|
|
|
from __future__ import absolute_import
|
|
|
|
import warnings
|
|
|
|
|
|
|
|
from apscheduler.jobstores.base import BaseJobStore, JobLookupError, ConflictingIdError
|
|
|
|
from apscheduler.util import maybe_ref, datetime_to_utc_timestamp, utc_timestamp_to_datetime
|
|
|
|
from apscheduler.job import Job
|
|
|
|
|
|
|
|
try:
|
|
|
|
import cPickle as pickle
|
|
|
|
except ImportError: # pragma: nocover
|
|
|
|
import pickle
|
|
|
|
|
|
|
|
try:
|
|
|
|
from bson.binary import Binary
|
|
|
|
from pymongo.errors import DuplicateKeyError
|
|
|
|
from pymongo import MongoClient, ASCENDING
|
|
|
|
except ImportError: # pragma: nocover
|
|
|
|
raise ImportError('MongoDBJobStore requires PyMongo installed')
|
|
|
|
|
|
|
|
|
|
|
|
class MongoDBJobStore(BaseJobStore):
|
|
|
|
"""
|
|
|
|
Stores jobs in a MongoDB database. Any leftover keyword arguments are directly passed to
|
|
|
|
pymongo's `MongoClient
|
|
|
|
<http://api.mongodb.org/python/current/api/pymongo/mongo_client.html#pymongo.mongo_client.MongoClient>`_.
|
|
|
|
|
|
|
|
Plugin alias: ``mongodb``
|
|
|
|
|
|
|
|
:param str database: database to store jobs in
|
|
|
|
:param str collection: collection to store jobs in
|
|
|
|
:param client: a :class:`~pymongo.mongo_client.MongoClient` instance to use instead of
|
|
|
|
providing connection arguments
|
|
|
|
:param int pickle_protocol: pickle protocol level to use (for serialization), defaults to the
|
|
|
|
highest available
|
|
|
|
"""
|
|
|
|
|
|
|
|
def __init__(self, database='apscheduler', collection='jobs', client=None,
|
|
|
|
pickle_protocol=pickle.HIGHEST_PROTOCOL, **connect_args):
|
|
|
|
super(MongoDBJobStore, self).__init__()
|
|
|
|
self.pickle_protocol = pickle_protocol
|
|
|
|
|
|
|
|
if not database:
|
|
|
|
raise ValueError('The "database" parameter must not be empty')
|
|
|
|
if not collection:
|
|
|
|
raise ValueError('The "collection" parameter must not be empty')
|
|
|
|
|
|
|
|
if client:
|
|
|
|
self.client = maybe_ref(client)
|
|
|
|
else:
|
|
|
|
connect_args.setdefault('w', 1)
|
|
|
|
self.client = MongoClient(**connect_args)
|
|
|
|
|
|
|
|
self.collection = self.client[database][collection]
|
|
|
|
|
|
|
|
def start(self, scheduler, alias):
|
|
|
|
super(MongoDBJobStore, self).start(scheduler, alias)
|
|
|
|
self.collection.create_index('next_run_time', sparse=True)
|
|
|
|
|
|
|
|
@property
|
|
|
|
def connection(self):
|
|
|
|
warnings.warn('The "connection" member is deprecated -- use "client" instead',
|
|
|
|
DeprecationWarning)
|
|
|
|
return self.client
|
|
|
|
|
|
|
|
def lookup_job(self, job_id):
|
|
|
|
document = self.collection.find_one(job_id, ['job_state'])
|
|
|
|
return self._reconstitute_job(document['job_state']) if document else None
|
|
|
|
|
|
|
|
def get_due_jobs(self, now):
|
|
|
|
timestamp = datetime_to_utc_timestamp(now)
|
|
|
|
return self._get_jobs({'next_run_time': {'$lte': timestamp}})
|
|
|
|
|
|
|
|
def get_next_run_time(self):
|
|
|
|
document = self.collection.find_one({'next_run_time': {'$ne': None}},
|
|
|
|
projection=['next_run_time'],
|
|
|
|
sort=[('next_run_time', ASCENDING)])
|
|
|
|
return utc_timestamp_to_datetime(document['next_run_time']) if document else None
|
|
|
|
|
|
|
|
def get_all_jobs(self):
|
|
|
|
jobs = self._get_jobs({})
|
|
|
|
self._fix_paused_jobs_sorting(jobs)
|
|
|
|
return jobs
|
|
|
|
|
|
|
|
def add_job(self, job):
|
|
|
|
try:
|
|
|
|
self.collection.insert_one({
|
|
|
|
'_id': job.id,
|
|
|
|
'next_run_time': datetime_to_utc_timestamp(job.next_run_time),
|
|
|
|
'job_state': Binary(pickle.dumps(job.__getstate__(), self.pickle_protocol))
|
|
|
|
})
|
|
|
|
except DuplicateKeyError:
|
|
|
|
raise ConflictingIdError(job.id)
|
|
|
|
|
|
|
|
def update_job(self, job):
|
|
|
|
changes = {
|
|
|
|
'next_run_time': datetime_to_utc_timestamp(job.next_run_time),
|
|
|
|
'job_state': Binary(pickle.dumps(job.__getstate__(), self.pickle_protocol))
|
|
|
|
}
|
|
|
|
result = self.collection.update_one({'_id': job.id}, {'$set': changes})
|
|
|
|
if result and result.matched_count == 0:
|
|
|
|
raise JobLookupError(job.id)
|
|
|
|
|
|
|
|
def remove_job(self, job_id):
|
|
|
|
result = self.collection.delete_one({'_id': job_id})
|
|
|
|
if result and result.deleted_count == 0:
|
|
|
|
raise JobLookupError(job_id)
|
|
|
|
|
|
|
|
def remove_all_jobs(self):
|
|
|
|
self.collection.delete_many({})
|
|
|
|
|
|
|
|
def shutdown(self):
|
|
|
|
self.client.close()
|
|
|
|
|
|
|
|
def _reconstitute_job(self, job_state):
|
|
|
|
job_state = pickle.loads(job_state)
|
|
|
|
job = Job.__new__(Job)
|
|
|
|
job.__setstate__(job_state)
|
|
|
|
job._scheduler = self._scheduler
|
|
|
|
job._jobstore_alias = self._alias
|
|
|
|
return job
|
|
|
|
|
|
|
|
def _get_jobs(self, conditions):
|
|
|
|
jobs = []
|
|
|
|
failed_job_ids = []
|
|
|
|
for document in self.collection.find(conditions, ['_id', 'job_state'],
|
|
|
|
sort=[('next_run_time', ASCENDING)]):
|
|
|
|
try:
|
|
|
|
jobs.append(self._reconstitute_job(document['job_state']))
|
|
|
|
except BaseException:
|
|
|
|
self._logger.exception('Unable to restore job "%s" -- removing it',
|
|
|
|
document['_id'])
|
|
|
|
failed_job_ids.append(document['_id'])
|
|
|
|
|
|
|
|
# Remove all the jobs we failed to restore
|
|
|
|
if failed_job_ids:
|
|
|
|
self.collection.delete_many({'_id': {'$in': failed_job_ids}})
|
|
|
|
|
|
|
|
return jobs
|
|
|
|
|
|
|
|
def __repr__(self):
|
|
|
|
return '<%s (client=%s)>' % (self.__class__.__name__, self.client)
|