def process(self, data_retrieve_params, timestamps):
print 'PROCESSING BASIC CONDITION'
print 'Timestamps:', timestamps
self.timestamps = copy(timestamps)
data_retriever = DataRetriever(data_dao=self.data_dao, data_retrieve_params=data_retrieve_params)
data = data_retriever.get_data(end=timestamps['current_execution_time'], start=timestamps['last_execution_time'])
return self.__get_rule_matching_data__(data=data)
def process(self, data_retrieve_params, timestamps):
print 'PROCESSING STATISTICS CONDITION'
self.timestamps = copy(timestamps)
self.baseline.update(self.conditions_builder.baseline)
self.time_limit = self.baseline['baselineSecondsBack']
if not self.is_full_period_from_last_trigger_time():
print ('There was not full period of ' + str(self.baseline['baselineSecondsBack']) + ' seconds from last trigger time')
return []
data_retriever = DataRetriever(data_dao=self.data_dao, data_retrieve_params=data_retrieve_params)
data = data_retriever.get_data(start=self.timestamps['last_obs_trigger_time'],
end=self.timestamps['current_execution_time'])
count = data.count()
if self.baseline['baselineMinimalInstances'] == 0 or count < self.baseline['baselineMinimalInstances']:
print "Number or observations in lower than baselineMinimalInstances."
return []
sum = data.reduce(lambda x, y: reduce_func(x, y))
avg = sum / count
stdev = data.map(lambda x: float(x[1])).stdev()
self.conditions_builder.convert_to_statistics(avg, stdev)
fun = self.conditions_builder.get_condition()
filtered_data = data.filter(lambda x: fun(x)).take(FILTERED_DATA_LIMIT)
if len(filtered_data) > 0:
self.timestamps['last_obs_trigger_time'] = self.timestamps['current_execution_time']
return filtered_data
class TimebasedConditionProcessor(TimePeriodProcessor):
def __init__(self, data_dao, conditions_builder, time_limit):
super(TimebasedConditionProcessor, self).__init__()
self.data_dao = data_dao
self.conditions_builder = conditions_builder
self.time_limit = time_limit
self.timestamps = None
self.data_retriever = None
self.matching_period = None
def process(self, data_retrieve_params, timestamps):
print 'PROCESSING TIME BASED CONDITION'
self.timestamps = copy(timestamps)
self.data_retriever = DataRetriever(data_dao=self.data_dao, data_retrieve_params=data_retrieve_params)
if not self.is_full_period_from_last_trigger_time():
print ('There was not full period of ' + str(self.time_limit) + ' seconds from last trigger time')
return []
self.__find_matching_period()
if self.__is_condition_fulfilled():
print ('Matching time period length: ' + str(self.matching_period))
map_fun = self.conditions_builder.get_condition()
data = self.data_retriever.get_data(start=self.matching_period_start_time,
end=self.__get_matching_period_end_time())
filtered_data = data.filter(lambda x: map_fun(x)['passed']).take(FILTERED_DATA_LIMIT)
self.timestamps['last_obs_trigger_time'] = self.__get_matching_period_end_time()
return filtered_data
return []
def __is_condition_fulfilled(self):
return self.time_limit * 1000 <= self.matching_period
def __get_matching_period_end_time(self):
return self.matching_period_start_time + self.time_limit + 1
def __find_matching_period(self):
data = self.data_retriever.get_data(start=self.timestamps['last_obs_trigger_time'],
end=self.timestamps['current_execution_time'])
matching_period_finder = MatchingPeriodFinder(data)
self.matching_period = matching_period_finder.get_largest(self.conditions_builder.get_condition())
self.matching_period_start_time = matching_period_finder.get_largest_start_time()
def process(self, data_retrieve_params, timestamps):
print 'PROCESSING TIME BASED CONDITION'
self.timestamps = copy(timestamps)
self.data_retriever = DataRetriever(data_dao=self.data_dao, data_retrieve_params=data_retrieve_params)
if not self.is_full_period_from_last_trigger_time():
print ('There was not full period of ' + str(self.time_limit) + ' seconds from last trigger time')
return []
self.__find_matching_period()
if self.__is_condition_fulfilled():
print ('Matching time period length: ' + str(self.matching_period))
map_fun = self.conditions_builder.get_condition()
data = self.data_retriever.get_data(start=self.matching_period_start_time,
end=self.__get_matching_period_end_time())
filtered_data = data.filter(lambda x: map_fun(x)['passed']).take(FILTERED_DATA_LIMIT)
self.timestamps['last_obs_trigger_time'] = self.__get_matching_period_end_time()
return filtered_data
return []