came back to edit a bug out of the snippet
Long story short:
Nothing gets me STOKED like some radness from things LBT related.
First few days working heavy with PO API/SDK and streamlit: THRILLED.
I’d needed to get some data from the overall job dataframe (LBT pandas? YES.)
about 6000 runs via 4 jobs; needed to get the user inputs as am building some reporting PCP plots for a parametric study, some borrowing and ‘abuse’ of some pollination-streamlit modules later and this is what has me so stoked, and wanted to share:
def download_inputs(job: Job, results_folder):
avoid_inputs = (
'additional-string', 'ddy', 'epw', 'model', 'units', 'viz-variables')
runs = job.runs
results_folder = results_folder # data/applicable_job_folder
df = job.runs_dataframe.dataframe
todrop = ['additional-string', 'ddy', 'epw', 'model', 'html', 'idf', 'job-id', 'run-status', 'sim-par', 'sql', 'zsz', 'err']
df = df.drop(todrop, axis=1)
df.rename(columns={'run-id':'run_id'}, inplace=True)
for run in runs:
output_folder = os.path.join(results_folder, run.id)
item = df.loc[df.index == run.id]
item.to_csv(os.path.join(output_folder, 'user_input.csv'))
return
def_url = 'full url to PO job'
api_key = None
job = job_selector(api_key = api_key, default=def_url)
res_fldr = 'data\local_job_dir_name'
download_inputs(job, res_fldr)
I REALLY like pandas and use it sometimes daily; these capabilities just have me super stoked!
thanks LBT crew! thanks for lbt pandas @antoinedao!!! suuper awesome
Feels like this post is still relevant then:
