What is the easiest way to do this transformation in Python?

I have a table in this format, I want to calculate group-wise weightage average such that if there is null value for a particular value_KPI its weightage should be distributed equally to other KPIs from same group.

something like this:

Please help me with python code to do this.

Let's define a simple helper function:

def distribute(g):
    nans = g['value_KPI'].isna()
    g.loc[~nans, 'Weightages'] += g.loc[nans, 'Weightages'].sum()/sum(~nans)
    g.loc[nans, 'Weightages'] = 'NaN'
    return g

Now we apply it to each group after groupby

df.groupby(['Groups']).apply(distribute)

output:

    Groups    KPIs      Weightages    value_KPI
--  --------  ------  ------------  -----------
 0  G1        KP1              1             45
 1  G1        KP2            nan            nan
 2  G1        KP3            nan            nan
 3  G2        KP4              0.4           34
 4  G2        KP5            nan            nan
 5  G2        KP6              0.3           90
 6  G2        KP7              0.3           45