Skip to content

examples

Jump to bottom
weilixu edited this page Jul 5, 2018 · 4 revisions

Use Cases

How to retrieve a model's results?

Once you have completed a simulation, you will receive a model API key or a model track token. The model API key looks like UIUD and the track token is something looks like 1-11-1111. The differences between these two keys are explained in the Getting started. Despite the differences, both keys work for retrieving a model's results.

import BuildSimHubAPI as bsh_api
project_api_key = 'f9dsadb3-253f-438c-a321-82aew4b9424e'
model_api_key = '609da2acf-de2-44fa-9883-a0df8bdsdb56'
# model_track_token = "1-11-111"

bsh = bsh_api.BuildSimHubAPIClient()
# This works too:
# results = bsh.model_results(project_api_key, model_track_token)
results = bsh.model_results(project_api_key, model_api_key)
print(str(results.net_site_eui()) + ' ' + results.last_parameter_unit)

The results object contains all the information about the simulation results. Detail list of data that you can extract from the BuildSim Cloud is listed in Simulation Results page.

How to retrieve a parametric study results?

Unlike the results from a single model, retrieving results from parametric study requires a project API key and a model API key.

import BuildSimHubAPI as bsh_api
project_api_key = 'f9dsadb3-253f-438c-a321-82aew4b9424e'
model_api_key = '609da2acf-de2-44fa-9883-a0df8bdsdb56'

bsh = bsh_api.BuildSimHubAPIClient()
# This works too:
results = bsh.parametric_results(project_api_key, model_api_key)
print(str(results.net_site_eui()) + ' ' + results.last_parameter_unit)

Use Pandas to retrieve simulation results

Parametric study usually involves hundreds of simulations. By using the parametric result object, we can extract a same result value from these simulations. However, parametric is a part of approach for exploring the best design. So we still have to decide which design we want to pursue after the study. Once we have determined a design among hundreds of simulations, how can we dive deeper to investigate the details of that particular simulation? Before we do that, we need a function to process the model list. You will need pandas package for this function.

import pandas as pd
def post_process_models(df):  
    param_list = list()  
    for index, row in df.iterrows():  
        msg = row['commit_msg']  
        parameters = msg.split(',')  
        data_dict = dict()  
        for k in range(len(parameters)):  
            title, val = parameters[k].split(':')  
            data_dict[title.strip()] = float(val.strip())  
        param_list.append(data_dict)  
    parameter_df = pd.DataFrame(param_list)  
    return pd.concat([df, parameter_df], axis=1)

Then, let's apply the filters on the models:

import BuildSimHubAPI as bshapi  

# paste your project_api_key and model_api_key here  
project_api_key = 'f9dsadb3-253f-438c-a321-82aew4b9424e'
model_api_key = '609da2acf-de2-44fa-9883-a0df8bdsdb56'

bsh = bshapi.BuildSimHubAPIClient()  
data_list = bsh.model_list(project_api_key, model_api_key)

post_df = post_process_models(df)  
val = post_df.loc[(post_df['HeatingEff'] == 0.88) & (post_df['LPD'] == 0.858)]['commit_id']  
model_id = val.values[0]

# extract the single model eui  
results = bsh.model_results(project_api_key, model_id)  
print(str(results.net_site_eui()) + ' ' + results.last_parameter_unit)

You have successfully extract the net site eui of the simulation, whose heating efficiency is 88% and lighting power density is 0.858 W/ft2.

How to download all the models generated in a parametric study

You can download all the models under a parametric study with a for loop.

import BuildSimHubAPI as bshapi
import os
import time
import pandas

# get your project key
project_api_key = 'f98aadb3-254f-428d-a321-82a6e4b9424c'
# location where you want to store the models
folder_dir = '/Users/download/'
model_api_key = 'bd70dc9c-633c-4a2c-8bac-4f4032a7d7'

bsh = bshapi.BuildSimHubAPIClient()
data_list = bsh.model_list(project_api_key, model_api_key)

df = pd.DataFrame(data_list)  
# drop the seed model
df = df[df.commit_msg != 'INIT']

for index, row in df.T.iteritems(): 
    model = bsh.model_results(project_api_key, row['commit_id'])
    print('Downloading the model: ' + key)
    content = model.download_model()
    time.sleep(5)
    # save to a file in local
    full_path = os.path.join(folder_dir, row['commit_msg'] + '.idf')
    print('Write the model to the file: ' + full_path)
    with open(full_path, 'w') as file:
        file.write(content)
        file.close()
    print('Done!')

BuildSimHub Inc

Home / Blog / Twitter@buildsim / LinkedIn@buildsim

Clone this wiki locally