# weighted mean
v = np.average(df[name].values[bi], None, np.nan_to_num(ws))
elif wt and mode in ["q1", "q2", "q3"]:
q = (["q1", "q2", "q3"].index(mode) + 1) * 0.25
v = weighted_quantiles(df[name].values, [q],
np.prod([df[_].values for _ in wt], 0))[0]
elif hasattr(pd.Series, mode):
fn = eval('pd.Series.' + mode)
v = fn(df[name])
elif mode == 'major':
if df[name].any():
v = df[name].value_counts().idxmax()
elif mode in ["q1", "q2", "q3"]:
q = (["q1", "q2", "q3"].index(mode) + 1) * 0.25
v = df[name].quantile(q)
elif mode.startswith('p') and str.isnumeric(mode[1:]):
v = np.percentile(df[name], float(mode[1:]))
r.append(v)
return (r)
def main(*args):
pd_save_dataframe(bm_breakdown(args[0], args[1], args[2], args[3]),
args[4])
if __name__ == "__main__" and sys.argv[0].endswith('bm_breakdown.py'):
usage_gui(__doc__)
#!python from _gui import usage_gui usage_gui(None)
