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
https://stackoverflow.com/questions/75412399/how-to-check-each-row-trend-with-some-tolerance-by-ignoring-the-np-nan-values-in

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.

You can use:

# filter only relevant columns
# use any method
df2 = df.drop(columns='Cell')

d = (df2.ffill(axis=1)
        .diff(axis=1)
     )

m = (d.where(df2.notna())
      .lt(0)
     )

df['is_increasing'] = ~m.any(axis=1)

df['failing'] = (
  m.mul(df2.columns)
   .where(m).stack()
   .groupby(level=0).agg(list)
)

df['diff'] = (d
   .where(m).stack()
   .groupby(level=0).agg(list)
 )

p = (df2.ffill(axis=1)
        .pct_change(axis=1)
        .mul(-100).round(2)
     )
df['diff%'] = (p
   .where(m).stack()
   .groupby(level=0).agg(list)
 )

print(df)

Output:

  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]        [-2.0]         [40.0]
2    C   2  6.0  5.0   NaN          False      [D3]        [-1.0]        [16.67]
3    D   6  NaN  5.0   NaN          False      [D3]        [-1.0]        [16.67]
4    E   6  3.0  NaN   2.0          False  [D2, D6]  [-3.0, -1.0]  [50.0, 33.33]