data.py 31.6 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:
"""

Markus Scheidgen's avatar
Markus Scheidgen committed
27
from typing import cast, List, Any, ContextManager, Tuple, Generator, Dict, cast
28
from mongoengine import StringField, DateTimeField, DictField, BooleanField
Markus Scheidgen's avatar
Markus Scheidgen committed
29
import logging
30
from structlog import wrap_logger
31
from contextlib import contextmanager
32
import os.path
33
34
from datetime import datetime
from pymongo import UpdateOne
Markus Scheidgen's avatar
Markus Scheidgen committed
35

36
from nomad import utils, coe_repo, config, infrastructure, search, datamodel
37
from nomad.files import PathObject, UploadFiles, ExtractError, ArchiveBasedStagingUploadFiles, PublicUploadFiles
38
from nomad.processing.base import Proc, process, task, PENDING, SUCCESS, FAILURE
39
from nomad.parsing import parser_dict, match_parser, LocalBackend
Markus Scheidgen's avatar
Markus Scheidgen committed
40
from nomad.normalizing import normalizers
41
from nomad.datamodel import UploadWithMetadata, Domain
Markus Scheidgen's avatar
Markus Scheidgen committed
42
43


Markus Scheidgen's avatar
Markus Scheidgen committed
44
class Calc(Proc):
Markus Scheidgen's avatar
Markus Scheidgen committed
45
46
47
48
49
50
51
52
53
54
    """
    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:
55
        calc_id: the calc_id of this calc
Markus Scheidgen's avatar
Markus Scheidgen committed
56
57
58
        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
59

60
        metadata: the metadata record wit calc and user metadata, see :class:`datamodel.CalcWithMetadata`
Markus Scheidgen's avatar
Markus Scheidgen committed
61
    """
62
    calc_id = StringField(primary_key=True)
Markus Scheidgen's avatar
Markus Scheidgen committed
63
64
65
66
    upload_id = StringField()
    mainfile = StringField()
    parser = StringField()

67
68
    metadata = DictField()

Markus Scheidgen's avatar
Markus Scheidgen committed
69
    meta: Any = {
Markus Scheidgen's avatar
Markus Scheidgen committed
70
        'indexes': [
71
72
73
74
75
            'upload_id',
            ('upload_id', 'mainfile'),
            ('upload_id', 'parser'),
            ('upload_id', 'tasks_status'),
            ('upload_id', 'process_status')
Markus Scheidgen's avatar
Markus Scheidgen committed
76
77
78
79
80
        ]
    }

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
81
        self._parser_backend: LocalBackend = None
82
83
        self._upload: Upload = None
        self._upload_files: ArchiveBasedStagingUploadFiles = None
84
        self._calc_proc_logwriter = None
85
        self._calc_proc_logwriter_ctx: ContextManager = None
Markus Scheidgen's avatar
Markus Scheidgen committed
86
87
88

    @classmethod
    def get(cls, id):
89
        return cls.get_by_id(id, 'calc_id')
Markus Scheidgen's avatar
Markus Scheidgen committed
90

Markus Scheidgen's avatar
Markus Scheidgen committed
91
    @property
92
93
    def mainfile_file(self) -> PathObject:
        return self.upload_files.raw_file_object(self.mainfile)
Markus Scheidgen's avatar
Markus Scheidgen committed
94

95
96
97
98
    @property
    def upload(self) -> 'Upload':
        if not self._upload:
            self._upload = Upload.get(self.upload_id)
99
            self._upload.worker_hostname = self.worker_hostname
100
101
        return self._upload

102
103
104
    @property
    def upload_files(self) -> ArchiveBasedStagingUploadFiles:
        if not self._upload_files:
105
106
            self._upload_files = ArchiveBasedStagingUploadFiles(
                self.upload_id, is_authorized=lambda: True, upload_path=self.upload.upload_path)
107
108
        return self._upload_files

Markus Scheidgen's avatar
Markus Scheidgen committed
109
    def get_logger(self, **kwargs):
110
111
112
113
        """
        Returns a wrapped logger that additionally saves all entries to the calculation
        processing log in the archive.
        """
114
115
116
        logger = super().get_logger()
        logger = logger.bind(
            upload_id=self.upload_id, mainfile=self.mainfile, calc_id=self.calc_id, **kwargs)
117

118
        if self._calc_proc_logwriter_ctx is None:
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
            try:
                self._calc_proc_logwriter_ctx = self.upload_files.archive_log_file(self.calc_id, 'wt')
                self._calc_proc_logwriter = self._calc_proc_logwriter_ctx.__enter__()  # pylint: disable=E1101
            except KeyError:
                # cannot open log file
                pass

        if self._calc_proc_logwriter_ctx is None:
            return logger
        else:
            def save_to_calc_log(logger, method_name, event_dict):
                if self._calc_proc_logwriter is not None:
                    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

            return wrap_logger(logger, processors=[save_to_calc_log])
144

Markus Scheidgen's avatar
Markus Scheidgen committed
145
    @process
146
    def process_calc(self):
147
        logger = self.get_logger()
148
        if self.upload is None:
149
            logger.error('calculation upload does not exist')
Markus Scheidgen's avatar
Markus Scheidgen committed
150
151

        try:
152
153
            # save preliminary minimum calc metadata in case processing fails
            # successful processing will replace it with the actual metadata
154
            calc_with_metadata = datamodel.CalcWithMetadata(
155
156
157
158
159
160
                upload_id=self.upload_id,
                calc_id=self.calc_id,
                calc_hash=self.upload_files.calc_hash(self.mainfile),
                mainfile=self.mainfile)
            calc_with_metadata.published = False
            calc_with_metadata.uploader = self.upload.uploader.to_popo()
161
            calc_with_metadata.upload_time = self.upload.upload_time
162
            calc_with_metadata.nomad_version = config.version
163
            calc_with_metadata.nomad_commit = config.commit
164
165
166
167
            calc_with_metadata.last_processing = datetime.now()
            calc_with_metadata.files = self.upload_files.calc_files(self.mainfile)
            self.metadata = calc_with_metadata.to_dict()

Markus Scheidgen's avatar
Markus Scheidgen committed
168
169
170
171
            self.parsing()
            self.normalizing()
            self.archiving()
        finally:
Markus Scheidgen's avatar
Markus Scheidgen committed
172
            # close loghandler that was not closed due to failures
173
            try:
174
175
176
                if self._calc_proc_logwriter is not None:
                    self._calc_proc_logwriter.close()
                    self._calc_proc_logwriter = None
177
178
179
            except Exception as e:
                logger.error('could not close calculation proc log', exc_info=e)

180
181
182
183
    def fail(self, *errors, log_level=logging.ERROR, **kwargs):
        # in case of failure, index a minimum set of metadata and mark
        # processing failure
        try:
184
            calc_with_metadata = datamodel.CalcWithMetadata(**self.metadata)
185
186
187
188
189
190
191
192
            calc_with_metadata.formula = config.services.not_processed_value
            calc_with_metadata.basis_set = config.services.not_processed_value
            calc_with_metadata.xc_functional = config.services.not_processed_value
            calc_with_metadata.system = config.services.not_processed_value
            calc_with_metadata.crystal_system = config.services.not_processed_value
            calc_with_metadata.spacegroup = config.services.not_processed_value
            calc_with_metadata.spacegroup_symbol = config.services.not_processed_value
            calc_with_metadata.code_version = config.services.not_processed_value
193
194
195
196
197
198
199

            calc_with_metadata.code_name = config.services.not_processed_value
            if self.parser is not None:
                parser = parser_dict[self.parser]
                if hasattr(parser, 'code_name'):
                    calc_with_metadata.code_name = parser.code_name

200
201
202
203
204
205
206
207
            calc_with_metadata.processed = False
            self.metadata = calc_with_metadata.to_dict()
            search.Entry.from_calc_with_metadata(calc_with_metadata).save()
        except Exception as e:
            self.get_logger().error('could not index after processing failure', exc_info=e)

        super().fail(*errors, log_level=log_level, **kwargs)

208
209
210
211
212
213
214
    def on_process_complete(self, process_name):
        # the save might be necessary to correctly read the join condition from the db
        self.save()
        # in case of error, the process_name might be unknown
        if process_name == 'process_calc' or process_name is None:
            self.upload.reload()
            self.upload.check_join()
Markus Scheidgen's avatar
Markus Scheidgen committed
215
216
217

    @task
    def parsing(self):
218
        context = dict(parser=self.parser, step=self.parser)
219
        logger = self.get_logger(**context)
220
        parser = parser_dict[self.parser]
221
        self.metadata['parser_name'] = self.parser
Markus Scheidgen's avatar
Markus Scheidgen committed
222

223
        with utils.timer(logger, 'parser executed', input_size=self.mainfile_file.size):
224
225
226
227
228
229
230
            try:
                self._parser_backend = parser.run(
                    self.upload_files.raw_file_object(self.mainfile).os_path, logger=logger)
            except Exception as e:
                self.fail(
                    'parser failed with exception', level=logging.ERROR,
                    exc_info=e, error=str(e), **context)
231
                return
Markus Scheidgen's avatar
Markus Scheidgen committed
232

233
234
        # add the non code specific calc metadata to the backend
        # all other quantities have been determined by parsers/normalizers
235
        self._parser_backend.openNonOverlappingSection('section_entry_info')
236
        self._parser_backend.addValue('upload_id', self.upload_id)
237
        self._parser_backend.addValue('calc_id', self.calc_id)
238
        self._parser_backend.addValue('calc_hash', self.metadata['calc_hash'])
239
        self._parser_backend.addValue('mainfile', self.mainfile)
240
        self._parser_backend.addValue('parser_name', self.parser)
241
242
243
244
245
246
247
248
249
250
251
252
        filepaths = self.metadata['files']
        self._parser_backend.addValue('number_of_files', len(filepaths))
        self._parser_backend.addValue('filepaths', filepaths)
        uploader = self.upload.uploader
        self._parser_backend.addValue(
            'entry_uploader_name', '%s, %s' % (uploader.first_name, uploader.last_name))
        self._parser_backend.addValue(
            'entry_uploader_id', str(uploader.user_id))
        self._parser_backend.addValue('entry_upload_time', int(self.upload.upload_time.timestamp()))
        self._parser_backend.closeNonOverlappingSection('section_entry_info')

        self.add_processor_info(self.parser)
253

Markus Scheidgen's avatar
Markus Scheidgen committed
254
        if self._parser_backend.status[0] != 'ParseSuccess':
255
            logger.error(self._parser_backend.status[1])
Markus Scheidgen's avatar
Markus Scheidgen committed
256
            error = self._parser_backend.status[1]
257
            self.fail(error, level=logging.INFO, **context)
258
259
260
261
262
263
264
265

    @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:
266
        self._parser_backend.openContext('/section_entry_info/0')
267
268
269
270
271
272
273
        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')
274
                self._parser_backend.addValue('number_of_archive_processor_warnings', len(warnings))
275
276
277
278
                self._parser_backend.addArrayValues('archive_processor_warnings', [str(warning) for warning in warnings])
            else:
                self._parser_backend.addValue('archive_processor_status', 'Success')
        else:
279
            errors = self._parser_backend.status[1]
280
281
282
            self._parser_backend.addValue('archive_processor_error', str(errors))

        self._parser_backend.closeNonOverlappingSection('section_archive_processing_info')
283
        self._parser_backend.closeContext('/section_entry_info/0')
Markus Scheidgen's avatar
Markus Scheidgen committed
284
285
286
287

    @task
    def normalizing(self):
        for normalizer in normalizers:
288
289
290
            if normalizer.domain != config.domain:
                continue

Markus Scheidgen's avatar
Markus Scheidgen committed
291
            normalizer_name = normalizer.__name__
292
            context = dict(normalizer=normalizer_name, step=normalizer_name)
293
            logger = self.get_logger(**context)
Markus Scheidgen's avatar
Markus Scheidgen committed
294
295

            with utils.timer(
296
                    logger, 'normalizer executed', input_size=self.mainfile_file.size):
297
                with self.use_parser_backend(normalizer_name) as backend:
298
299
300
301
302
303
304
                    try:
                        normalizer(backend).normalize(logger=logger)
                    except Exception as e:
                        self.fail(
                            'normalizer failed with exception', level=logging.ERROR,
                            exc_info=e, error=str(e), **context)
                        self._parser_backend.status = ['ParseFailure', str(e)]
Markus Scheidgen's avatar
Markus Scheidgen committed
305

306
307
            failed = self._parser_backend.status[0] != 'ParseSuccess'
            if failed:
308
                logger.error(self._parser_backend.status[1])
Markus Scheidgen's avatar
Markus Scheidgen committed
309
                error = self._parser_backend.status[1]
310
                self.fail(error, level=logging.WARNING, error=error, **context)
311
312
313
314
                break
            else:
                logger.debug(
                    'completed normalizer successfully', normalizer=normalizer_name)
Markus Scheidgen's avatar
Markus Scheidgen committed
315
316
317

    @task
    def archiving(self):
318
319
        logger = self.get_logger()

320
        calc_with_metadata = datamodel.CalcWithMetadata(**self.metadata)
321
        calc_with_metadata.apply_domain_metadata(self._parser_backend)
322
        calc_with_metadata.processed = True
323

324
325
        # persist the calc metadata
        with utils.timer(logger, 'saved calc metadata', step='metadata'):
326
            self.metadata = calc_with_metadata.to_dict()
327
328

        # index in search
329
        with utils.timer(logger, 'indexed', step='index'):
330
            search.Entry.from_calc_with_metadata(calc_with_metadata).save()
331

332
        # persist the archive
Markus Scheidgen's avatar
Markus Scheidgen committed
333
        with utils.timer(
334
                logger, 'archived', step='archive',
Markus Scheidgen's avatar
Markus Scheidgen committed
335
                input_size=self.mainfile_file.size) as log_data:
336
            with self.upload_files.archive_file(self.calc_id, 'wt') as out:
337
                self._parser_backend.write_json(out, pretty=True, root_sections=Domain.instance.root_sections)
338

339
            log_data.update(archive_size=self.upload_files.archive_file_object(self.calc_id).size)
Markus Scheidgen's avatar
Markus Scheidgen committed
340
341
342
343

        # close loghandler
        if self._calc_proc_logwriter is not None:
            with utils.timer(
344
                    logger, 'archived log', step='logs',
Markus Scheidgen's avatar
Markus Scheidgen committed
345
                    input_size=self.mainfile_file.size) as log_data:
346
                self._calc_proc_logwriter_ctx.__exit__(None, None, None)  # pylint: disable=E1101
347
                self._calc_proc_logwriter = None
Markus Scheidgen's avatar
Markus Scheidgen committed
348

349
                log_data.update(log_size=self.upload_files.archive_log_file_object(self.calc_id).size)
Markus Scheidgen's avatar
Markus Scheidgen committed
350

351
    def __str__(self):
352
        return 'calc %s calc_id=%s upload_id%s' % (super().__str__(), self.calc_id, self.upload_id)
353

354

355
class Upload(Proc):
Markus Scheidgen's avatar
Markus Scheidgen committed
356
357
358
359
360
361
    """
    Represents uploads in the databases. Provides persistence access to the files storage,
    and processing state.

    Attributes:
        name: optional user provided upload name
362
363
        upload_path: the path were the uploaded files was stored
        temporary: True if the uploaded file should be removed after extraction
364
        metadata: optional user provided additional meta data
Markus Scheidgen's avatar
Markus Scheidgen committed
365
366
        upload_id: the upload id generated by the database
        upload_time: the timestamp when the system realised the upload
367
        user_id: the id of the user that created this upload
Markus Scheidgen's avatar
Markus Scheidgen committed
368
369
370
371
    """
    id_field = 'upload_id'

    upload_id = StringField(primary_key=True)
372
373
    upload_path = StringField(default=None)
    temporary = BooleanField(default=False)
Markus Scheidgen's avatar
Markus Scheidgen committed
374
375
376

    name = StringField(default=None)
    upload_time = DateTimeField()
377
    user_id = StringField(required=True)
378
379
    published = BooleanField(default=False)
    publish_time = DateTimeField()
Markus Scheidgen's avatar
Markus Scheidgen committed
380
381
382

    meta: Any = {
        'indexes': [
383
            'user_id', 'tasks_status', 'process_status', 'published'
Markus Scheidgen's avatar
Markus Scheidgen committed
384
385
386
387
388
        ]
    }

    def __init__(self, **kwargs):
        super().__init__(**kwargs)
389
        self._upload_files: ArchiveBasedStagingUploadFiles = None
Markus Scheidgen's avatar
Markus Scheidgen committed
390

391
392
    @property
    def metadata(self) -> dict:
393
394
395
396
397
398
        # TODO user_metadata needs to be stored in the public bucket, since staging data might not be shared
        try:
            upload_files = PublicUploadFiles(self.upload_id, is_authorized=lambda: True)
        except KeyError:
            return None
        return upload_files.user_metadata
399
400
401

    @metadata.setter
    def metadata(self, data: dict) -> None:
402
403
404
        # TODO user_metadata needs to be stored in the public bucket, since staging data might not be shared
        upload_files = PublicUploadFiles(self.upload_id, is_authorized=lambda: True, create=True)
        upload_files.user_metadata = data
405

Markus Scheidgen's avatar
Markus Scheidgen committed
406
    @classmethod
407
408
409
410
411
412
413
    def get(cls, id: str, include_published: bool = False) -> 'Upload':
        upload = cls.get_by_id(id, 'upload_id')
        # TODO published uploads should not be hidden by this and API
        if upload is not None and (not upload.published or include_published):
            return upload

        raise KeyError()
Markus Scheidgen's avatar
Markus Scheidgen committed
414
415

    @classmethod
416
    def user_uploads(cls, user: coe_repo.User) -> List['Upload']:
Markus Scheidgen's avatar
Markus Scheidgen committed
417
        """ Returns all uploads for the given user. Currently returns all uploads. """
418
        return cls.objects(user_id=str(user.user_id), published=False)
Markus Scheidgen's avatar
Markus Scheidgen committed
419

420
421
422
423
    @property
    def uploader(self):
        return coe_repo.User.from_user_id(self.user_id)

Markus Scheidgen's avatar
Markus Scheidgen committed
424
425
    def get_logger(self, **kwargs):
        logger = super().get_logger()
426
        logger = logger.bind(upload_id=self.upload_id, upload_name=self.name, **kwargs)
Markus Scheidgen's avatar
Markus Scheidgen committed
427
428
429
430
431
432
433
434
        return logger

    @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.
435
436

        Arguments:
437
            user (coe_repo.User): The user that created the upload.
Markus Scheidgen's avatar
Markus Scheidgen committed
438
        """
439
        user: coe_repo.User = kwargs['user']
440
441
442
        del(kwargs['user'])
        if 'upload_id' not in kwargs:
            kwargs.update(upload_id=utils.create_uuid())
443
        kwargs.update(user_id=str(user.user_id))
Markus Scheidgen's avatar
Markus Scheidgen committed
444
        self = super().create(**kwargs)
445

Markus Scheidgen's avatar
Markus Scheidgen committed
446
        self._continue_with('uploading')
447

Markus Scheidgen's avatar
Markus Scheidgen committed
448
449
        return self

450
451
452
453
454
455
456
457
458
459
460
461
462
463
    def delete(self):
        """ Deletes this upload process state entry and its calcs. """
        Calc.objects(upload_id=self.upload_id).delete()
        super().delete()

    @process
    def delete_upload(self):
        """
        Deletes of the upload, including its processing state and
        staging files.
        """
        logger = self.get_logger()

        with utils.lnr(logger, 'staged upload delete failed'):
464
465
466
467
468
469

            with utils.timer(
                    logger, 'upload deleted from repo db', step='repo',
                    upload_size=self.upload_files.size):
                coe_repo.Upload.delete(self.upload_id)

470
            with utils.timer(
471
                    logger, 'upload deleted from index', step='index',
472
                    upload_size=self.upload_files.size):
473
                search.delete_upload(self.upload_id)
474

475
            with utils.timer(
476
                    logger, 'staged upload deleted', step='files',
477
478
479
480
481
                    upload_size=self.upload_files.size):
                self.upload_files.delete()
                self.delete()

        return True  # do not save the process status on the delete upload
482

483
484
485
486
487
488
489
490
491
492
493
494
495
496
    @process
    def enforce_consistency(self):
        """
        Takes the proc data of this upload as truth and updates coe repository db and
        ES index accordingly. It takes userdata from coe repository db, if exists as truth.
        """
        # retrive data from coe repository

        # diff coe repository data with proc data

        # update coe repository data with diffs

        # update the elastic search index (with diffs?)

497
    @process
498
    def publish_upload(self):
499
500
501
502
503
504
        """
        Moves the upload out of staging to add it to the coe repository. It will
        pack the staging upload files in to public upload files, add entries to the
        coe repository db and remove this instance and its calculation from the
        processing state db.
        """
505
506
        assert self.processed_calcs > 0

507
        logger = self.get_logger()
508
        logger.info('started to publish')
509

510
        with utils.lnr(logger, 'publish failed'):
511
512
            upload_with_metadata = self.to_upload_with_metadata()

513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
            if config.repository_db.publish_enabled:
                with utils.timer(
                        logger, 'upload added to repository', step='repo',
                        upload_size=self.upload_files.size):
                    coe_repo.Upload.publish(upload_with_metadata)

            if config.repository_db.publish_enabled:
                coe_upload = coe_repo.Upload.from_upload_id(upload_with_metadata.upload_id)
                if coe_upload is not None:
                    calcs = [coe_calc.to_calc_with_metadata() for coe_calc in coe_upload.calcs]
                else:
                    calcs = []
            else:
                calcs = upload_with_metadata.calcs

528
            with utils.timer(
529
                    logger, 'upload metadata updated', step='metadata',
530
                    upload_size=self.upload_files.size):
531
532

                def create_update(calc):
533
                    calc.published = True
534
535
536
537
538
                    return UpdateOne(
                        {'_id': calc.calc_id},
                        {'$set': {'metadata': calc.to_dict()}})

                Calc._get_collection().bulk_write([create_update(calc) for calc in calcs])
539

540
541
542
            with utils.timer(
                    logger, 'staged upload files packed', step='pack',
                    upload_size=self.upload_files.size):
543
                self.upload_files.pack(upload_with_metadata)
544
545
546
547

            with utils.timer(
                    logger, 'index updated', step='index',
                    upload_size=self.upload_files.size):
548
                search.publish(calcs)
549
550
551
552
553

            with utils.timer(
                    logger, 'staged upload deleted', step='delete staged',
                    upload_size=self.upload_files.size):
                self.upload_files.delete()
554
555
556
                self.published = True
                self.publish_time = datetime.now()
                self.save()
557

Markus Scheidgen's avatar
Markus Scheidgen committed
558
    @process
559
    def process_upload(self):
Markus Scheidgen's avatar
Markus Scheidgen committed
560
561
562
563
564
565
566
        self.extracting()
        self.parse_all()

    @task
    def uploading(self):
        pass

567
    @property
568
569
    def upload_files(self) -> UploadFiles:
        upload_files_class = ArchiveBasedStagingUploadFiles if not self.published else PublicUploadFiles
570
        kwargs = dict(upload_path=self.upload_path) if not self.published else {}
571
572
573

        if not self._upload_files or not isinstance(self._upload_files, upload_files_class):
            self._upload_files = upload_files_class(
574
                self.upload_id, is_authorized=lambda: True, **kwargs)
575

576
        return self._upload_files
577

578
579
580
581
582
    @property
    def staging_upload_files(self) -> ArchiveBasedStagingUploadFiles:
        assert not self.published
        return cast(ArchiveBasedStagingUploadFiles, self.upload_files)

Markus Scheidgen's avatar
Markus Scheidgen committed
583
584
    @task
    def extracting(self):
585
586
587
588
589
        """
        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.
        """
590
591
592
593
594
        # extract the uploaded file
        self._upload_files = ArchiveBasedStagingUploadFiles(
            upload_id=self.upload_id, is_authorized=lambda: True, create=True,
            upload_path=self.upload_path)

Markus Scheidgen's avatar
Markus Scheidgen committed
595
596
        logger = self.get_logger()
        try:
Markus Scheidgen's avatar
Markus Scheidgen committed
597
598
            with utils.timer(
                    logger, 'upload extracted', step='extracting',
599
600
                    upload_size=self.upload_files.size):
                self.upload_files.extract()
601
602
603
604

            if self.temporary:
                os.remove(self.upload_path)
                self.upload_path = None
605

606
        except KeyError:
607
            self.fail('processing requested for non existing upload', log_level=logging.ERROR)
608
609
610
            return
        except ExtractError:
            self.fail('bad .zip/.tar file', log_level=logging.INFO)
Markus Scheidgen's avatar
Markus Scheidgen committed
611
612
            return

613
    def match_mainfiles(self) -> Generator[Tuple[str, object], None, None]:
614
615
616
617
618
619
620
        """
        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
        """
621
        directories_with_match: Dict[str, str] = dict()
622
623
        upload_files = self.staging_upload_files
        for filename in upload_files.raw_file_manifest():
624
            try:
625
                parser = match_parser(filename, upload_files)
626
                if parser is not None:
627
628
                    directory = os.path.dirname(filename)
                    if directory in directories_with_match:
629
630
631
                        # TODO this might give us the chance to store directory based relationship
                        # between calcs for the future?
                        pass
632
633
                    else:
                        directories_with_match[directory] = filename
634
635

                    yield filename, parser
636
637
638
639
            except Exception as e:
                self.get_logger().error(
                    'exception while matching pot. mainfile',
                    mainfile=filename, exc_info=e)
640

Markus Scheidgen's avatar
Markus Scheidgen committed
641
642
    @task
    def parse_all(self):
643
        """
644
        Identified mainfile/parser combinations among the upload's files, creates
645
646
        respective :class:`Calc` instances, and triggers their processing.
        """
647
648
        logger = self.get_logger()

Markus Scheidgen's avatar
Markus Scheidgen committed
649
650
        with utils.timer(
                logger, 'upload extracted', step='matching',
651
652
                upload_size=self.upload_files.size):
            for filename, parser in self.match_mainfiles():
653
                calc = Calc.create(
654
                    calc_id=self.upload_files.calc_id(filename),
655
                    mainfile=filename, parser=parser.name,
656
                    worker_hostname=self.worker_hostname,
657
658
                    upload_id=self.upload_id)

659
                calc.process_calc()
Markus Scheidgen's avatar
Markus Scheidgen committed
660

661
662
663
664
665
666
667
668
669
670
671
672
    def on_process_complete(self, process_name):
        if process_name == 'process_upload':
            self.check_join()

    def check_join(self):
        total_calcs = self.total_calcs
        processed_calcs = self.processed_calcs

        self.get_logger().debug('check join', processed_calcs=processed_calcs, total_calcs=total_calcs)
        if not self.process_running and processed_calcs >= total_calcs:
            self.get_logger().debug('join')
            self.join()
673
674
675

    def join(self):
        self.cleanup()
Markus Scheidgen's avatar
Markus Scheidgen committed
676
677
678

    @task
    def cleanup(self):
679
680
        search.refresh()

681
682
683
684
685
686
687
        # send email about process finish
        user = self.uploader
        name = '%s %s' % (user.first_name, user.last_name)
        message = '\n'.join([
            'Dear %s,' % name,
            '',
            'your data %suploaded %s has completed processing.' % (
Markus Scheidgen's avatar
Markus Scheidgen committed
688
                self.name if self.name else '', self.upload_time.isoformat()),  # pylint: disable=no-member
Markus Scheidgen's avatar
Markus Scheidgen committed
689
            'You can review your data on your upload page: %s/uploads' % config.api_url()[:-3]
690
        ])
691
692
693
694
695
696
697
        try:
            infrastructure.send_mail(
                name=name, email=user.email, message=message, subject='Processing completed')
        except Exception as e:
            # probably due to email configuration problems
            # don't fail or present this error to clients
            self.logger.error('could not send after processing email', exc_info=e)
Markus Scheidgen's avatar
Markus Scheidgen committed
698

699
700
701
    def get_calc(self, calc_id) -> Calc:
        return Calc.objects(upload_id=self.upload_id, calc_id=calc_id).first()

Markus Scheidgen's avatar
Markus Scheidgen committed
702
    @property
703
    def processed_calcs(self):
704
        return Calc.objects(upload_id=self.upload_id, tasks_status__in=[SUCCESS, FAILURE]).count()
705
706
707
708
709
710
711

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

    @property
    def failed_calcs(self):
712
        return Calc.objects(upload_id=self.upload_id, tasks_status=FAILURE).count()
713

714
715
    @property
    def pending_calcs(self):
716
        return Calc.objects(upload_id=self.upload_id, tasks_status=PENDING).count()
717

718
719
720
    def all_calcs(self, start, end, order_by=None):
        query = Calc.objects(upload_id=self.upload_id)[start:end]
        return query.order_by(order_by) if order_by is not None else query
721
722
723

    @property
    def calcs(self):
Markus Scheidgen's avatar
Markus Scheidgen committed
724
        return Calc.objects(upload_id=self.upload_id, tasks_status=SUCCESS)
Markus Scheidgen's avatar
Markus Scheidgen committed
725
726

    def to_upload_with_metadata(self) -> UploadWithMetadata:
727
728
729
730
        # prepare user metadata per upload and per calc
        calc_metadatas: Dict[str, Any] = dict()
        upload_metadata: Dict[str, Any] = dict()

Markus Scheidgen's avatar
Markus Scheidgen committed
731
        if self.metadata is not None:
732
733
734
            upload_metadata.update(self.metadata)
            if 'calculations' in upload_metadata:
                del(upload_metadata['calculations'])
Markus Scheidgen's avatar
Markus Scheidgen committed
735

Markus Scheidgen's avatar
Markus Scheidgen committed
736
            for calc in self.metadata.get('calculations', []):  # pylint: disable=no-member
737
                calc_metadatas[calc['mainfile']] = calc
Markus Scheidgen's avatar
Markus Scheidgen committed
738

739
        user_upload_time = upload_metadata.get('_upload_time', None)
Markus Scheidgen's avatar
Markus Scheidgen committed
740
741
742
743
744
        result = UploadWithMetadata(
            upload_id=self.upload_id,
            uploader=utils.POPO(id=int(self.user_id)),
            upload_time=self.upload_time if user_upload_time is None else user_upload_time)

745
746
747
748
749
        def get_metadata(calc: Calc):
            """
            Assemble metadata from calc's processed calc metadata and the uploads
            user metadata.
            """
750
            calc_data = calc.metadata
751
            calc_with_metadata = datamodel.CalcWithMetadata(**calc_data)
752
753
            calc_metadata = dict(upload_metadata)
            calc_metadata.update(calc_metadatas.get(calc.mainfile, {}))
754
755
756
757
            calc_with_metadata.apply_user_metadata(calc_metadata)

            return calc_with_metadata

758
        result.calcs = [get_metadata(calc) for calc in Calc.objects(upload_id=self.upload_id)]
759

Markus Scheidgen's avatar
Markus Scheidgen committed
760
        return result
761

762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
    def compress_and_set_metadata(self, metadata: Dict[str, Any]) -> None:
        """
        Stores the given user metadata in the upload document. This is the metadata
        adhering to the API model (``UploadMetaData``). Most quantities can be stored
        for the upload and for each calculation. This method will try to move same values
        from the calculation to the upload to "compress" the data.
        """
        compressed = {
            key: value for key, value in metadata.items() if key != 'calculations'}
        calculations: List[Dict[str, Any]] = []
        compressed['calculations'] = calculations

        for calc in metadata.get('calculations', []):
            compressed_calc: Dict[str, Any] = {}
            calculations.append(compressed_calc)
            for key, value in calc.items():
                if key in ['_pid', 'mainfile']:
                    # these quantities are explicitly calc specific and have to stay with
                    # the calc
                    compressed_calc[key] = value
                else:
                    if key not in compressed:
                        compressed[key] = value
                    elif compressed[key].__repr__ != value.__repr__:
                        compressed_calc[key] = value
                    else:
                        compressed[key] = value

        self.metadata = compressed

792
    def __str__(self):
793
        return 'upload %s upload_id%s' % (super().__str__(), self.upload_id)