runIDFs() method suggestion

[rodrigokl]

Hey guys,

Before submitting to a pull request, I would like to suggest a small update to the runIDFs and prepare_run methods, both of which create/read the 'multi_runs' folder at os.mkdir('multi_runs'), but when I try to run multiple parallel instances (for example, in a 20 thread processor, 4 runs with 5 simulations each), we replace the simulations as they run, losing the outputs at the end.

To fix this, my suggestion is to add some output_dir to 'multi_runs' like this:

def runIDFs(jobs, processors=1, debug=False, output_dir='multi_runs')
 
 def prepare_run(run_id, run_data, output_dir='multi_runs')

[santoshphilip]

@rodrigokl , Just acknowledging that I have seen the issue.
From your description, it sounds like a nasty bug.

I may need your help in recreating is issue on my end.

I am caught up in some other work right now. Will check in with you in a couple of days.

[rodrigokl]

Hello @santoshphilip, here is a suggestion for the 'RunIDFs' and 'prepare_run' methods.
run_functions.zip

[santoshphilip]

@rodrigokl sorry for the delay. Somehow I did not get a notification email on your post. I start to look into this.