data.py 38.6 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
35
import hashlib
Markus Scheidgen's avatar
Markus Scheidgen committed
36

37
from nomad import utils, coe_repo, config, infrastructure, search, datamodel
38
from nomad.files import PathObject, UploadFiles, ExtractError, ArchiveBasedStagingUploadFiles, PublicUploadFiles, StagingUploadFiles
39
from nomad.processing.base import Proc, process, task, PENDING, SUCCESS, FAILURE
40
from nomad.parsing import parser_dict, match_parser, LocalBackend
Markus Scheidgen's avatar
Markus Scheidgen committed
41
from nomad.normalizing import normalizers
42
from nomad.datamodel import UploadWithMetadata, Domain
Markus Scheidgen's avatar
Markus Scheidgen committed
43
44


Markus Scheidgen's avatar
Markus Scheidgen committed
45
class Calc(Proc):
Markus Scheidgen's avatar
Markus Scheidgen committed
46
47
48
49
50
51
52
53
54
55
    """
    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:
56
        calc_id: the calc_id of this calc
Markus Scheidgen's avatar
Markus Scheidgen committed
57
58
59
        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
60

61
        metadata: the metadata record wit calc and user metadata, see :class:`datamodel.CalcWithMetadata`
Markus Scheidgen's avatar
Markus Scheidgen committed
62
    """
63
    calc_id = StringField(primary_key=True)
Markus Scheidgen's avatar
Markus Scheidgen committed
64
65
66
67
    upload_id = StringField()
    mainfile = StringField()
    parser = StringField()

68
69
    metadata = DictField()

Markus Scheidgen's avatar
Markus Scheidgen committed
70
    meta: Any = {
Markus Scheidgen's avatar
Markus Scheidgen committed
71
        'indexes': [
72
73
74
75
76
            'upload_id',
            ('upload_id', 'mainfile'),
            ('upload_id', 'parser'),
            ('upload_id', 'tasks_status'),
            ('upload_id', 'process_status')
Markus Scheidgen's avatar
Markus Scheidgen committed
77
78
79
80
81
        ]
    }

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

    @classmethod
    def get(cls, id):
90
        return cls.get_by_id(id, 'calc_id')
Markus Scheidgen's avatar
Markus Scheidgen committed
91

Markus Scheidgen's avatar
Markus Scheidgen committed
92
    @property
93
94
    def mainfile_file(self) -> PathObject:
        return self.upload_files.raw_file_object(self.mainfile)
Markus Scheidgen's avatar
Markus Scheidgen committed
95

96
97
98
99
    @property
    def upload(self) -> 'Upload':
        if not self._upload:
            self._upload = Upload.get(self.upload_id)
100
            self._upload.worker_hostname = self.worker_hostname
101
102
        return self._upload

103
104
105
    @property
    def upload_files(self) -> ArchiveBasedStagingUploadFiles:
        if not self._upload_files:
106
107
            self._upload_files = ArchiveBasedStagingUploadFiles(
                self.upload_id, is_authorized=lambda: True, upload_path=self.upload.upload_path)
108
109
        return self._upload_files

Markus Scheidgen's avatar
Markus Scheidgen committed
110
    def get_logger(self, **kwargs):
111
112
113
114
        """
        Returns a wrapped logger that additionally saves all entries to the calculation
        processing log in the archive.
        """
115
116
117
        logger = super().get_logger()
        logger = logger.bind(
            upload_id=self.upload_id, mainfile=self.mainfile, calc_id=self.calc_id, **kwargs)
118

119
        if self._calc_proc_logwriter_ctx is None:
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
            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])
145

146
147
148
149
150
151
152
153
154
155
    @process
    def re_process_calc(self):
        """
        Processes a calculation again. This means there is already metadata and
        instead of creating it initially, we are just updating the existing
        records.
        """
        logger = self.get_logger()

        try:
156
157
158
159
160
161
162
163
            calc_with_metadata = datamodel.CalcWithMetadata(**self.metadata)
            calc_with_metadata.upload_id = self.upload_id
            calc_with_metadata.calc_id = self.calc_id
            calc_with_metadata.calc_hash = self.upload_files.calc_hash(self.mainfile)
            calc_with_metadata.mainfile = self.mainfile
            calc_with_metadata.nomad_version = config.version
            calc_with_metadata.nomad_commit = config.commit
            calc_with_metadata.last_processing = datetime.now()
164
            calc_with_metadata.files = self.upload_files.calc_files(self.mainfile)
165
            self.metadata = calc_with_metadata.to_dict()
166
167
168
169
170
171
172
173
174
175
176
177
178

            self.parsing()
            self.normalizing()
            self.archiving()
        finally:
            # close loghandler that was not closed due to failures
            try:
                if self._calc_proc_logwriter is not None:
                    self._calc_proc_logwriter.close()
                    self._calc_proc_logwriter = None
            except Exception as e:
                logger.error('could not close calculation proc log', exc_info=e)

Markus Scheidgen's avatar
Markus Scheidgen committed
179
    @process
180
    def process_calc(self):
181
182
183
184
        """
        Processes a new calculation that has no prior records in the mongo, elastic,
        or filesystem storage. It will create an initial set of (user) metadata.
        """
185
        logger = self.get_logger()
186
        if self.upload is None:
187
            logger.error('calculation upload does not exist')
Markus Scheidgen's avatar
Markus Scheidgen committed
188
189

        try:
190
191
            # save preliminary minimum calc metadata in case processing fails
            # successful processing will replace it with the actual metadata
192
            calc_with_metadata = datamodel.CalcWithMetadata(
193
194
195
196
197
198
                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()
199
            calc_with_metadata.upload_time = self.upload.upload_time
200
            calc_with_metadata.nomad_version = config.version
201
            calc_with_metadata.nomad_commit = config.commit
202
203
204
205
            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()

206
207
208
209
210
            if len(calc_with_metadata.files) >= config.auxfile_cutoff:
                self.warning(
                    'This calc has many aux files in its directory. '
                    'Have you placed many calculations in the same directory?')

Markus Scheidgen's avatar
Markus Scheidgen committed
211
212
213
214
            self.parsing()
            self.normalizing()
            self.archiving()
        finally:
Markus Scheidgen's avatar
Markus Scheidgen committed
215
            # close loghandler that was not closed due to failures
216
            try:
217
218
219
                if self._calc_proc_logwriter is not None:
                    self._calc_proc_logwriter.close()
                    self._calc_proc_logwriter = None
220
221
222
            except Exception as e:
                logger.error('could not close calculation proc log', exc_info=e)

223
224
225
226
    def fail(self, *errors, log_level=logging.ERROR, **kwargs):
        # in case of failure, index a minimum set of metadata and mark
        # processing failure
        try:
227
            calc_with_metadata = datamodel.CalcWithMetadata(**self.metadata)
228
229
230
231
232
233
234
235
            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
236
237
238
239
240
241
242

            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

243
244
245
246
247
248
249
250
            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)

251
252
253
254
    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
255
        if process_name == 'process_calc' or process_name == 're_process_calc' or process_name is None:
256
257
            self.upload.reload()
            self.upload.check_join()
Markus Scheidgen's avatar
Markus Scheidgen committed
258
259
260

    @task
    def parsing(self):
261
        context = dict(parser=self.parser, step=self.parser)
262
        logger = self.get_logger(**context)
263
        parser = parser_dict[self.parser]
264
        self.metadata['parser_name'] = self.parser
Markus Scheidgen's avatar
Markus Scheidgen committed
265

266
        with utils.timer(logger, 'parser executed', input_size=self.mainfile_file.size):
267
268
269
270
271
272
273
            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)
274
                return
Markus Scheidgen's avatar
Markus Scheidgen committed
275

276
277
        # add the non code specific calc metadata to the backend
        # all other quantities have been determined by parsers/normalizers
278
        self._parser_backend.openNonOverlappingSection('section_entry_info')
279
        self._parser_backend.addValue('upload_id', self.upload_id)
280
        self._parser_backend.addValue('calc_id', self.calc_id)
281
        self._parser_backend.addValue('calc_hash', self.metadata['calc_hash'])
282
        self._parser_backend.addValue('mainfile', self.mainfile)
283
        self._parser_backend.addValue('parser_name', self.parser)
284
285
286
287
288
289
290
291
292
293
294
295
        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)
296

Markus Scheidgen's avatar
Markus Scheidgen committed
297
        if self._parser_backend.status[0] != 'ParseSuccess':
298
            logger.error(self._parser_backend.status[1])
Markus Scheidgen's avatar
Markus Scheidgen committed
299
            error = self._parser_backend.status[1]
300
            self.fail(error, level=logging.INFO, **context)
301
302
303
304
305
306
307
308

    @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:
309
        self._parser_backend.openContext('/section_entry_info/0')
310
311
312
313
314
315
316
        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')
317
                self._parser_backend.addValue('number_of_archive_processor_warnings', len(warnings))
318
319
320
321
                self._parser_backend.addArrayValues('archive_processor_warnings', [str(warning) for warning in warnings])
            else:
                self._parser_backend.addValue('archive_processor_status', 'Success')
        else:
322
            errors = self._parser_backend.status[1]
323
324
325
            self._parser_backend.addValue('archive_processor_error', str(errors))

        self._parser_backend.closeNonOverlappingSection('section_archive_processing_info')
326
        self._parser_backend.closeContext('/section_entry_info/0')
Markus Scheidgen's avatar
Markus Scheidgen committed
327
328
329
330

    @task
    def normalizing(self):
        for normalizer in normalizers:
331
332
333
            if normalizer.domain != config.domain:
                continue

Markus Scheidgen's avatar
Markus Scheidgen committed
334
            normalizer_name = normalizer.__name__
335
            context = dict(normalizer=normalizer_name, step=normalizer_name)
336
            logger = self.get_logger(**context)
Markus Scheidgen's avatar
Markus Scheidgen committed
337
338

            with utils.timer(
339
                    logger, 'normalizer executed', input_size=self.mainfile_file.size):
340
                with self.use_parser_backend(normalizer_name) as backend:
341
342
343
344
345
346
347
                    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
348

349
350
            failed = self._parser_backend.status[0] != 'ParseSuccess'
            if failed:
351
                logger.error(self._parser_backend.status[1])
Markus Scheidgen's avatar
Markus Scheidgen committed
352
                error = self._parser_backend.status[1]
353
                self.fail(error, level=logging.WARNING, error=error, **context)
354
355
356
357
                break
            else:
                logger.debug(
                    'completed normalizer successfully', normalizer=normalizer_name)
Markus Scheidgen's avatar
Markus Scheidgen committed
358
359
360

    @task
    def archiving(self):
361
362
        logger = self.get_logger()

363
        calc_with_metadata = datamodel.CalcWithMetadata(**self.metadata)
364
        calc_with_metadata.apply_domain_metadata(self._parser_backend)
365
        calc_with_metadata.processed = True
366

367
368
        # persist the calc metadata
        with utils.timer(logger, 'saved calc metadata', step='metadata'):
369
            self.metadata = calc_with_metadata.to_dict()
370
371

        # index in search
372
        with utils.timer(logger, 'indexed', step='index'):
373
            search.Entry.from_calc_with_metadata(calc_with_metadata).save()
374

375
        # persist the archive
Markus Scheidgen's avatar
Markus Scheidgen committed
376
        with utils.timer(
377
                logger, 'archived', step='archive',
Markus Scheidgen's avatar
Markus Scheidgen committed
378
                input_size=self.mainfile_file.size) as log_data:
379
            with self.upload_files.archive_file(self.calc_id, 'wt') as out:
380
                self._parser_backend.write_json(out, pretty=True, root_sections=Domain.instance.root_sections)
381

382
            log_data.update(archive_size=self.upload_files.archive_file_object(self.calc_id).size)
Markus Scheidgen's avatar
Markus Scheidgen committed
383
384
385
386

        # close loghandler
        if self._calc_proc_logwriter is not None:
            with utils.timer(
387
                    logger, 'archived log', step='logs',
Markus Scheidgen's avatar
Markus Scheidgen committed
388
                    input_size=self.mainfile_file.size) as log_data:
389
                self._calc_proc_logwriter_ctx.__exit__(None, None, None)  # pylint: disable=E1101
390
                self._calc_proc_logwriter = None
Markus Scheidgen's avatar
Markus Scheidgen committed
391

392
                log_data.update(log_size=self.upload_files.archive_log_file_object(self.calc_id).size)
Markus Scheidgen's avatar
Markus Scheidgen committed
393

394
    def __str__(self):
395
        return 'calc %s calc_id=%s upload_id%s' % (super().__str__(), self.calc_id, self.upload_id)
396

397

398
class Upload(Proc):
Markus Scheidgen's avatar
Markus Scheidgen committed
399
400
401
402
403
404
    """
    Represents uploads in the databases. Provides persistence access to the files storage,
    and processing state.

    Attributes:
        name: optional user provided upload name
405
406
        upload_path: the path were the uploaded files was stored
        temporary: True if the uploaded file should be removed after extraction
407
        metadata: optional user provided additional meta data
Markus Scheidgen's avatar
Markus Scheidgen committed
408
409
        upload_id: the upload id generated by the database
        upload_time: the timestamp when the system realised the upload
410
        user_id: the id of the user that created this upload
Markus Scheidgen's avatar
Markus Scheidgen committed
411
412
413
414
    """
    id_field = 'upload_id'

    upload_id = StringField(primary_key=True)
415
416
    upload_path = StringField(default=None)
    temporary = BooleanField(default=False)
Markus Scheidgen's avatar
Markus Scheidgen committed
417
418
419

    name = StringField(default=None)
    upload_time = DateTimeField()
420
    user_id = StringField(required=True)
421
422
    published = BooleanField(default=False)
    publish_time = DateTimeField()
423
    last_update = DateTimeField()
Markus Scheidgen's avatar
Markus Scheidgen committed
424
425
426

    meta: Any = {
        'indexes': [
427
            'user_id', 'tasks_status', 'process_status', 'published', 'upload_time'
Markus Scheidgen's avatar
Markus Scheidgen committed
428
429
430
431
432
        ]
    }

    def __init__(self, **kwargs):
        super().__init__(**kwargs)
433
        self._upload_files: ArchiveBasedStagingUploadFiles = None
Markus Scheidgen's avatar
Markus Scheidgen committed
434

435
436
    @property
    def metadata(self) -> dict:
437
438
439
440
441
442
        # 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
443
444
445

    @metadata.setter
    def metadata(self, data: dict) -> None:
446
447
448
        # 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
449

Markus Scheidgen's avatar
Markus Scheidgen committed
450
    @classmethod
451
    def get(cls, id: str, include_published: bool = True) -> 'Upload':
452
        upload = cls.get_by_id(id, 'upload_id')
453
        if upload is not None:
454
455
456
            return upload

        raise KeyError()
Markus Scheidgen's avatar
Markus Scheidgen committed
457
458

    @classmethod
459
460
461
    def user_uploads(cls, user: coe_repo.User, **kwargs) -> List['Upload']:
        """ Returns all uploads for the given user. Kwargs are passed to mongo query. """
        return cls.objects(user_id=str(user.user_id), **kwargs)
Markus Scheidgen's avatar
Markus Scheidgen committed
462

463
464
465
466
    @property
    def uploader(self):
        return coe_repo.User.from_user_id(self.user_id)

Markus Scheidgen's avatar
Markus Scheidgen committed
467
468
    def get_logger(self, **kwargs):
        logger = super().get_logger()
469
        logger = logger.bind(upload_id=self.upload_id, upload_name=self.name, **kwargs)
Markus Scheidgen's avatar
Markus Scheidgen committed
470
471
472
473
474
475
476
477
        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.
478
479

        Arguments:
480
            user (coe_repo.User): The user that created the upload.
Markus Scheidgen's avatar
Markus Scheidgen committed
481
        """
482
        user: coe_repo.User = kwargs['user']
483
484
485
        del(kwargs['user'])
        if 'upload_id' not in kwargs:
            kwargs.update(upload_id=utils.create_uuid())
486
        kwargs.update(user_id=str(user.user_id))
Markus Scheidgen's avatar
Markus Scheidgen committed
487
        self = super().create(**kwargs)
488

Markus Scheidgen's avatar
Markus Scheidgen committed
489
        self._continue_with('uploading')
490

Markus Scheidgen's avatar
Markus Scheidgen committed
491
492
        return self

493
494
495
496
497
    def delete(self):
        """ Deletes this upload process state entry and its calcs. """
        Calc.objects(upload_id=self.upload_id).delete()
        super().delete()

498
    def delete_upload_local(self, with_coe_repo: bool = False):
499
500
        """
        Deletes of the upload, including its processing state and
501
        staging files. Local version without celery processing.
502
503
504
505
        """
        logger = self.get_logger()

        with utils.lnr(logger, 'staged upload delete failed'):
506

507
508
509
510
511
            if with_coe_repo and self.published:
                with utils.timer(
                        logger, 'upload deleted from repo db', step='repo',
                        upload_size=self.upload_files.size):
                    coe_repo.Upload.delete(self.upload_id)
512

513
            with utils.timer(
514
                    logger, 'upload deleted from index', step='index',
515
                    upload_size=self.upload_files.size):
516
                search.delete_upload(self.upload_id)
517

518
            with utils.timer(
519
                    logger, 'staged upload deleted', step='files',
520
521
522
523
                    upload_size=self.upload_files.size):
                self.upload_files.delete()
                self.delete()

524
525
526
527
528
529
530
531
    @process
    def delete_upload(self, with_coe_repo: bool = False):
        """
        Deletes of the upload, including its processing state and
        staging files. This starts the celery process of deleting the upload.
        """
        self.delete_upload_local(with_coe_repo=with_coe_repo)

532
        return True  # do not save the process status on the delete upload
533

534
    @process
535
    def publish_upload(self):
536
537
538
539
540
        """
        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.
541
542
543

        If the upload is already published (i.e. re-publish), it will update user metadata from
        repository db, publish to repository db if not exists, update the search index.
544
        """
545
546
        assert self.processed_calcs > 0

547
        logger = self.get_logger()
548
        logger.info('started to publish')
549

550
551
        with utils.lnr(logger, '(re-)publish failed'):
            upload_with_metadata = self.to_upload_with_metadata(self.metadata)
552
            calcs = upload_with_metadata.calcs
553

554
            if config.repository_db.publish_enabled:
555
556
557
558
559
560
561
                if config.repository_db.mode == 'coe' and isinstance(self.upload_files, StagingUploadFiles):
                    with utils.timer(
                            logger, 'coe extracted raw-file copy created', step='repo',
                            upload_size=self.upload_files.size):

                        self.upload_files.create_extracted_copy()

562
563
564
565
566
567
568
                coe_upload = coe_repo.Upload.from_upload_id(upload_with_metadata.upload_id)
                if coe_upload is None:
                    with utils.timer(
                            logger, 'upload added to repository', step='repo',
                            upload_size=self.upload_files.size):
                        coe_upload = coe_repo.Upload.publish(upload_with_metadata)

569
                with utils.timer(
570
                        logger, 'upload PIDs read from repository', step='repo',
571
                        upload_size=self.upload_files.size):
572
573
                    for calc, coe_calc in zip(calcs, coe_upload.calcs):
                        calc.pid = coe_calc.coe_calc_id
574

575
            with utils.timer(
576
                    logger, 'upload metadata updated', step='metadata',
577
                    upload_size=self.upload_files.size):
578
579

                def create_update(calc):
580
                    calc.published = True
581
582
583
584
585
                    return UpdateOne(
                        {'_id': calc.calc_id},
                        {'$set': {'metadata': calc.to_dict()}})

                Calc._get_collection().bulk_write([create_update(calc) for calc in calcs])
586

587
588
589
590
591
            if isinstance(self.upload_files, StagingUploadFiles):
                with utils.timer(
                        logger, 'staged upload files packed', step='pack',
                        upload_size=self.upload_files.size):
                    self.upload_files.pack(upload_with_metadata)
592
593
594
595

            with utils.timer(
                    logger, 'index updated', step='index',
                    upload_size=self.upload_files.size):
596
                search.publish(calcs)
597

598
599
600
601
602
603
604
605
606
607
608
            if isinstance(self.upload_files, StagingUploadFiles):
                with utils.timer(
                        logger, 'staged upload deleted', step='delete staged',
                        upload_size=self.upload_files.size):
                    self.upload_files.delete()
                    self.published = True
                    self.publish_time = datetime.now()
                    self.last_update = datetime.now()
                    self.save()
            else:
                self.last_update = datetime.now()
609
                self.save()
610

611
612
613
614
615
616
617
618
619
620
621
622
    @process
    def re_process_upload(self):
        """
        Runs the distributed process of fully reparsing/renormalizing an existing and
        already published upload. Will renew the archive part of the upload and update
        mongo and elastic search entries.

        TODO this implementation does not do any re-matching. This will be more complex
        due to handling of new or missing matches.
        """
        assert self.published

623
624
625
        logger = self.get_logger()
        logger.info('started to re-process')

626
627
628
629
630
631
632
633
634
635
        self.reset()
        # mock the steps of actual processing
        self._continue_with('uploading')

        # extract the published raw files into a staging upload files instance
        self._continue_with('extracting')
        public_upload_files = cast(PublicUploadFiles, self.upload_files)
        public_upload_files.to_staging_upload_files(create=True)

        self._continue_with('parse_all')
636
        for calc in Calc.objects(upload_id=self.upload_id):
637
638
639
640
641
642
            calc.reset()
            calc.re_process_calc()

        # the packing and removing of the staging upload files, will be trigged by
        # the 'cleanup' task after processing all calcs

Markus Scheidgen's avatar
Markus Scheidgen committed
643
    @process
644
    def process_upload(self):
Markus Scheidgen's avatar
Markus Scheidgen committed
645
646
647
648
649
650
651
        self.extracting()
        self.parse_all()

    @task
    def uploading(self):
        pass

652
    @property
653
654
    def upload_files(self) -> UploadFiles:
        upload_files_class = ArchiveBasedStagingUploadFiles if not self.published else PublicUploadFiles
655
        kwargs = dict(upload_path=self.upload_path) if not self.published else {}
656
657
658

        if not self._upload_files or not isinstance(self._upload_files, upload_files_class):
            self._upload_files = upload_files_class(
659
                self.upload_id, is_authorized=lambda: True, **kwargs)
660

661
        return self._upload_files
662

663
664
665
666
667
    @property
    def staging_upload_files(self) -> ArchiveBasedStagingUploadFiles:
        assert not self.published
        return cast(ArchiveBasedStagingUploadFiles, self.upload_files)

Markus Scheidgen's avatar
Markus Scheidgen committed
668
669
    @task
    def extracting(self):
670
671
672
673
674
        """
        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.
        """
675
676
677
678
679
        # 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
680
681
        logger = self.get_logger()
        try:
Markus Scheidgen's avatar
Markus Scheidgen committed
682
683
            with utils.timer(
                    logger, 'upload extracted', step='extracting',
684
685
                    upload_size=self.upload_files.size):
                self.upload_files.extract()
686
687
688
689

            if self.temporary:
                os.remove(self.upload_path)
                self.upload_path = None
690

691
        except KeyError:
692
            self.fail('processing requested for non existing upload', log_level=logging.ERROR)
693
694
695
            return
        except ExtractError:
            self.fail('bad .zip/.tar file', log_level=logging.INFO)
Markus Scheidgen's avatar
Markus Scheidgen committed
696
697
            return

698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
    def _preprocess_files(self, path):
        """
        Some files need preprocessing. Currently we need to add a stripped POTCAR version
        and always restrict/embargo the original.
        """
        if os.path.basename(path) == 'POTCAR':
            # create checksum
            hash = hashlib.sha224()
            with open(self.staging_upload_files.raw_file_object(path).os_path, 'rb') as orig_f:
                for line in orig_f.readlines():
                    hash.update(line)

            checksum = hash.hexdigest()

            # created stripped POTCAR
            stripped_path = path + '.stripped'
            with open(self.staging_upload_files.raw_file_object(stripped_path).os_path, 'wt') as stripped_f:
                stripped_f.write('Stripped POTCAR file. Checksum of original file (sha224): %s\n' % checksum)
            os.system(
                '''
                    awk < %s >> %s '
                    BEGIN { dump=1 }
                    /End of Dataset/ { dump=1 }
                    dump==1 { print }
                    /END of PSCTR/ { dump=0 }'
                ''' % (
                    self.staging_upload_files.raw_file_object(path).os_path,
                    self.staging_upload_files.raw_file_object(stripped_path).os_path))

727
    def match_mainfiles(self) -> Generator[Tuple[str, object], None, None]:
728
729
730
731
732
733
734
        """
        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
        """
735
        directories_with_match: Dict[str, str] = dict()
736
737
        upload_files = self.staging_upload_files
        for filename in upload_files.raw_file_manifest():
738
            self._preprocess_files(filename)
739
            try:
740
                parser = match_parser(filename, upload_files)
741
                if parser is not None:
742
743
                    directory = os.path.dirname(filename)
                    if directory in directories_with_match:
744
745
746
                        # TODO this might give us the chance to store directory based relationship
                        # between calcs for the future?
                        pass
747
748
                    else:
                        directories_with_match[directory] = filename
749
750

                    yield filename, parser
751
752
753
754
            except Exception as e:
                self.get_logger().error(
                    'exception while matching pot. mainfile',
                    mainfile=filename, exc_info=e)
755

Markus Scheidgen's avatar
Markus Scheidgen committed
756
757
    @task
    def parse_all(self):
758
        """
759
        Identified mainfile/parser combinations among the upload's files, creates
760
761
        respective :class:`Calc` instances, and triggers their processing.
        """
762
763
        logger = self.get_logger()

Markus Scheidgen's avatar
Markus Scheidgen committed
764
765
        with utils.timer(
                logger, 'upload extracted', step='matching',
766
767
                upload_size=self.upload_files.size):
            for filename, parser in self.match_mainfiles():
768
                calc = Calc.create(
769
                    calc_id=self.upload_files.calc_id(filename),
770
                    mainfile=filename, parser=parser.name,
771
                    worker_hostname=self.worker_hostname,
772
773
                    upload_id=self.upload_id)

774
                calc.process_calc()
Markus Scheidgen's avatar
Markus Scheidgen committed
775

776
    def on_process_complete(self, process_name):
777
        if process_name == 'process_upload' or process_name == 're_process_upload':
778
779
780
781
782
783
784
785
786
787
            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()
788
789
790

    def join(self):
        self.cleanup()
Markus Scheidgen's avatar
Markus Scheidgen committed
791

792
793
794
795
796
797
798
    @property
    def gui_url(self):
        base = config.api_url()[:-3]
        if base.endswith('/'):
            base = base[:-1]
        return '%s/uploads/' % base

799
    def _cleanup_after_processing(self):
800
801
802
803
804
805
        # send email about process finish
        user = self.uploader
        name = '%s %s' % (user.first_name, user.last_name)
        message = '\n'.join([
            'Dear %s,' % name,
            '',
806
            'your data %suploaded at %s has completed processing.' % (
807
                '"%s" ' % self.name if self.name else '', self.upload_time.isoformat()),  # pylint: disable=no-member
808
809
810
811
812
            'You can review your data on your upload page: %s' % self.gui_url,
            '',
            'If you encouter any issues with your upload, please let us know and replay to this email.',
            '',
            'The nomad team'
813
        ])
814
815
816
817
818
819
820
        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
821

822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
    def _cleanup_after_re_processing(self):
        logger = self.get_logger()
        logger.info('started to repack re-processed upload')

        staging_upload_files = self.upload_files.to_staging_upload_files()

        with utils.timer(
                logger, 'reprocessed staged upload packed', step='delete staged',
                upload_size=self.upload_files.size):

            staging_upload_files.pack(self.to_upload_with_metadata())

        with utils.timer(
                logger, 'reprocessed staged upload deleted', step='delete staged',
                upload_size=self.upload_files.size):

            staging_upload_files.delete()
            self.last_update = datetime.now()
            self.save()

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

        if self.current_process == 're_process_upload':
            self._cleanup_after_re_processing()
        else:
            self._cleanup_after_processing()

851
852
853
    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
854
    @property
855
    def processed_calcs(self):
856
        return Calc.objects(upload_id=self.upload_id, tasks_status__in=[SUCCESS, FAILURE]).count()
857
858
859
860
861
862
863

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

    @property
    def failed_calcs(self):
864
        return Calc.objects(upload_id=self.upload_id, tasks_status=FAILURE).count()
865

866
867
    @property
    def pending_calcs(self):
868
        return Calc.objects(upload_id=self.upload_id, tasks_status=PENDING).count()
869

870
871
872
    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
873

874
875
876
877
878
879
    @property
    def outdated_calcs(self):
        return Calc.objects(
            upload_id=self.upload_id, tasks_status=SUCCESS,
            metadata__nomad_version__ne=config.version)

880
881
    @property
    def calcs(self):
Markus Scheidgen's avatar
Markus Scheidgen committed
882
        return Calc.objects(upload_id=self.upload_id, tasks_status=SUCCESS)
Markus Scheidgen's avatar
Markus Scheidgen committed
883

884
    def to_upload_with_metadata(self, user_metadata: dict = None) -> UploadWithMetadata:
885
        # prepare user metadata per upload and per calc
886
887
888
        if user_metadata is not None:
            calc_metadatas: Dict[str, Any] = dict()
            upload_metadata: Dict[str, Any] = dict()
889

890
            upload_metadata.update(user_metadata)
891
892
            if 'calculations' in upload_metadata:
                del(upload_metadata['calculations'])
Markus Scheidgen's avatar
Markus Scheidgen committed
893

894
            for calc in user_metadata.get('calculations', []):  # pylint: disable=no-member
895
                calc_metadatas[calc['mainfile']] = calc
Markus Scheidgen's avatar
Markus Scheidgen committed
896

897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
            user_upload_time = upload_metadata.get('_upload_time', None)

            def get_metadata(calc: Calc):
                """
                Assemble metadata from calc's processed calc metadata and the uploads
                user metadata.
                """
                calc_data = calc.metadata
                calc_with_metadata = datamodel.CalcWithMetadata(**calc_data)
                calc_metadata = dict(upload_metadata)
                calc_metadata.update(calc_metadatas.get(calc.mainfile, {}))
                calc_with_metadata.apply_user_metadata(calc_metadata)

                return calc_with_metadata
        else:
            user_upload_time = None

            def get_metadata(calc: Calc):
                return datamodel.CalcWithMetadata(**calc.metadata)

Markus Scheidgen's avatar
Markus Scheidgen committed
917
918
919
920
921
        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)

922
        result.calcs = [get_metadata(calc) for calc in Calc.objects(upload_id=self.upload_id)]
923

Markus Scheidgen's avatar
Markus Scheidgen committed
924
        return result
925

926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
    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

956
    def __str__(self):
957
        return 'upload %s upload_id%s' % (super().__str__(), self.upload_id)