data.py 14.9 KB
Newer Older
Markus Scheidgen's avatar
Markus Scheidgen committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
# 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:
"""

from typing import List, Any
from datetime import datetime
29
from elasticsearch.exceptions import NotFoundError
Markus Scheidgen's avatar
Markus Scheidgen committed
30
from mongoengine import StringField, BooleanField, DateTimeField, DictField, IntField
Markus Scheidgen's avatar
Markus Scheidgen committed
31
import logging
32
import base64
Markus Scheidgen's avatar
Markus Scheidgen committed
33
import time
Markus Scheidgen's avatar
Markus Scheidgen committed
34

35
from nomad import config, utils
Markus Scheidgen's avatar
Markus Scheidgen committed
36
from nomad.files import UploadFile, ArchiveFile
Markus Scheidgen's avatar
Markus Scheidgen committed
37
from nomad.repo import RepoCalc
Markus Scheidgen's avatar
Markus Scheidgen committed
38
from nomad.user import User
39
from nomad.processing.base import Proc, Chord, process, task, PENDING, SUCCESS, FAILURE, RUNNING
Markus Scheidgen's avatar
Markus Scheidgen committed
40
from nomad.parsing import parsers, parser_dict
Markus Scheidgen's avatar
Markus Scheidgen committed
41
from nomad.normalizing import normalizers
Markus Scheidgen's avatar
Markus Scheidgen committed
42
from nomad.utils import lnr
Markus Scheidgen's avatar
Markus Scheidgen committed
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72


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': [
73
            'upload_id', 'mainfile', 'code', 'parser', 'status'
Markus Scheidgen's avatar
Markus Scheidgen committed
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
        ]
    }

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self._parser_backend = None
        self._upload = None

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

    def delete(self):
        """
        Delete this calculation and all associated data. This includes all files,
        the archive, and this search index entry.
90
        TODO is this needed? Or do we always delete hole uploads in bulk.
Markus Scheidgen's avatar
Markus Scheidgen committed
91
92
93
        """
        # delete the archive
        if self.archive_id is not None:
94
            ArchiveFile(self.archive_id).delete()
Markus Scheidgen's avatar
Markus Scheidgen committed
95
96

        # delete the search index entry
97
98
99
100
101
102
        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
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130

        # 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,
            upload_hash=upload_hash, calc_hash=calc_hash, **kwargs)
        return logger

    @property
    def json_dict(self):
        """ A json serializable dictionary representation. """
        data = {
            'archive_id': self.archive_id,
            'mainfile': self.mainfile,
            'upload_id': self.upload_id,
            'parser': self.parser
        }
        data.update(super().json_dict)
        return {key: value for key, value in data.items() if value is not None}

    @process
    def process(self):
        self._upload = Upload.get(self.upload_id)
        if self._upload is None:
Markus Scheidgen's avatar
Markus Scheidgen committed
131
            self.get_logger().error('calculation upload does not exist')
Markus Scheidgen's avatar
Markus Scheidgen committed
132
133
134
135
136
137

        try:
            self.parsing()
            self.normalizing()
            self.archiving()
        finally:
138
            self._upload.completed_child()
Markus Scheidgen's avatar
Markus Scheidgen committed
139
140
141

    @task
    def parsing(self):
142
143
        parser = parser_dict[self.parser]
        self._parser_backend = parser.run(self.mainfile_tmp_path, logger=self.get_logger())
Markus Scheidgen's avatar
Markus Scheidgen committed
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
        if self._parser_backend.status[0] != 'ParseSuccess':
            error = self._parser_backend.status[1]
            self.fail(error, level=logging.DEBUG)

    @task
    def normalizing(self):
        for normalizer in normalizers:
            normalizer_name = normalizer.__name__
            normalizer(self._parser_backend).normalize()
            if self._parser_backend.status[0] != 'ParseSuccess':
                error = self._parser_backend.status[1]
                self.fail(error, normalizer=normalizer_name, level=logging.WARNING)
                return
            self.get_logger().debug(
                'completed normalizer successfully', normalizer=normalizer_name)

    @task
    def archiving(self):
        upload_hash, calc_hash = self.archive_id.split('/')
163
        additional = dict(
Markus Scheidgen's avatar
Markus Scheidgen committed
164
            mainfile=self.mainfile,
165
166
167
168
            upload_time=self._upload.upload_time,
            staging=True,
            restricted=False,
            user_id=self._upload.user_id)
169
170
171
172
173
174
175
        # persist to elastic search
        RepoCalc.create_from_backend(
            self._parser_backend,
            additional=additional,
            upload_hash=upload_hash,
            calc_hash=calc_hash,
            upload_id=self.upload_id)
Markus Scheidgen's avatar
Markus Scheidgen committed
176
177

        # persist the archive
178
        with ArchiveFile(self.archive_id).write_archive_json() as out:
Markus Scheidgen's avatar
Markus Scheidgen committed
179
180
181
            self._parser_backend.write_json(out, pretty=True)


182
class Upload(Chord):
Markus Scheidgen's avatar
Markus Scheidgen committed
183
184
185
186
187
188
189
190
191
192
193
194
    """
    Represents uploads in the databases. Provides persistence access to the files storage,
    and processing state.

    Attributes:
        name: optional user provided upload name
        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
195
        user_id: the id of the user that created this upload
Markus Scheidgen's avatar
Markus Scheidgen committed
196
197
198
199
200
201
202
203
204
205
206
207
208
209
    """
    id_field = 'upload_id'

    upload_id = StringField(primary_key=True)

    name = StringField(default=None)
    additional_metadata = DictField(default=None)

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

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

210
    user_id = StringField(required=True)
211
212
    upload_url = StringField(default=None)
    upload_command = StringField(default=None)
Markus Scheidgen's avatar
Markus Scheidgen committed
213

214
215
    _initiated_parsers = IntField(default=-1)

Markus Scheidgen's avatar
Markus Scheidgen committed
216
217
    meta: Any = {
        'indexes': [
218
            'upload_hash', 'user_id', 'status'
Markus Scheidgen's avatar
Markus Scheidgen committed
219
220
221
222
223
224
225
226
227
228
229
230
231
232
        ]
    }

    def __init__(self, **kwargs):
        super().__init__(**kwargs)
        self._upload = None

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

    @classmethod
    def user_uploads(cls, user: User) -> List['Upload']:
        """ Returns all uploads for the given user. Currently returns all uploads. """
Markus Scheidgen's avatar
Markus Scheidgen committed
233
        return cls.objects(user_id=user.email, in_staging=True)
Markus Scheidgen's avatar
Markus Scheidgen committed
234
235
236
237
238
239
240
241
242
243
244
245
246
247

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

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

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

        with lnr(logger, 'delete upload file'):
            try:
248
                UploadFile(self.upload_id).delete()
Markus Scheidgen's avatar
Markus Scheidgen committed
249
250
251
252
253
254
255
256
257
            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'):
258
            # delete archive files
259
            ArchiveFile.delete_archives(upload_hash=self.upload_hash)
260
261

            # delete repo entries
Markus Scheidgen's avatar
Markus Scheidgen committed
262
            RepoCalc.delete_upload(upload_id=self.upload_id)
263
264
265

            # delete calc processings
            Calc.objects(upload_id=self.upload_id).delete()
Markus Scheidgen's avatar
Markus Scheidgen committed
266
267
268
269

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

270
271
    @classmethod
    def _external_objects_url(cls, url):
272
        """ Replaces the given internal object storage url with an URL that allows
273
274
            external access.
        """
275
        return 'http://%s:%s%s%s' % (config.services.api_host, config.services.api_port, config.services.api_base_path, url)
276

Markus Scheidgen's avatar
Markus Scheidgen committed
277
278
279
280
281
282
    @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.
283
284
285

        Arguments:
            user (User): The user that created the upload.
Markus Scheidgen's avatar
Markus Scheidgen committed
286
        """
287
288
289
290
291
        user: User = kwargs['user']
        del(kwargs['user'])
        if 'upload_id' not in kwargs:
            kwargs.update(upload_id=utils.create_uuid())
        kwargs.update(user_id=user.email)
Markus Scheidgen's avatar
Markus Scheidgen committed
292
        self = super().create(**kwargs)
293
294
295
296

        basic_auth_token = base64.b64encode(b'%s:' % user.generate_auth_token()).decode('utf-8')

        self.upload_url = cls._external_objects_url('/uploads/%s/file' % self.upload_id)
Markus Scheidgen's avatar
Markus Scheidgen committed
297
        self.upload_command = 'curl -H "Authorization: Basic %s" "%s" --upload-file local_file' % (
298
299
            basic_auth_token, self.upload_url)

Markus Scheidgen's avatar
Markus Scheidgen committed
300
        self._continue_with('uploading')
301

Markus Scheidgen's avatar
Markus Scheidgen committed
302
303
304
305
306
307
308
309
310
        return self

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

Markus Scheidgen's avatar
Markus Scheidgen committed
311
    def unstage(self):
312
        self.get_logger().info('unstage')
Markus Scheidgen's avatar
Markus Scheidgen committed
313
        self.in_staging = False
314
        RepoCalc.unstage(upload_id=self.upload_id)
Markus Scheidgen's avatar
Markus Scheidgen committed
315
316
        self.save()

Markus Scheidgen's avatar
Markus Scheidgen committed
317
318
319
320
321
322
323
    @property
    def json_dict(self) -> dict:
        """ A json serializable dictionary representation. """
        data = {
            'name': self.name,
            'additional_metadata': self.additional_metadata,
            'upload_id': self.upload_id,
324
            'upload_url': self.upload_url,
325
            'upload_command': self.upload_command,
Markus Scheidgen's avatar
Markus Scheidgen committed
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
            'upload_time': self.upload_time.isoformat() if self.upload_time is not None else None,
            'is_stale': self.is_stale,
        }
        data.update(super().json_dict)
        return {key: value for key, value in data.items() if value is not None}

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

    @task
    def uploading(self):
        pass

    @task
    def extracting(self):
        logger = self.get_logger()
        try:
345
346
            self._upload = UploadFile(self.upload_id)
            self._upload.extract()
Markus Scheidgen's avatar
Markus Scheidgen committed
347
348
349
350
351
352
353
            logger.debug('opened upload')
        except KeyError as e:
            self.fail('process request for non existing upload', level=logging.INFO)
            return

        try:
            self.upload_hash = self._upload.hash()
354
        except Exception as e:
Markus Scheidgen's avatar
Markus Scheidgen committed
355
356
357
            self.fail('could not create upload hash', e)
            return

Markus Scheidgen's avatar
Markus Scheidgen committed
358
        if RepoCalc.upload_exists(self.upload_hash):
Markus Scheidgen's avatar
Markus Scheidgen committed
359
360
361
362
363
364
            self.fail('The same file was already uploaded and processed.', level=logging.INFO)
            return

    @task
    def parse_all(self):
        # TODO: deal with multiple possible parser specs
365
        total_calcs = 0
Markus Scheidgen's avatar
Markus Scheidgen committed
366
367
368
369
370
371
372
373
374
375
376
377
        for filename in self._upload.filelist:
            for parser in parsers:
                try:
                    if parser.is_mainfile(filename, lambda fn: self._upload.open_file(fn)):
                        tmp_mainfile = self._upload.get_path(filename)
                        calc = Calc.create(
                            archive_id='%s/%s' % (self.upload_hash, utils.hash(filename)),
                            mainfile=filename, parser=parser.name,
                            mainfile_tmp_path=tmp_mainfile,
                            upload_id=self.upload_id)

                        calc.process()
378
                        total_calcs += 1
Markus Scheidgen's avatar
Markus Scheidgen committed
379
                except Exception as e:
380
                    self.warning(
Markus Scheidgen's avatar
Markus Scheidgen committed
381
382
383
                        'exception while matching pot. mainfile',
                        mainfile=filename, exc_info=e)

384
385
386
387
388
        # 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
389
390
391
392

    @task
    def cleanup(self):
        try:
393
            upload = UploadFile(self.upload_id)
Markus Scheidgen's avatar
Markus Scheidgen committed
394
        except KeyError as e:
Markus Scheidgen's avatar
Markus Scheidgen committed
395
            self.fail('Upload does not exist', exc_info=e)
Markus Scheidgen's avatar
Markus Scheidgen committed
396
397
            return

398
        upload.remove_extract()
Markus Scheidgen's avatar
Markus Scheidgen committed
399
400
401
        self.get_logger().debug('closed upload')

    @property
402
403
404
405
406
407
408
409
410
411
412
    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()

413
414
415
416
    @property
    def pending_calcs(self):
        return Calc.objects(upload_id=self.upload_id, status=PENDING).count()

417
418
    def all_calcs(self, start, end, order_by='mainfile'):
        return Calc.objects(upload_id=self.upload_id)[start:end].order_by(order_by)
419

Markus Scheidgen's avatar
Markus Scheidgen committed
420
    @staticmethod
421
422
423
424
425
    def repair_all():
        """
        Utitlity function that will look for suspiciously looking conditions in
        all uncompleted downloads. It ain't a perfect world.
        """
426
427
428
        # TODO this was added as a quick fix to #37.
        # Even though it might be strictly necessary, there should be a tested backup
        # solution for it Chords to not work properly due to failed to fail processings
429
430
431
432
433
434
435
436
437
438
        uploads = Upload.objects(status__in=[PENDING, RUNNING])
        for upload in uploads:
            completed = upload.processed_calcs
            total = upload.total
            pending = upload.pending_calcs

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