input_group.next()

              for i, (input, index_filename, previous_data_index) in \
                      enumerate(zip(input_group, index_filenames,
                          previous_data_indices)):

                with open(index_filename, 'at') as indexfile:
                  if input.data_index != previous_data_index:
                    indexfile.write("%d %d\n" % (
                      input.data_index,
                      input.data_index_end,
                      )
                    )
                  previous_data_indices[i] = input.data_index

                # creates the checksums for all indexes
                chksum = hashFileContents(index_filename)
                with open(index_filename + '.checksum', 'wt') as f:
                  f.write(chksum)

          except Exception as e:
            logger.critical("Failed to retrieve the next data: %s", e)
            retcode += 1
            continue

426
427
428
429
430
431
432
433
        db_process.kill()
        db_process.wait()
        db_process.rm()

        db_socket.setsockopt(zmq.LINGER, 0)
        db_socket.close()
        zmq_context.term()

André Anjos's avatar
André Anjos committed
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
  return retcode


def view_output(configuration, dataset_name, excluded_outputs=''):

  parts = dataset_name.split('/')

  if len(parts) == 4:
    db_name       = os.path.join(*parts[:2])
    protocol_name = parts[2]
    set_name      = parts[3]
  else:
    logger.error("Database view specification should have the format " \
            "`<database>/<version>/<protocol>/<set>', the value you " \
            "passed (%s) is not valid", dataset_name)
    return 1

  excluded_outputs = map(lambda x: x.strip(), excluded_outputs.split(','))

  dataformat_cache = {}
  database = Database(configuration.path, db_name, dataformat_cache)
  if not database.valid:
    logger.error("Failed to load the database `%s':", db_name)
    for e in database.errors: logger.error('  * %s', e)
    return 1

  if protocol_name not in database.protocol_names:
    logger.error("Unknown protocol `%s' for database `%s'. Valid protocols " \
            "are `%s'", protocol_name, db_name,
            ', '.join(database.protocol_names))
    return 1

  if set_name not in database.set_names(protocol_name):
    set_names = database.set_names(protocol_name)
    logger.error("Unknown set `%s' for database `%s' and protocol `%s'. " \
            "Valid set names for this protocol are `%s'", set_name, db_name,
                protocol_name, ', '.join(set_names))
    return 1

  view = database.view(protocol_name, set_name)

  try:
    view.prepare_outputs()
    view.setup()
  except Exception:
    logger.error("Failed to setup the set `%s/%s/%s'", db_name, protocol_name,
        set_name)
    import traceback
    for e in traceback.format_exc().split('\n'): logger.error('  %s', e)
    return 1

  # Setup the inputs
  input_group = view.input_group(exclude_outputs=excluded_outputs)

  # Display the data
  while input_group.hasMoreData():

    try:
      input_group.next()
    except Exception as e:
      logger.error("Failed to retrieve the next data: %s", e)
      return 1

    logger.extra(80 * '-')

    for input in input_group:
      header = str(input.name)
      if input.data_index == input.data_index_end:
        header += '[%d]' % input.data_index
      else:
        header += '[%d:%d]' % (input.data_index, input.data_index_end)
      header += ": "

      json_data = input.data.as_dict()
      for name, value in json_data.items():
        json_data[name] = common.stringify(value)
        json_data_ = simplejson.dumps(json_data[name], indent=2,
                cls=NumpyJSONEncoder).\
                replace('"BEAT_LIST_DELIMITER[', '[')\
                .replace(']BEAT_LIST_DELIMITER"', ']')\
                .replace('"...",', '...')\
                .replace('"BEAT_LIST_SIZE(', '(')\
                .replace(')BEAT_LIST_SIZE"', ')')
        print(header + json_data_)

  return 0


def process(args):

  if args['list']:
    if args['--remote']:
      with common.make_webapi(args['config']) as webapi:
        return common.display_remote_list(webapi, 'database')
    else:
      return common.display_local_list(args['config'].path, 'database')

  elif args['check']:
    return common.check(args['config'].path, 'database', args['<name>'])

  elif args['pull']:
    with common.make_webapi(args['config']) as webapi:
      return pull(webapi, args['config'].path, args['<name>'],
              args['--force'], 0, {})

  elif args['push']:
    with common.make_webapi(args['config']) as webapi:
      return common.push(webapi, args['config'].path, 'database',
              args['<name>'], ['name', 'declaration', 'code', 'description'],
              {}, args['--force'], args['--dry-run'], 0)

  elif args['diff']:
    with common.make_webapi(args['config']) as webapi:
      return common.diff(webapi, args['config'].path, 'database',
              args['<name>'][0], ['declaration', 'code', 'description'])

  elif args['status']:
    with common.make_webapi(args['config']) as webapi:
      return common.status(webapi, args['config'].path, 'database')[0]

  elif args['version']:
    return common.new_version(args['config'].path, 'database',
            args['<name>'][0])

  elif args['view']:
    if args['--exclude']:
      return view_output(args['config'], args['<set_name>'], args['--exclude'])
    else:
      return view_output(args['config'], args['<set_name>'])

  elif args['index']:
    return index_output(args['config'], args['<name>'], args['--list'],
            args['--delete'], args['--checksum'])

  # Should not happen
  logger.error("unrecognized `databases' subcommand")
  return 1