I realise that now. Apologies about that @chris
Here is an overview of what we are trying to achieve.
We are wanting to simulate hundreds of thousands of design options for a series of rooms. We have successfully generated all of the permutations using GH/RH and have exported them all as .hbjson files. We have also generated the DA and DF input files (the .json files needed to run the recipe command). Each of these design permutations are sitting in a directory per permutation.
What we have done is designed a wrapper that stores all of that information along with the methods required to execute and postprocess the data into a series of required outputs (.csv files).
The wrapper contains the following:
- The inputs folder path as a string
- The df json file path as a string
- The da json file path as a string
- The .wea file path as a string
- The .hbjson file path as a string
- Methods for retrieving the cmd command (essentially string concatenation)
- Methods for execution of the CLI command using subprocess.run().
- Methods for post processing the data
We are calling the CLI command and outputting the result to a temporary directory as all we need is the .res file for the DF sim and the .ill/.hoy files for the DA sim.
The CLI command we are using is the lbt-recipes run recipe name command. This was as per your recommendation here: Using honeybee + radiance with CLI - #4 by chris
The issue we are bumping into is that when we try to execute the subprocess.run() method to run the simulations concurrently, we get luigi connection errors as it cannot connect to the localhost.
When trying to run the simulations in sequence we get the same error (as well as a recursion error).