Difference to previous row in python on a specific group in percentage

Question

I have a similiar question to this one and to this one. But the main difference is that I do not have any timestamp column, my data looks the following:

df = df.groupby(['A', 'B', 'Year', 'Month']['Sales'].sum().reset_index().sort_values(by=(['A','B','Year','Month']))

A    B  Year Month Sales
Der Die 2019  1     2
Der Die 2019  2     3
...
Die Das 2019  1    40
Die Das 2019  2    30

How can i get the difference for each group (pair of A and B) to the previous month (in this case previos row), that I get the following output:

A    B  Year Month Sales Diff
Der Die 2019  1     2    0
Der Die 2019  2     3    3/2
...
Die Das 2019  1    40  0
Die Das 2019  2    30  30/40

Note that in case of missing values, i want to have a 0 in it.

When I use:

df['diff'] = df.groupby(['A', 'B'])['Sales'].diff().fillna()

I only get the difference, but not the percentage:

A    B  Year Month Sales Diff
Der Die 2019  1     2    0
Der Die 2019  2     3    3-2
...
Die Das 2019  1    40  0
Die Das 2019  2    30  30-40