I have a Flask app that sends tasks to an RQ worker which will uses Pykaldi. I am installing Pykaldi via conda and the package is quite large.
At the moment, I have my Flask app and an RQ worker running on different containers but both have Pykaldi installed which consumes space. The reason is that the flask container needs to enqueue jobs to the redis queue. To do this, it calls the background task's code which uses Pykaldi.
Is there any way to optimize this and avoid installing libraries that will, at the end of the day, only be used on the worker side?
