Add Sum Subtotal to multiple layers of pandas pivot

Question

I am looking to add subtotals as done in adding subtotals to multiple layers of pandas pivot-table but ONLY including the sum aggregate, no others. I want to group by 'city', then 'name'.

When I modify the above code to only calculate the sum of rating I get the following error: NotImplementedError: > 1 ndim Categorical are not supported at this time

Any suggestions? Here is my code:

data = {"name":['paul','daniel','paul','john','daniel','daniel','john'],
        "food":['cream','chocolate','chocolate','cream','jam','butter','jam'],
        "city":['LA','NY','LA','NY','LA','NY','NY'],
        "rating":[2,3,4,5,1,3,9]}

df = pd.DataFrame(data)


df1 = pd.pivot_table(df,values = 'rating', index=['city', 'name', 'food'], aggfunc=['sum'], margins=True, margins_name="Total")



df2 = df.groupby(['city','name']).agg(['sum'])
df2 = df2.rename(index=lambda x: x+'_total', level=1)
df2 = df2.swaplevel(0, 1, axis=1)
df2 = df2.assign(food='').set_index('food', append=True)

df3 = df.groupby('city').agg(['sum'])
df3.index = pd.MultiIndex.from_arrays([df3.index + '_total', len(df3.index) * ['']])
df3 = df3.assign(name='', food='').set_index(['name','food'], append=True)
df3 = df3.swaplevel(0,1, axis=1)

df_out = pd.concat([df1,df2,df3]).sort_index()
df_out

I am ultimately trying to replicate a pivot table like below and write to excel.

Adding dtypes:

df.info() 
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 7 entries, 0 to 6
Data columns (total 4 columns):
 #   Column  Non-Null Count  Dtype 
---  ------  --------------  ----- 
 0   name    7 non-null      object
 1   food    7 non-null      object
 2   city    7 non-null      object
 3   rating  7 non-null      int64 
dtypes: int64(1), object(3)
memory usage: 352.0+ bytes

df1.info() 
<class 'pandas.core.frame.DataFrame'>
MultiIndex: 8 entries, ('LA', 'daniel', 'jam') to ('Total', '', '')
Data columns (total 1 columns):
 #   Column         Non-Null Count  Dtype
---  ------         --------------  -----
 0   (sum, rating)  8 non-null      int64
dtypes: int64(1)
memory usage: 795.0+ bytes

df2.info() 
<class 'pandas.core.frame.DataFrame'>
MultiIndex: 4 entries, ('LA', 'daniel_total', '') to ('NY', 'john_total', '')
Data columns (total 2 columns):
 #   Column         Non-Null Count  Dtype 
---  ------         --------------  ----- 
 0   (sum, food)    4 non-null      object
 1   (sum, rating)  4 non-null      int64 
dtypes: int64(1), object(1)
memory usage: 607.0+ bytes