data.py 22.9 KB
Newer Older
Markus Scheidgen's avatar
Markus Scheidgen 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
# Copyright 2018 Markus Scheidgen
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#   http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an"AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

"""
This module comprises a set of persistent document classes that hold all user related
data. These are information about users, their uploads and datasets, the associated
calculations, and files


.. autoclass:: Calc
    :members:
.. autoclass:: Upload
    :members:
"""

27
from typing import List, Any, ContextManager, Tuple, Generator
Markus Scheidgen's avatar
Markus Scheidgen committed
28
from datetime import datetime
29
from elasticsearch.exceptions import NotFoundError
Markus Scheidgen's avatar
Markus Scheidgen committed
30
from mongoengine import StringField, BooleanField, DateTimeField, DictField, IntField
Markus Scheidgen's avatar
Markus Scheidgen committed
31
import logging
32
import base64
Markus Scheidgen's avatar
Markus Scheidgen committed
33
import time
34
from structlog import wrap_logger
35
from contextlib import contextmanager
Markus Scheidgen's avatar
Markus Scheidgen committed
36

37
from nomad import config, utils, coe_repo
Markus Scheidgen's avatar
Markus Scheidgen committed
38
from nomad.files import UploadFile, ArchiveFile, ArchiveLogFile, File
Markus Scheidgen's avatar
Markus Scheidgen committed
39
from nomad.repo import RepoCalc
40
from nomad.processing.base import Proc, Chord, process, task, PENDING, SUCCESS, FAILURE, RUNNING
Markus Scheidgen's avatar
Markus Scheidgen committed
41
from nomad.parsing import parsers, parser_dict
Markus Scheidgen's avatar
Markus Scheidgen committed
42
from nomad.normalizing import normalizers
Markus Scheidgen's avatar
Markus Scheidgen committed
43
from nomad.utils import lnr
Markus Scheidgen's avatar
Markus Scheidgen committed
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


class NotAllowedDuringProcessing(Exception): pass


class Calc(Proc):
    """
    Instances of this class represent calculations. This class manages the elastic
    search index entry, files, and archive for the respective calculation.

    It also contains the calculations processing and its state.

    The attribute list, does not include the various repository properties generated
    while parsing, including ``program_name``, ``program_version``, etc.

    Attributes:
        archive_id: the hash based archive id of the calc
        parser: the name of the parser used to process this calc
        upload_id: the id of the upload used to create this calculation
        mainfile: the mainfile (including path in upload) that was used to create this calc
        mainfile_tmp_path: path to the mainfile extracted for processing
    """
    archive_id = StringField(primary_key=True)
    upload_id = StringField()
    mainfile = StringField()
    parser = StringField()
    mainfile_tmp_path = StringField()

    meta: Any = {
        'indices': [
74
            'upload_id', 'mainfile', 'code', 'parser', 'status'
Markus Scheidgen's avatar
Markus Scheidgen committed
75
76
77
78
79
80
81
        ]
    }

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self._parser_backend = None
        self._upload = None
82
        self._calc_proc_logwriter = None
Markus Scheidgen's avatar
Markus Scheidgen committed
83
        self._calc_proc_logfile = None
84
        self._calc_proc_logwriter_ctx: ContextManager = None
Markus Scheidgen's avatar
Markus Scheidgen committed
85
86
87
88
89

    @classmethod
    def get(cls, id):
        return cls.get_by_id(id, 'archive_id')

Markus Scheidgen's avatar
Markus Scheidgen committed
90
91
92
93
    @property
    def mainfile_file(self) -> File:
        return File(self.mainfile_tmp_path)

94
95
96
97
98
99
    @property
    def upload(self) -> 'Upload':
        if not self._upload:
            self._upload = Upload.get(self.upload_id)
        return self._upload

Markus Scheidgen's avatar
Markus Scheidgen committed
100
101
102
103
    def delete(self):
        """
        Delete this calculation and all associated data. This includes all files,
        the archive, and this search index entry.
104
        TODO is this needed? Or do we always delete hole uploads in bulk.
Markus Scheidgen's avatar
Markus Scheidgen committed
105
106
107
        """
        # delete the archive
        if self.archive_id is not None:
108
            ArchiveFile(self.archive_id).delete()
Markus Scheidgen's avatar
Markus Scheidgen committed
109
110

        # delete the search index entry
111
112
113
114
115
116
        try:
            elastic_entry = RepoCalc.get(self.archive_id)
            if elastic_entry is not None:
                elastic_entry.delete()
        except NotFoundError:
            pass
Markus Scheidgen's avatar
Markus Scheidgen committed
117
118
119
120
121
122
123
124
125

        # delete this mongo document
        super().delete()

    def get_logger(self, **kwargs):
        upload_hash, calc_hash = self.archive_id.split('/')
        logger = super().get_logger()
        logger = logger.bind(
            upload_id=self.upload_id, mainfile=self.mainfile,
126
            upload_hash=upload_hash, calc_hash=calc_hash,
Markus Scheidgen's avatar
Markus Scheidgen committed
127
            archive_id=self.archive_id, **kwargs)
128

Markus Scheidgen's avatar
Markus Scheidgen committed
129
130
        return logger

131
132
133
134
135
136
137
138
    def get_calc_logger(self, **kwargs):
        """
        Returns a wrapped logger that additionally saves all entries to the calculation
        processing log in the archive.
        """
        logger = self.get_logger(**kwargs)

        if self._calc_proc_logwriter is None:
Markus Scheidgen's avatar
Markus Scheidgen committed
139
            self._calc_proc_logfile = ArchiveLogFile(self.archive_id)
140
141
            self._calc_proc_logwriter_ctx = self._calc_proc_logfile.open('wt')
            self._calc_proc_logwriter = self._calc_proc_logwriter_ctx.__enter__()  # pylint: disable=E1101
142

143
        def save_to_calc_log(logger, method_name, event_dict):
144
145
146
147
148
149
150
151
152
153
154
            program = event_dict.get('normalizer', 'parser')
            event = event_dict.get('event', '')
            entry = '[%s] %s: %s' % (method_name, program, event)
            if len(entry) > 120:
                self._calc_proc_logwriter.write(entry[:120])
                self._calc_proc_logwriter.write('...')
            else:
                self._calc_proc_logwriter.write(entry)
            self._calc_proc_logwriter.write('\n')
            return event_dict

155
        return wrap_logger(logger, processors=[save_to_calc_log])
156

Markus Scheidgen's avatar
Markus Scheidgen committed
157
158
159
160
161
162
163
164
165
166
167
168
169
170
    @property
    def json_dict(self):
        """ A json serializable dictionary representation. """
        data = {
            'archive_id': self.archive_id,
            'mainfile': self.mainfile,
            'upload_id': self.upload_id,
            'parser': self.parser
        }
        data.update(super().json_dict)
        return {key: value for key, value in data.items() if value is not None}

    @process
    def process(self):
171
        logger = self.get_logger()
172
        if self.upload is None:
173
            logger.error('calculation upload does not exist')
Markus Scheidgen's avatar
Markus Scheidgen committed
174
175
176
177
178
179

        try:
            self.parsing()
            self.normalizing()
            self.archiving()
        finally:
Markus Scheidgen's avatar
Markus Scheidgen committed
180
            # close loghandler that was not closed due to failures
181
            try:
182
183
184
                if self._calc_proc_logwriter is not None:
                    self._calc_proc_logwriter.close()
                    self._calc_proc_logwriter = None
185
186
187
188
            except Exception as e:
                logger.error('could not close calculation proc log', exc_info=e)

            # inform parent proc about completion
189
            self.upload.completed_child()
Markus Scheidgen's avatar
Markus Scheidgen committed
190
191
192

    @task
    def parsing(self):
193
194
        context = dict(parser=self.parser, step=self.parser)
        logger = self.get_calc_logger(**context)
195
        parser = parser_dict[self.parser]
Markus Scheidgen's avatar
Markus Scheidgen committed
196

197
        with utils.timer(logger, 'parser executed', input_size=self.mainfile_file.size):
198
            self._parser_backend = parser.run(self.mainfile_tmp_path, logger=logger)
Markus Scheidgen's avatar
Markus Scheidgen committed
199

200
201
202
203
204
205
        self._parser_backend.openNonOverlappingSection('section_calculation_info')
        self._parser_backend.addValue('upload_id', self.upload_id)
        self._parser_backend.addValue('archive_id', self.archive_id)
        self._parser_backend.addValue('main_file', self.mainfile)
        self._parser_backend.addValue('parser_name', self.parser)

Markus Scheidgen's avatar
Markus Scheidgen committed
206
        if self._parser_backend.status[0] != 'ParseSuccess':
207
            logger.error(self._parser_backend.status[1])
Markus Scheidgen's avatar
Markus Scheidgen committed
208
            error = self._parser_backend.status[1]
209
            self._parser_backend.addValue('parse_status', 'ParseFailure')
210
            self.fail(error, level=logging.DEBUG, **context)
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
        else:
            self._parser_backend.addValue('parse_status', 'ParseSuccess')

        self._parser_backend.closeNonOverlappingSection('section_calculation_info')

        self.add_processor_info(self.parser)

    @contextmanager
    def use_parser_backend(self, processor_name):
        self._parser_backend.reset_status()
        yield self._parser_backend
        self.add_processor_info(processor_name)

    def add_processor_info(self, processor_name: str) -> None:
        self._parser_backend.openContext('/section_calculation_info/0')
        self._parser_backend.openNonOverlappingSection('section_archive_processing_info')
        self._parser_backend.addValue('archive_processor_name', processor_name)

        if self._parser_backend.status[0] == 'ParseSuccess':
            warnings = getattr(self._parser_backend, '_warnings', [])
            if len(warnings) > 0:
                self._parser_backend.addValue('archive_processor_status', 'WithWarnings')
                self._parser_backend.addValue('archive_processor_warning_number', len(warnings))
                self._parser_backend.addArrayValues('archive_processor_warnings', [str(warning) for warning in warnings])
            else:
                self._parser_backend.addValue('archive_processor_status', 'Success')
        else:
            errors = self._parser_backend.status[1]
            self._parser_backend.addValue('archive_processor_error', str(errors))

        self._parser_backend.closeNonOverlappingSection('section_archive_processing_info')
        self._parser_backend.closeContext('/section_calculation_info/0')
Markus Scheidgen's avatar
Markus Scheidgen committed
243
244
245
246
247

    @task
    def normalizing(self):
        for normalizer in normalizers:
            normalizer_name = normalizer.__name__
248
249
            context = dict(normalizer=normalizer_name, step=normalizer_name)
            logger = self.get_calc_logger(**context)
Markus Scheidgen's avatar
Markus Scheidgen committed
250
251

            with utils.timer(
252
                    logger, 'normalizer executed', input_size=self.mainfile_file.size):
253
254
                with self.use_parser_backend(normalizer_name) as backend:
                    normalizer(backend).normalize(logger=logger)
Markus Scheidgen's avatar
Markus Scheidgen committed
255

256
257
            failed = self._parser_backend.status[0] != 'ParseSuccess'
            if failed:
258
                logger.error(self._parser_backend.status[1])
Markus Scheidgen's avatar
Markus Scheidgen committed
259
                error = self._parser_backend.status[1]
260
                self.fail(error, level=logging.WARNING, **context)
261
262
263
264
                break
            else:
                logger.debug(
                    'completed normalizer successfully', normalizer=normalizer_name)
Markus Scheidgen's avatar
Markus Scheidgen committed
265
266
267

    @task
    def archiving(self):
268
269
270
271
272
        logger = self.get_logger()

        upload_hash, calc_hash = self.archive_id.split('/')
        additional = dict(
            mainfile=self.mainfile,
273
            upload_time=self.upload.upload_time,
274
275
            staging=True,
            restricted=False,
276
277
            user_id=self.upload.user_id,
            aux_files=list(self.upload.upload_file.get_siblings(self.mainfile)))
278
279

        with utils.timer(logger, 'indexed', step='index'):
280
281
282
283
284
285
            # persist to elastic search
            RepoCalc.create_from_backend(
                self._parser_backend,
                additional=additional,
                upload_hash=upload_hash,
                calc_hash=calc_hash,
286
                upload_id=self.upload_id).persist()
287

Markus Scheidgen's avatar
Markus Scheidgen committed
288
289
290
291
        with utils.timer(
                logger, 'archived', step='archive',
                input_size=self.mainfile_file.size) as log_data:

292
            # persist the archive
Markus Scheidgen's avatar
Markus Scheidgen committed
293
294
            archive_file = ArchiveFile(self.archive_id)
            with archive_file.write_archive_json() as out:
295
296
                self._parser_backend.write_json(out, pretty=True)

Markus Scheidgen's avatar
Markus Scheidgen committed
297
298
299
300
301
302
303
            log_data.update(archive_size=archive_file.size)

        # close loghandler
        if self._calc_proc_logwriter is not None:
            with utils.timer(
                    logger, 'archived log', step='archive_log',
                    input_size=self.mainfile_file.size) as log_data:
304
                self._calc_proc_logwriter_ctx.__exit__(None, None, None)  # pylint: disable=E1101
305
                self._calc_proc_logwriter = None
Markus Scheidgen's avatar
Markus Scheidgen committed
306

Markus Scheidgen's avatar
Markus Scheidgen committed
307
308
                log_data.update(log_size=self._calc_proc_logfile.size)

Markus Scheidgen's avatar
Markus Scheidgen committed
309

310
class Upload(Chord):
Markus Scheidgen's avatar
Markus Scheidgen committed
311
312
313
314
315
316
    """
    Represents uploads in the databases. Provides persistence access to the files storage,
    and processing state.

    Attributes:
        name: optional user provided upload name
317
        local_path: optional local path, e.g. for files that are already somewhere on the server
Markus Scheidgen's avatar
Markus Scheidgen committed
318
319
320
321
322
323
        additional_metadata: optional user provided additional meta data
        upload_id: the upload id generated by the database
        in_staging: true if the upload is still in staging and can be edited by the uploader
        is_private: true if the upload and its derivitaves are only visible to the uploader
        upload_time: the timestamp when the system realised the upload
        upload_hash: the hash of the uploaded file
324
        user_id: the id of the user that created this upload
Markus Scheidgen's avatar
Markus Scheidgen committed
325
326
327
328
329
330
    """
    id_field = 'upload_id'

    upload_id = StringField(primary_key=True)

    name = StringField(default=None)
331
    local_path = StringField(default=None)
Markus Scheidgen's avatar
Markus Scheidgen committed
332
333
334
335
336
337
338
339
    additional_metadata = DictField(default=None)

    in_staging = BooleanField(default=True)
    is_private = BooleanField(default=False)

    upload_time = DateTimeField()
    upload_hash = StringField(default=None)

340
    user_id = StringField(required=True)
341
342
    upload_url = StringField(default=None)
    upload_command = StringField(default=None)
Markus Scheidgen's avatar
Markus Scheidgen committed
343

344
345
    coe_repo_upload_id = IntField(default=None)

346
347
    _initiated_parsers = IntField(default=-1)

Markus Scheidgen's avatar
Markus Scheidgen committed
348
349
    meta: Any = {
        'indexes': [
350
            'upload_hash', 'user_id', 'status'
Markus Scheidgen's avatar
Markus Scheidgen committed
351
352
353
354
355
        ]
    }

    def __init__(self, **kwargs):
        super().__init__(**kwargs)
356
        self._upload_file = None
Markus Scheidgen's avatar
Markus Scheidgen committed
357
358
359
360
361
362

    @classmethod
    def get(cls, id):
        return cls.get_by_id(id, 'upload_id')

    @classmethod
363
    def user_uploads(cls, user: coe_repo.User) -> List['Upload']:
Markus Scheidgen's avatar
Markus Scheidgen committed
364
        """ Returns all uploads for the given user. Currently returns all uploads. """
365
        return cls.objects(user_id=str(user.user_id), in_staging=True)
Markus Scheidgen's avatar
Markus Scheidgen committed
366
367
368
369
370
371
372
373
374
375
376
377
378
379

    def get_logger(self, **kwargs):
        logger = super().get_logger()
        logger = logger.bind(upload_id=self.upload_id, **kwargs)
        return logger

    def delete(self):
        logger = self.get_logger(task='delete')

        if not (self.completed or self.is_stale or self.current_task == 'uploading'):
            raise NotAllowedDuringProcessing()

        with lnr(logger, 'delete upload file'):
            try:
380
                UploadFile(self.upload_id, local_path=self.local_path).delete()
Markus Scheidgen's avatar
Markus Scheidgen committed
381
382
383
384
385
386
387
388
389
            except KeyError:
                if self.current_task == 'uploading':
                    logger.debug(
                        'Upload exist, but file does not exist. '
                        'It was probably aborted and deleted.')
                else:
                    logger.debug('Upload exist, but uploaded file does not exist.')

        with lnr(logger, 'deleting calcs'):
390
            # delete archive files
391
            ArchiveFile.delete_archives(upload_hash=self.upload_hash)
392
393

            # delete repo entries
Markus Scheidgen's avatar
Markus Scheidgen committed
394
            RepoCalc.delete_upload(upload_id=self.upload_id)
395
396
397

            # delete calc processings
            Calc.objects(upload_id=self.upload_id).delete()
Markus Scheidgen's avatar
Markus Scheidgen committed
398
399
400
401

        with lnr(logger, 'deleting upload'):
            super().delete()

402
403
    @classmethod
    def _external_objects_url(cls, url):
404
        """ Replaces the given internal object storage url with an URL that allows
405
406
            external access.
        """
407
        return 'http://%s:%s%s%s' % (config.services.api_host, config.services.api_port, config.services.api_base_path, url)
408

Markus Scheidgen's avatar
Markus Scheidgen committed
409
410
411
412
413
414
    @classmethod
    def create(cls, **kwargs) -> 'Upload':
        """
        Creates a new upload for the given user, a user given name is optional.
        It will populate the record with a signed url and pending :class:`UploadProc`.
        The upload will be already saved to the database.
415
416

        Arguments:
417
            user (coe_repo.User): The user that created the upload.
Markus Scheidgen's avatar
Markus Scheidgen committed
418
        """
419
        user: coe_repo.User = kwargs['user']
420
421
422
        del(kwargs['user'])
        if 'upload_id' not in kwargs:
            kwargs.update(upload_id=utils.create_uuid())
423
        kwargs.update(user_id=str(user.user_id))
Markus Scheidgen's avatar
Markus Scheidgen committed
424
        self = super().create(**kwargs)
425

426
        basic_auth_token = base64.b64encode(b'%s:' % user.get_auth_token()).decode('utf-8')
427
428

        self.upload_url = cls._external_objects_url('/uploads/%s/file' % self.upload_id)
Markus Scheidgen's avatar
Markus Scheidgen committed
429
        self.upload_command = 'curl -H "Authorization: Basic %s" "%s" --upload-file local_file' % (
430
431
            basic_auth_token, self.upload_url)

Markus Scheidgen's avatar
Markus Scheidgen committed
432
        self._continue_with('uploading')
433

Markus Scheidgen's avatar
Markus Scheidgen committed
434
435
436
437
438
439
440
441
442
        return self

    @property
    def is_stale(self) -> bool:
        if self.current_task == 'uploading' and self.upload_time is None:
            return (datetime.now() - self.create_time).days > 1
        else:
            return False

Markus Scheidgen's avatar
Markus Scheidgen committed
443
    def unstage(self):
444
        self.get_logger().info('unstage')
Markus Scheidgen's avatar
Markus Scheidgen committed
445
        self.in_staging = False
446
        RepoCalc.unstage(upload_id=self.upload_id)
447
        coe_repo.add_upload(self, restricted=False)  # TODO allow users to choose restricted
Markus Scheidgen's avatar
Markus Scheidgen committed
448
449
        self.save()

Markus Scheidgen's avatar
Markus Scheidgen committed
450
451
452
453
454
    @property
    def json_dict(self) -> dict:
        """ A json serializable dictionary representation. """
        data = {
            'name': self.name,
455
            'local_path': self.local_path,
Markus Scheidgen's avatar
Markus Scheidgen committed
456
457
            'additional_metadata': self.additional_metadata,
            'upload_id': self.upload_id,
458
            'upload_hash': self.upload_hash,
459
            'upload_url': self.upload_url,
460
            'upload_command': self.upload_command,
Markus Scheidgen's avatar
Markus Scheidgen committed
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
            'upload_time': self.upload_time.isoformat() if self.upload_time is not None else None,
            'is_stale': self.is_stale,
        }
        data.update(super().json_dict)
        return {key: value for key, value in data.items() if value is not None}

    @process
    def process(self):
        self.extracting()
        self.parse_all()

    @task
    def uploading(self):
        pass

476
477
478
479
480
481
482
    @property
    def upload_file(self):
        """ The :class:`UploadFile` instance that represents the uploaded file of this upload. """
        if not self._upload_file:
            self._upload_file = UploadFile(self.upload_id, local_path=self.local_path)
        return self._upload_file

Markus Scheidgen's avatar
Markus Scheidgen committed
483
484
    @task
    def extracting(self):
485
486
487
488
489
490
        """
        Task performed before the actual parsing/normalizing. Extracting and bagging
        the uploaded files, computing all keys, create an *upload* entry in the NOMAD-coe
        repository db, etc.
        """
        # extract the uploaded file, this will also create a bagit bag.
Markus Scheidgen's avatar
Markus Scheidgen committed
491
492
        logger = self.get_logger()
        try:
Markus Scheidgen's avatar
Markus Scheidgen committed
493
494
            with utils.timer(
                    logger, 'upload extracted', step='extracting',
495
496
                    upload_size=self.upload_file.size):
                self.upload_file.extract()
Markus Scheidgen's avatar
Markus Scheidgen committed
497
498
499
500
        except KeyError as e:
            self.fail('process request for non existing upload', level=logging.INFO)
            return

501
        # create and save a hash for the upload
Markus Scheidgen's avatar
Markus Scheidgen committed
502
        try:
503
            self.upload_hash = self.upload_file.upload_hash()
504
        except Exception as e:
Markus Scheidgen's avatar
Markus Scheidgen committed
505
506
507
            self.fail('could not create upload hash', e)
            return

508
        # check if the file was already uploaded and processed before
Markus Scheidgen's avatar
Markus Scheidgen committed
509
        if RepoCalc.upload_exists(self.upload_hash):
Markus Scheidgen's avatar
Markus Scheidgen committed
510
511
512
            self.fail('The same file was already uploaded and processed.', level=logging.INFO)
            return

513
514
515
516
517
518
519
520
521
522
523
524
    def match_mainfiles(self) -> Generator[Tuple[File, str, object], None, None]:
        """
        Generator function that matches all files in the upload to all parsers to
        determine the upload's mainfiles.

        Returns:
            Tuples of mainfile, filename, and parsers
        """
        for filename in self.upload_file.filelist:
            potential_mainfile = self.upload_file.get_file(filename)
            for parser in parsers:
                try:
525
                    with potential_mainfile.open('r') as mainfile_f:
526
527
528
                        if parser.is_mainfile(filename, lambda fn: mainfile_f):
                            yield potential_mainfile, filename, parser
                except Exception as e:
529
                    self.get_logger().error(
530
531
532
                        'exception while matching pot. mainfile',
                        mainfile=filename, exc_info=e)

Markus Scheidgen's avatar
Markus Scheidgen committed
533
534
    @task
    def parse_all(self):
535
        """
536
        Identified mainfile/parser combinations among the upload's files, creates
537
538
        respective :class:`Calc` instances, and triggers their processing.
        """
539
540
        logger = self.get_logger()

Markus Scheidgen's avatar
Markus Scheidgen committed
541
        # TODO: deal with multiple possible parser specs
Markus Scheidgen's avatar
Markus Scheidgen committed
542
543
        with utils.timer(
                logger, 'upload extracted', step='matching',
544
545
                upload_size=self.upload_file.size,
                upload_filecount=len(self.upload_file.filelist)):
546
            total_calcs = 0
547
548
549
550
551
552
553
554
555
            for mainfile, filename, parser in self.match_mainfiles():
                calc = Calc.create(
                    archive_id='%s/%s' % (self.upload_hash, utils.hash(filename)),
                    mainfile=filename, parser=parser.name,
                    mainfile_tmp_path=mainfile.os_path,
                    upload_id=self.upload_id)

                calc.process()
                total_calcs += 1
Markus Scheidgen's avatar
Markus Scheidgen committed
556

557
558
559
560
561
        # have to save the total_calcs information for chord management
        self.spwaned_childred(total_calcs)

    def join(self):
        self.cleanup()
Markus Scheidgen's avatar
Markus Scheidgen committed
562
563
564
565

    @task
    def cleanup(self):
        try:
Markus Scheidgen's avatar
Markus Scheidgen committed
566
            upload = UploadFile(self.upload_id, local_path=self.local_path)
567
568
569
570
571
            with utils.timer(
                    self.get_logger(), 'upload persisted', step='cleaning',
                    upload_size=upload.size):
                upload.persist()

Markus Scheidgen's avatar
Markus Scheidgen committed
572
573
574
575
            with utils.timer(
                    self.get_logger(), 'processing cleaned up', step='cleaning',
                    upload_size=upload.size):
                upload.remove_extract()
Markus Scheidgen's avatar
Markus Scheidgen committed
576
        except KeyError as e:
Markus Scheidgen's avatar
Markus Scheidgen committed
577
            self.fail('Upload does not exist', exc_info=e)
Markus Scheidgen's avatar
Markus Scheidgen committed
578
579
580
581
582
            return

        self.get_logger().debug('closed upload')

    @property
583
584
585
586
587
588
589
590
591
592
593
    def processed_calcs(self):
        return Calc.objects(upload_id=self.upload_id, status__in=[SUCCESS, FAILURE]).count()

    @property
    def total_calcs(self):
        return Calc.objects(upload_id=self.upload_id).count()

    @property
    def failed_calcs(self):
        return Calc.objects(upload_id=self.upload_id, status=FAILURE).count()

594
595
596
597
    @property
    def pending_calcs(self):
        return Calc.objects(upload_id=self.upload_id, status=PENDING).count()

598
599
    def all_calcs(self, start, end, order_by='mainfile'):
        return Calc.objects(upload_id=self.upload_id)[start:end].order_by(order_by)
600

Markus Scheidgen's avatar
Markus Scheidgen committed
601
    @staticmethod
602
603
604
605
606
    def repair_all():
        """
        Utitlity function that will look for suspiciously looking conditions in
        all uncompleted downloads. It ain't a perfect world.
        """
607
608
609
        # TODO this was added as a quick fix to #37.
        # Even though it might be strictly necessary, there should be a tested backup
        # solution for it Chords to not work properly due to failed to fail processings
610
611
612
613
614
615
616
617
618
619
        uploads = Upload.objects(status__in=[PENDING, RUNNING])
        for upload in uploads:
            completed = upload.processed_calcs
            total = upload.total
            pending = upload.pending_calcs

            if completed + pending == total:
                time.sleep(2)
                if pending == upload.pending_calcs:
                    Calc.objects(upload_id=upload.upload_id, status=PENDING).delete()