submission_processor.py 4.58 KB
Newer Older
Peter Stanko's avatar
Peter Stanko committed
1
2
import json
import random
Peter Stanko's avatar
Peter Stanko committed
3
from pathlib import Path
Peter Stanko's avatar
Peter Stanko committed
4
from typing import Union
Peter Stanko's avatar
Peter Stanko committed
5
6
7

from storage import UploadedEntity

Peter Stanko's avatar
Peter Stanko committed
8
9
10
import portal.tools.worker_client
from portal import logger, tools
from portal.database import Project, Submission, SubmissionState, Worker
Peter Stanko's avatar
Peter Stanko committed
11
12
from portal.service import general

Peter Stanko's avatar
Peter Stanko committed
13
log = logger.get_logger(__name__)
Peter Stanko's avatar
Peter Stanko committed
14
15
16
17
18
19
20
21
22
23
24
25
26


class SubmissionProcessor:
    def __init__(self, submission: Submission, params: dict = None):
        self._submission = submission
        self._params = params

    @property
    def submission(self) -> Submission:
        return self._submission

    @property
    def params(self) -> dict:
Peter Stanko's avatar
Peter Stanko committed
27
        return self.submission.parameters
Peter Stanko's avatar
Peter Stanko committed
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70

    @property
    def project(self) -> Project:
        return self.submission.project

    @property
    def celery(self):
        from portal.async_celery import celery_app
        return celery_app

    @property
    def storage(self):
        from portal import storage
        return storage

    def reset_task_id(self, state=None):
        if state is not None:
            self.submission.state = state

        self.submission.async_task_id = None
        general.write_entity(self.submission)

    def submission_enqueue_ended(self):
        log.info(f"[ASYNC] Submission enqueue ended: {self.submission}")
        self.reset_task_id(state=SubmissionState.QUEUED)

    def get_delay_for_submission(self):
        log.debug(f"[ASYNC] Submission delay: {self.submission}")
        time_wait = self.project.config.submissions_cancellation_period
        return time_wait

    def dispatch_submission_processing(self):
        delay = self.get_delay_for_submission()
        args = (self.submission.id, self.params)
        from .tasks import start_processing_submission
        start_processing_submission.apply_async(args=args, countdown=delay)

    def submission_store_ended(self, version: str):
        log.info(f"[ASYNC] Submission preparation ended: {self.submission}")
        self.submission.source_hash = version
        self.reset_task_id(state=SubmissionState.READY)

    def download_submission(self):
Peter Stanko's avatar
Peter Stanko committed
71
72
        file_params = self.params['file_params']
        log.info(f"[ASYNC] Uploading submission: {self.submission} with {file_params}")
Peter Stanko's avatar
Peter Stanko committed
73
        updated_entity: UploadedEntity = self.storage. \
Peter Stanko's avatar
Peter Stanko committed
74
            submissions.create(entity_id=self.submission.id, **file_params)
Peter Stanko's avatar
Peter Stanko committed
75
76
77
78
79
80
81
82
83
84
        self.submission_store_ended(version=updated_entity.version)

    def clone(self, target):
        log.info(f"[ASYNC] Cloning submission: {self.submission} to {target}")
        self.storage.submissions.clone(self.submission.id, target.id)
        self.submission_store_ended(version=self.submission.source_hash)

    def send_to_worker(self):
        # TODO: implement processing
        log.info(f"[ASYNC] Sending submission to worker: {self.submission}")
Peter Stanko's avatar
Peter Stanko committed
85
86
        worker = self.schedule_submission_to_worker()
        self.execute_submission(worker)
Peter Stanko's avatar
Peter Stanko committed
87

Peter Stanko's avatar
Peter Stanko committed
88
89
90
91
    def upload_result(self, path, file_params):
        log.info(f"[ASYNC] Uploading result for the submission "
                 f"{self.submission.id} with {file_params}")
        self.storage.results.create(entity_id=self.submission.id, **file_params)
Peter Stanko's avatar
Peter Stanko committed
92
93
94
95
96
97
98
        Path(path).unlink()
        self.reset_task_id(SubmissionState.FINISHED)

    def process_submission(self):
        log.info(f"[ASYNC] Processing submission: {self.submission}")
        self.download_submission()
        self.dispatch_submission_processing()
99
100
101
102
103
104
105
106

    def revoke_task(self):
        log.info(f'[ASYNC] Submission processing cancelled {self.submission}')
        task_id = self.submission.async_task_id
        if task_id:
            self.celery.control.revoke(task_id=task_id, terminate=True)
            self.reset_task_id(state=SubmissionState.CANCELLED)
        # TODO: Storage clean up
Peter Stanko's avatar
Peter Stanko committed
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128

    # TODO implement - @mdujava
    # STUB: Select initialized worker
    def schedule_submission_to_worker(self) -> Worker:
        """Based on the features (worker tags) and preferences in project config
        schedule submission for the execution on initialized worker

        Returns(Worker): Worker instance on which the submission will be executed
        """
        workers = Worker.query.filter(Worker.is_initialized).all()
        if not workers:
            self._worker_not_available()
        worker = random.choice(workers)  # randomly select a worker
        return worker

    def execute_submission(self, worker: Worker):
        worker_client = tools.worker_client.WorkerClient(worker=worker)
        worker_client.execute_submission(self.submission)

    def _worker_not_available(self):
        log.warning(f"[PROC] Worker is no available for submission: {self.submission}")
        pass