Signal Transformation Computation is too slow

[antoinecarme]

We copy the dataset and use a loop over the whole dataset which is too baaaaad :

'''

def specific_invert(self, df):
    df_orig = df.copy();
    df_orig.iloc[0] = df.iloc[0];
    for i in range(1,df.shape[0]):
        df_orig.iloc[i] = df.iloc[i] -  df.iloc[i - 1]
    return df_orig;

'''

There is room for improvement.

[antoinecarme]

Fix : use aggreagtes like cumsum, cumprod on shifted series etc

for the previous example , an equivalent rewrite is :

def specific_invert(self, df):
    df_orig = df - df.shift(1);
    df_orig.iloc[0] = df.iloc[0];
    return df_orig;

[antoinecarme]

detect all such errors in all transformations.