def apply_on_factors(func, factors, *args):
values = map(lambda x: __delayed(func)(x, *args), factors)
return __compute(*values, scheduler="multiprocessing")
def apply_on_dataframes(func, df_list, *args):
values = map(lambda x: __delayed(func)(x, *args), df_list)
return __compute(*values, scheduler="multiprocessing")
def rolling_apply_on_dataframe_index(func, df, window, step=1, *args):
values = map(
lambda x: __delayed(func)(x, *args),
map(lambda x: df.iloc[x - window:x],
range(df.shape[0], window - 1, -step)))
return __compute(*values, scheduler="multiprocessing")
def apply_on_dataframe_columns(func, df, *args):
values = map(lambda x: __delayed(func)(x, *args),
map(lambda x: df[x], df.columns))
return __compute(*values, scheduler="multiprocessing")
def loading_data_by_split_code_list(func, code_list, window=1000):
values = map(
lambda codes: __delayed(func)(codes),
map(lambda x: code_list[x:x + window], range(0, len(code_list),
window)))
return __pd.concat(__compute(*values, scheduler="threading"))