Method inside method, is there a better way?

[kaanaksit]

I am a happy user of dispy. First of all, thank you for making it happen!

Context

Assume there are three files that represents my python program:

• main.py, where I gather some shell arguments and load some json dictionary and start the entire program,
  • routine.py, where I pass some variables from main.py and start my compute method/definition/function,
  • simulation.py, where the compute method/definition/function is.

Inside routine.py, I import compute from simulation.py and start a cluster accordingly.

Sturggle

I need some more methods/definitions/functions from simulation.py during compute, and my solution was to embed these other things into the compute as in below example:

def compute():
      def other_function_1():
            return 5
      def other_function_2():
           return 10
      result =  other_function_1()+other_function_2()
      return result

I find my solution to be very ugly.If I leave the other_functions outside of compute, I always get this can not find module error even if I try importing simulate.py inside compute.

Is there a better way to do this? Your help and suggestion is very much appreciated!

[pgiri]

One possibility (easier approach?) is to put functions in a module (separate file), send it with depends and have compute function import that module and use functions in that module.

[kaanaksit]

Thank you. That resolved my case. I followed your suggestion and

• import the new module first where I generate the cluster,
• added the new module as depends=[newmodule] during cluster generation,
• inside compute, I also import that module and use it.

[pgiri]

Great! Note that there is no need to import new module in client program (where cluster is generated), unless functions in new module are used in processing results etc.