Awesome
JobtasticMixins
Package includes:
AVGTimeRedis class
The class that helps automate calculate an avarage time for different kind tasks and saves result into Redis DB
Install
pip install jobtastic-mixins
or install from Github
pip install https://github.com/abbasovalex/JobtasticMixins/archive/master.zip
Usage
- You must will add the mixin class as first argument
- You should set
default_avg_time
in the seconds - You must use
self.update_progress()
without arguments
class YourTask(AVGTimeRedis, JobtasticTask):
default_avg_time = 60 # 1 minute
...
def calculate_result(self, ..., **kwargs):
...
default_avg_time
used when task still never calculated
It will be encompass arithmetical mean after task was executed.
Forecast accuracy depends from count executed tasks
Example with AVGTimeRedis class
The class that helps automate calculate an avarage time for different kind tasks and saves result into Redis DB
Let's take a look at the example task using with JobtasticMixins and AVGTimeRedis class:
from time import sleep
from jobtastic import JobtasticTask
from jobtasticmixins import AVGTimeRedis
class LotsOfDivisionTask(AVGTimeRedis, JobtasticTask):
"""
Division is hard. Make Celery do it a bunch.
"""
significant_kwargs = [
('numerators', str),
('denominators', str),
]
herd_avoidance_timeout = 60
cache_duration = 0
# optional variable was added. by default is 30 seconds
default_avg_time = 90
def calculate_result(self, numerators, denominators, **kwargs):
results = []
for count, divisors in enumerate(zip(numerators, denominators)):
numerator, denominator = divisors
results.append(numerator / denominator)
# it will be auto calculated
self.update_progress()
sleep(0.1)
# set finish=True for avoid trouble
self.update_progress(finish=True)
return results
Under the hood:
- AVGTimeRedis gets settings.BROKER_URL and connects to Redis
- It counts the tasks and the workers and uses to calculating
More details you can see into source