You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
bazarr/libs/apscheduler/triggers/cron/__init__.py

240 lines
9.0 KiB

from datetime import datetime, timedelta
from tzlocal import get_localzone
import six
from apscheduler.triggers.base import BaseTrigger
from apscheduler.triggers.cron.fields import (
BaseField, MonthField, WeekField, DayOfMonthField, DayOfWeekField, DEFAULT_VALUES)
from apscheduler.util import (
datetime_ceil, convert_to_datetime, datetime_repr, astimezone, localize, normalize)
class CronTrigger(BaseTrigger):
"""
Triggers when current time matches all specified time constraints,
similarly to how the UNIX cron scheduler works.
:param int|str year: 4-digit year
:param int|str month: month (1-12)
:param int|str day: day of month (1-31)
:param int|str week: ISO week (1-53)
:param int|str day_of_week: number or name of weekday (0-6 or mon,tue,wed,thu,fri,sat,sun)
:param int|str hour: hour (0-23)
:param int|str minute: minute (0-59)
:param int|str second: second (0-59)
:param datetime|str start_date: earliest possible date/time to trigger on (inclusive)
:param datetime|str end_date: latest possible date/time to trigger on (inclusive)
:param datetime.tzinfo|str timezone: time zone to use for the date/time calculations (defaults
to scheduler timezone)
:param int|None jitter: delay the job execution by ``jitter`` seconds at most
.. note:: The first weekday is always **monday**.
"""
FIELD_NAMES = ('year', 'month', 'day', 'week', 'day_of_week', 'hour', 'minute', 'second')
FIELDS_MAP = {
'year': BaseField,
'month': MonthField,
'week': WeekField,
'day': DayOfMonthField,
'day_of_week': DayOfWeekField,
'hour': BaseField,
'minute': BaseField,
'second': BaseField
}
__slots__ = 'timezone', 'start_date', 'end_date', 'fields', 'jitter'
def __init__(self, year=None, month=None, day=None, week=None, day_of_week=None, hour=None,
minute=None, second=None, start_date=None, end_date=None, timezone=None,
jitter=None):
if timezone:
self.timezone = astimezone(timezone)
elif isinstance(start_date, datetime) and start_date.tzinfo:
self.timezone = start_date.tzinfo
elif isinstance(end_date, datetime) and end_date.tzinfo:
self.timezone = end_date.tzinfo
else:
self.timezone = get_localzone()
self.start_date = convert_to_datetime(start_date, self.timezone, 'start_date')
self.end_date = convert_to_datetime(end_date, self.timezone, 'end_date')
self.jitter = jitter
values = dict((key, value) for (key, value) in six.iteritems(locals())
if key in self.FIELD_NAMES and value is not None)
self.fields = []
assign_defaults = False
for field_name in self.FIELD_NAMES:
if field_name in values:
exprs = values.pop(field_name)
is_default = False
assign_defaults = not values
elif assign_defaults:
exprs = DEFAULT_VALUES[field_name]
is_default = True
else:
exprs = '*'
is_default = True
field_class = self.FIELDS_MAP[field_name]
field = field_class(field_name, exprs, is_default)
self.fields.append(field)
@classmethod
def from_crontab(cls, expr, timezone=None):
"""
Create a :class:`~CronTrigger` from a standard crontab expression.
See https://en.wikipedia.org/wiki/Cron for more information on the format accepted here.
:param expr: minute, hour, day of month, month, day of week
:param datetime.tzinfo|str timezone: time zone to use for the date/time calculations (
defaults to scheduler timezone)
:return: a :class:`~CronTrigger` instance
"""
values = expr.split()
if len(values) != 5:
raise ValueError('Wrong number of fields; got {}, expected 5'.format(len(values)))
return cls(minute=values[0], hour=values[1], day=values[2], month=values[3],
day_of_week=values[4], timezone=timezone)
def _increment_field_value(self, dateval, fieldnum):
"""
Increments the designated field and resets all less significant fields to their minimum
values.
:type dateval: datetime
:type fieldnum: int
:return: a tuple containing the new date, and the number of the field that was actually
incremented
:rtype: tuple
"""
values = {}
i = 0
while i < len(self.fields):
field = self.fields[i]
if not field.REAL:
if i == fieldnum:
fieldnum -= 1
i -= 1
else:
i += 1
continue
if i < fieldnum:
values[field.name] = field.get_value(dateval)
i += 1
elif i > fieldnum:
values[field.name] = field.get_min(dateval)
i += 1
else:
value = field.get_value(dateval)
maxval = field.get_max(dateval)
if value == maxval:
fieldnum -= 1
i -= 1
else:
values[field.name] = value + 1
i += 1
difference = datetime(**values) - dateval.replace(tzinfo=None)
return normalize(dateval + difference), fieldnum
def _set_field_value(self, dateval, fieldnum, new_value):
values = {}
for i, field in enumerate(self.fields):
if field.REAL:
if i < fieldnum:
values[field.name] = field.get_value(dateval)
elif i > fieldnum:
values[field.name] = field.get_min(dateval)
else:
values[field.name] = new_value
return localize(datetime(**values), self.timezone)
def get_next_fire_time(self, previous_fire_time, now):
if previous_fire_time:
start_date = min(now, previous_fire_time + timedelta(microseconds=1))
if start_date == previous_fire_time:
start_date += timedelta(microseconds=1)
else:
start_date = max(now, self.start_date) if self.start_date else now
fieldnum = 0
next_date = datetime_ceil(start_date).astimezone(self.timezone)
while 0 <= fieldnum < len(self.fields):
field = self.fields[fieldnum]
curr_value = field.get_value(next_date)
next_value = field.get_next_value(next_date)
if next_value is None:
# No valid value was found
next_date, fieldnum = self._increment_field_value(next_date, fieldnum - 1)
elif next_value > curr_value:
# A valid, but higher than the starting value, was found
if field.REAL:
next_date = self._set_field_value(next_date, fieldnum, next_value)
fieldnum += 1
else:
next_date, fieldnum = self._increment_field_value(next_date, fieldnum)
else:
# A valid value was found, no changes necessary
fieldnum += 1
# Return if the date has rolled past the end date
if self.end_date and next_date > self.end_date:
return None
if fieldnum >= 0:
next_date = self._apply_jitter(next_date, self.jitter, now)
return min(next_date, self.end_date) if self.end_date else next_date
def __getstate__(self):
return {
'version': 2,
'timezone': self.timezone,
'start_date': self.start_date,
'end_date': self.end_date,
'fields': self.fields,
'jitter': self.jitter,
}
def __setstate__(self, state):
# This is for compatibility with APScheduler 3.0.x
if isinstance(state, tuple):
state = state[1]
if state.get('version', 1) > 2:
raise ValueError(
'Got serialized data for version %s of %s, but only versions up to 2 can be '
'handled' % (state['version'], self.__class__.__name__))
self.timezone = state['timezone']
self.start_date = state['start_date']
self.end_date = state['end_date']
self.fields = state['fields']
self.jitter = state.get('jitter')
def __str__(self):
options = ["%s='%s'" % (f.name, f) for f in self.fields if not f.is_default]
return 'cron[%s]' % (', '.join(options))
def __repr__(self):
options = ["%s='%s'" % (f.name, f) for f in self.fields if not f.is_default]
if self.start_date:
options.append("start_date=%r" % datetime_repr(self.start_date))
if self.end_date:
options.append("end_date=%r" % datetime_repr(self.end_date))
if self.jitter:
options.append('jitter=%s' % self.jitter)
return "<%s (%s, timezone='%s')>" % (
self.__class__.__name__, ', '.join(options), self.timezone)