Allow using a different class for scheduler and worker

Would you please make such changes to allow overriding the class used for scheduler and worker?

--- a/huey/consumer.py
+++ b/huey/consumer.py

 class Consumer(object):
+
+    scheduler_class = Scheduler
+    worker_class = Worker
+
     def __init__(self, huey, workers=1, periodic=True, initial_delay=0.1,
                  backoff=1.15, max_delay=10.0, utc=True, scheduler_interval=1,
                  worker_type='thread', check_worker_health=True,

     def _create_worker(self):
-        return Worker(
+        return self.worker_class(
             huey=self.huey,
             default_delay=self.default_delay,
             max_delay=self.max_delay,

     def _create_scheduler(self):
-        return Scheduler(
+        return self.scheduler_class(
             huey=self.huey,
             interval=self.scheduler_interval,
             utc=self.utc,

Because currently I have to override Consumer._create_worker and Consumer._create_scheduler copying the original implementation and changing the class used which is not DRY and I risk to miss the moment when the original implementation changes.

[coleifer]

I didn’t think it was a good idea at first. After a day I thought it would probably be a fine addition: db68a555080270ff4195ccd861542e7a4164cc6e

[coleifer]

Do I understand correctly that … you prefer to keep Huey non-extendable so it can serve mostly your purposes?

It was quite easy to extend Huey in exactly the way you wished. This just makes it even easier – two attributes instead of two methods for the simple case.

I do keep Huey extensible as much as I think is useful. I rely on people who use the library to help find the places where this becomes difficult. Look at the storage backend abstractions – they are extremely amenable to extension or customization.