User defined job.run.metadata derived from 'post-processed results'?

tfedyna · January 10, 2022, 3:31pm

Hello! Having a grand time with polination-streamlit!

I was curious if this already exists, and or what all would need to be done specifically for these features in a recipe/plugin:

user_inputs via user intput component in grasshopper are added to the metadata of each run in a job; creating that really nice synchronicity with pollination_streamlit.dataframe.

Currently: is there a means to post process simulation results; and add them to the run.metadata; to further extend and lean on pollination_streamlit.dataframe?

At-least for what I am doing currently; that would definitely save a ton of time performance wise minimizing volume of I/O with streamlit app, ; and being able to not use results.sql from app; or perhaps not even generate eplusout.sql if not needed (from what I understand writing eplusout.sql; atleast with Eplus v9.4; was a time intensive part of simulation; and not doing sql output seemed to speed things up.

and it would be really cool to be able to do api calls to large parametric runs; and basically just ducktyping off job.runs_dataframe for plotting explicitly.

Understanding that; with the results JSON returning commands via honeybee plugin: what I’m talking about is kind of just lazy just thought it might be kinda nifty to further streamline some things!

First few days I’ve been able to get really heavy into all this so apologies for any day-one-stuff overlooks on my part!

best
-trevor

tfedyna · January 10, 2022, 3:50pm

fairly sure after some more reading in pollinations github: a plugin would be the move, atleast for the ‘basic functionality’, still curious as to the appendment to the job.run.metadata!

mostapha · March 3, 2022, 11:50pm

Hi @tfedyna, apologies for the late reply. Here is an example of how you can extract them from inputs. It’s not really elegant but it works:

github.com

pollination/sample-apps/blob/48af71a983c594770d35bfb6161138b576fad015/energy-use/app.py#L93-L123

      
        
            def extract_user_inputs(job: Job):
                """Extract the various combinations of user inputs for a job."""
                # variables to be used while looping through the runs
                input_map, all_inputs = {}, {}
                avoid_inputs = (
                    'additional-string', 'ddy', 'epw', 'model', 'units', 'viz-variables')
            
            
    # loop through the runs and find all unique user inputs
                for run in job.runs:
                    inp_key = []
                    for inp in run.status.inputs:
                        if inp.name not in avoid_inputs:
                            try:
                                inp_key.append(float(inp.value))
                            except (TypeError, ValueError):  # not numerical values
                                inp_key.append(inp.value)
                            try:
                                all_inputs[inp.name].add(inp.value)
                            except KeyError:
                                all_inputs[inp.name] = {inp.value}

This file has been truncated. show original

tfedyna · March 6, 2022, 5:36pm

thanks Mostapha!!! most clutch