data.py 20 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
28
from elasticsearch.exceptions import NotFoundError
Markus Scheidgen's avatar
Markus Scheidgen committed
29
from mongoengine import StringField, BooleanField, DateTimeField, DictField, IntField
Markus Scheidgen's avatar
Markus Scheidgen committed
30
import logging
31
from structlog import wrap_logger
32
from contextlib import contextmanager
Markus Scheidgen's avatar
Markus Scheidgen committed
33

34
from nomad import utils, coe_repo
Markus Scheidgen's avatar
Markus Scheidgen committed
35
from nomad.files import UploadFile, ArchiveFile, ArchiveLogFile, File
Markus Scheidgen's avatar
Markus Scheidgen committed
36
from nomad.repo import RepoCalc
37
from nomad.processing.base import Proc, Chord, process, task, PENDING, SUCCESS, FAILURE
Markus Scheidgen's avatar
Markus Scheidgen committed
38
from nomad.parsing import parsers, parser_dict
Markus Scheidgen's avatar
Markus Scheidgen committed
39
from nomad.normalizing import normalizers
Markus Scheidgen's avatar
Markus Scheidgen committed
40
from nomad.utils import lnr
Markus Scheidgen's avatar
Markus Scheidgen committed
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


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': [
71
            'upload_id', 'mainfile', 'code', 'parser', 'status'
Markus Scheidgen's avatar
Markus Scheidgen committed
72
73
74
75
76
77
78
        ]
    }

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self._parser_backend = None
        self._upload = None
79
        self._calc_proc_logwriter = None
Markus Scheidgen's avatar
Markus Scheidgen committed
80
        self._calc_proc_logfile = None
81
        self._calc_proc_logwriter_ctx: ContextManager = None
Markus Scheidgen's avatar
Markus Scheidgen committed
82
83
84
85
86

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

Markus Scheidgen's avatar
Markus Scheidgen committed
87
88
89
90
    @property
    def mainfile_file(self) -> File:
        return File(self.mainfile_tmp_path)

91
92
93
94
95
96
    @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
97
98
99
100
    def delete(self):
        """
        Delete this calculation and all associated data. This includes all files,
        the archive, and this search index entry.
101
        TODO is this needed? Or do we always delete hole uploads in bulk.
Markus Scheidgen's avatar
Markus Scheidgen committed
102
103
104
        """
        # delete the archive
        if self.archive_id is not None:
105
            ArchiveFile(self.archive_id).delete()
Markus Scheidgen's avatar
Markus Scheidgen committed
106
107

        # delete the search index entry
108
109
110
111
112
113
        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
114
115
116
117
118
119
120
121
122

        # 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,
123
            upload_hash=upload_hash, calc_hash=calc_hash,
Markus Scheidgen's avatar
Markus Scheidgen committed
124
            archive_id=self.archive_id, **kwargs)
125

Markus Scheidgen's avatar
Markus Scheidgen committed
126
127
        return logger

128
129
130
131
132
133
134
135
    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
136
            self._calc_proc_logfile = ArchiveLogFile(self.archive_id)
137
138
            self._calc_proc_logwriter_ctx = self._calc_proc_logfile.open('wt')
            self._calc_proc_logwriter = self._calc_proc_logwriter_ctx.__enter__()  # pylint: disable=E1101
139

140
        def save_to_calc_log(logger, method_name, event_dict):
141
142
143
144
145
146
147
148
149
150
151
            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

152
        return wrap_logger(logger, processors=[save_to_calc_log])
153

Markus Scheidgen's avatar
Markus Scheidgen committed
154
155
    @process
    def process(self):
156
        logger = self.get_logger()
157
        if self.upload is None:
158
            logger.error('calculation upload does not exist')
Markus Scheidgen's avatar
Markus Scheidgen committed
159
160
161
162
163
164

        try:
            self.parsing()
            self.normalizing()
            self.archiving()
        finally:
Markus Scheidgen's avatar
Markus Scheidgen committed
165
            # close loghandler that was not closed due to failures
166
            try:
167
168
169
                if self._calc_proc_logwriter is not None:
                    self._calc_proc_logwriter.close()
                    self._calc_proc_logwriter = None
170
171
172
173
            except Exception as e:
                logger.error('could not close calculation proc log', exc_info=e)

            # inform parent proc about completion
174
            self.upload.completed_child()
Markus Scheidgen's avatar
Markus Scheidgen committed
175
176
177

    @task
    def parsing(self):
178
179
        context = dict(parser=self.parser, step=self.parser)
        logger = self.get_calc_logger(**context)
180
        parser = parser_dict[self.parser]
Markus Scheidgen's avatar
Markus Scheidgen committed
181

182
        with utils.timer(logger, 'parser executed', input_size=self.mainfile_file.size):
183
            self._parser_backend = parser.run(self.mainfile_tmp_path, logger=logger)
Markus Scheidgen's avatar
Markus Scheidgen committed
184

185
186
187
188
189
190
        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
191
        if self._parser_backend.status[0] != 'ParseSuccess':
192
            logger.error(self._parser_backend.status[1])
Markus Scheidgen's avatar
Markus Scheidgen committed
193
            error = self._parser_backend.status[1]
194
            self._parser_backend.addValue('parse_status', 'ParseFailure')
195
            self.fail(error, level=logging.DEBUG, **context)
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
        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
228
229
230
231
232

    @task
    def normalizing(self):
        for normalizer in normalizers:
            normalizer_name = normalizer.__name__
233
234
            context = dict(normalizer=normalizer_name, step=normalizer_name)
            logger = self.get_calc_logger(**context)
Markus Scheidgen's avatar
Markus Scheidgen committed
235
236

            with utils.timer(
237
                    logger, 'normalizer executed', input_size=self.mainfile_file.size):
238
239
                with self.use_parser_backend(normalizer_name) as backend:
                    normalizer(backend).normalize(logger=logger)
Markus Scheidgen's avatar
Markus Scheidgen committed
240

241
242
            failed = self._parser_backend.status[0] != 'ParseSuccess'
            if failed:
243
                logger.error(self._parser_backend.status[1])
Markus Scheidgen's avatar
Markus Scheidgen committed
244
                error = self._parser_backend.status[1]
245
                self.fail(error, level=logging.WARNING, **context)
246
247
248
249
                break
            else:
                logger.debug(
                    'completed normalizer successfully', normalizer=normalizer_name)
Markus Scheidgen's avatar
Markus Scheidgen committed
250
251
252

    @task
    def archiving(self):
253
254
255
256
257
        logger = self.get_logger()

        upload_hash, calc_hash = self.archive_id.split('/')
        additional = dict(
            mainfile=self.mainfile,
258
            upload_time=self.upload.upload_time,
259
260
            staging=True,
            restricted=False,
261
262
            user_id=self.upload.user_id,
            aux_files=list(self.upload.upload_file.get_siblings(self.mainfile)))
263
264

        with utils.timer(logger, 'indexed', step='index'):
265
266
267
268
269
270
            # persist to elastic search
            RepoCalc.create_from_backend(
                self._parser_backend,
                additional=additional,
                upload_hash=upload_hash,
                calc_hash=calc_hash,
271
                upload_id=self.upload_id).persist()
272

Markus Scheidgen's avatar
Markus Scheidgen committed
273
274
275
276
        with utils.timer(
                logger, 'archived', step='archive',
                input_size=self.mainfile_file.size) as log_data:

277
            # persist the archive
Markus Scheidgen's avatar
Markus Scheidgen committed
278
279
            archive_file = ArchiveFile(self.archive_id)
            with archive_file.write_archive_json() as out:
280
281
                self._parser_backend.write_json(out, pretty=True)

Markus Scheidgen's avatar
Markus Scheidgen committed
282
283
284
285
286
287
288
            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:
289
                self._calc_proc_logwriter_ctx.__exit__(None, None, None)  # pylint: disable=E1101
290
                self._calc_proc_logwriter = None
Markus Scheidgen's avatar
Markus Scheidgen committed
291

Markus Scheidgen's avatar
Markus Scheidgen committed
292
293
                log_data.update(log_size=self._calc_proc_logfile.size)

Markus Scheidgen's avatar
Markus Scheidgen committed
294

295
class Upload(Chord):
Markus Scheidgen's avatar
Markus Scheidgen committed
296
297
298
299
300
301
    """
    Represents uploads in the databases. Provides persistence access to the files storage,
    and processing state.

    Attributes:
        name: optional user provided upload name
302
        local_path: optional local path, e.g. for files that are already somewhere on the server
Markus Scheidgen's avatar
Markus Scheidgen committed
303
304
305
306
307
308
        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
309
        user_id: the id of the user that created this upload
Markus Scheidgen's avatar
Markus Scheidgen committed
310
311
312
313
314
315
    """
    id_field = 'upload_id'

    upload_id = StringField(primary_key=True)

    name = StringField(default=None)
316
    local_path = StringField(default=None)
Markus Scheidgen's avatar
Markus Scheidgen committed
317
318
319
320
321
322
323
324
    additional_metadata = DictField(default=None)

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

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

325
    user_id = StringField(required=True)
Markus Scheidgen's avatar
Markus Scheidgen committed
326

327
328
    coe_repo_upload_id = IntField(default=None)

329
330
    _initiated_parsers = IntField(default=-1)

Markus Scheidgen's avatar
Markus Scheidgen committed
331
332
    meta: Any = {
        'indexes': [
333
            'upload_hash', 'user_id', 'status'
Markus Scheidgen's avatar
Markus Scheidgen committed
334
335
336
337
338
        ]
    }

    def __init__(self, **kwargs):
        super().__init__(**kwargs)
339
        self._upload_file = None
Markus Scheidgen's avatar
Markus Scheidgen committed
340
341
342
343
344
345

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

    @classmethod
346
    def user_uploads(cls, user: coe_repo.User) -> List['Upload']:
Markus Scheidgen's avatar
Markus Scheidgen committed
347
        """ Returns all uploads for the given user. Currently returns all uploads. """
348
        return cls.objects(user_id=str(user.user_id), in_staging=True)
Markus Scheidgen's avatar
Markus Scheidgen committed
349
350
351
352
353
354
355
356
357

    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')

358
        if not (self.completed or self.current_task == 'uploading'):
Markus Scheidgen's avatar
Markus Scheidgen committed
359
360
361
362
            raise NotAllowedDuringProcessing()

        with lnr(logger, 'delete upload file'):
            try:
363
                UploadFile(self.upload_id, local_path=self.local_path).delete()
Markus Scheidgen's avatar
Markus Scheidgen committed
364
365
366
367
368
369
370
371
372
            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'):
373
            # delete archive files
374
            ArchiveFile.delete_archives(upload_hash=self.upload_hash)
375
376

            # delete repo entries
Markus Scheidgen's avatar
Markus Scheidgen committed
377
            RepoCalc.delete_upload(upload_id=self.upload_id)
378
379
380

            # delete calc processings
            Calc.objects(upload_id=self.upload_id).delete()
Markus Scheidgen's avatar
Markus Scheidgen committed
381
382
383
384
385
386
387
388
389
390

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

    @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.
391
392

        Arguments:
393
            user (coe_repo.User): The user that created the upload.
Markus Scheidgen's avatar
Markus Scheidgen committed
394
        """
395
        user: coe_repo.User = kwargs['user']
396
397
398
        del(kwargs['user'])
        if 'upload_id' not in kwargs:
            kwargs.update(upload_id=utils.create_uuid())
399
        kwargs.update(user_id=str(user.user_id))
Markus Scheidgen's avatar
Markus Scheidgen committed
400
        self = super().create(**kwargs)
401

Markus Scheidgen's avatar
Markus Scheidgen committed
402
        self._continue_with('uploading')
403

Markus Scheidgen's avatar
Markus Scheidgen committed
404
405
        return self

406
    def unstage(self, meta_data):
407
        self.get_logger().info('unstage')
408
409
410
411

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

Markus Scheidgen's avatar
Markus Scheidgen committed
412
        self.in_staging = False
413
        RepoCalc.unstage(upload_id=self.upload_id)
414
        coe_repo.add_upload(self, meta_data)
Markus Scheidgen's avatar
Markus Scheidgen committed
415
416
        self.save()

Markus Scheidgen's avatar
Markus Scheidgen committed
417
418
419
420
421
422
423
424
425
    @process
    def process(self):
        self.extracting()
        self.parse_all()

    @task
    def uploading(self):
        pass

426
427
428
429
430
431
432
    @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
433
434
    @task
    def extracting(self):
435
436
437
438
439
440
        """
        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
441
442
        logger = self.get_logger()
        try:
Markus Scheidgen's avatar
Markus Scheidgen committed
443
444
            with utils.timer(
                    logger, 'upload extracted', step='extracting',
445
446
                    upload_size=self.upload_file.size):
                self.upload_file.extract()
Markus Scheidgen's avatar
Markus Scheidgen committed
447
448
449
450
        except KeyError as e:
            self.fail('process request for non existing upload', level=logging.INFO)
            return

451
        # create and save a hash for the upload
Markus Scheidgen's avatar
Markus Scheidgen committed
452
        try:
453
            self.upload_hash = self.upload_file.upload_hash()
454
        except Exception as e:
Markus Scheidgen's avatar
Markus Scheidgen committed
455
456
457
            self.fail('could not create upload hash', e)
            return

458
        # check if the file was already uploaded and processed before
Markus Scheidgen's avatar
Markus Scheidgen committed
459
        if RepoCalc.upload_exists(self.upload_hash):
Markus Scheidgen's avatar
Markus Scheidgen committed
460
461
462
            self.fail('The same file was already uploaded and processed.', level=logging.INFO)
            return

463
464
465
466
467
468
469
470
471
472
473
474
    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:
475
                    with potential_mainfile.open('r') as mainfile_f:
476
477
478
                        if parser.is_mainfile(filename, lambda fn: mainfile_f):
                            yield potential_mainfile, filename, parser
                except Exception as e:
479
                    self.get_logger().error(
480
481
482
                        'exception while matching pot. mainfile',
                        mainfile=filename, exc_info=e)

Markus Scheidgen's avatar
Markus Scheidgen committed
483
484
    @task
    def parse_all(self):
485
        """
486
        Identified mainfile/parser combinations among the upload's files, creates
487
488
        respective :class:`Calc` instances, and triggers their processing.
        """
489
490
        logger = self.get_logger()

Markus Scheidgen's avatar
Markus Scheidgen committed
491
        # TODO: deal with multiple possible parser specs
Markus Scheidgen's avatar
Markus Scheidgen committed
492
493
        with utils.timer(
                logger, 'upload extracted', step='matching',
494
495
                upload_size=self.upload_file.size,
                upload_filecount=len(self.upload_file.filelist)):
496
            total_calcs = 0
497
498
499
500
501
502
503
504
505
            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
506

507
508
509
510
511
        # 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
512
513
514
515

    @task
    def cleanup(self):
        try:
Markus Scheidgen's avatar
Markus Scheidgen committed
516
            upload = UploadFile(self.upload_id, local_path=self.local_path)
517
518
519
520
521
            with utils.timer(
                    self.get_logger(), 'upload persisted', step='cleaning',
                    upload_size=upload.size):
                upload.persist()

Markus Scheidgen's avatar
Markus Scheidgen committed
522
523
524
525
            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
526
        except KeyError as e:
Markus Scheidgen's avatar
Markus Scheidgen committed
527
            self.fail('Upload does not exist', exc_info=e)
Markus Scheidgen's avatar
Markus Scheidgen committed
528
529
530
531
532
            return

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

    @property
533
534
535
536
537
538
539
540
541
542
543
    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()

544
545
546
547
    @property
    def pending_calcs(self):
        return Calc.objects(upload_id=self.upload_id, status=PENDING).count()

548
549
    def all_calcs(self, start, end, order_by='mainfile'):
        return Calc.objects(upload_id=self.upload_id)[start:end].order_by(order_by)