How to check the row trend and add the difference and difference percentage in separate columns for the failing cases

An extension to the problem statement how to check each row trend with some tolerance by ignoring the np.nan values in pandas dataframe

import pandas as pd
import numpy as np
    d = {'Cell':['A','B','C','D','E'],'D1':[5, 2, 2, 6,6], 'D2':[np.nan, 5, 6, np.nan,3], 'D3':[7,np.nan, 5, 5,np.nan], 'D6':[17, 3, np.nan,np.nan,2]}
    df = pd.DataFrame(d)


Cell  D1   D2   D3    D6
0    A   5  NaN  7.0  17.0
1    B   2  5.0  NaN   3.0
2    C   2  6.0  5.0   NaN
3    D   6  NaN  5.0   NaN
4    E   6  3.0  NaN   2.0

i want output like this with additional columns diff and diff% along with is_increasing and failing columns

  Cell  D1   D2   D3    D6  is_increasing?   failing  diff          diff%
0    A   5  NaN  7.0  17.0            True       NaN  NaN           NaN
1    B   2  5.0  NaN   3.0            False      [D6]  [-2]         [40%]
2    C   2  6.0  5.0   NaN           False      [D3]  [-1]        [16.6%]
3    D   6  NaN  5.0   NaN           False      [D3]  [-1]        [16.6%]
4    E   6  3.0  NaN   2.0           False  [D2, D6]  [-3,-1]   [50%,33%]

Explanation of the columns:

is_increasing --> whether the values are strictly increasing or not
failing --> columns whether strictly increasing is not followed when compared with previous value
diff --> difference of the values where there is failing cases
diff% --> difference in terms of percentages for the failing cases

between (6,5) numbers in the columns

diff column --> 5-6=-1
diff%--> 1-(5/6)=16.6%

Please let me the solution to this problem, i tried different ways but not able to come up with solution.

source https://stackoverflow.com/questions/75440543/how-to-check-the-row-trend-and-add-the-difference-and-difference-percentage-in-s