DataFrame cell value grouped by 'groupby ()'

I have the following DataFrame of pandas that I call reacciones :

    Modelo          Apoyo   Grupo   Fx      Fy      Fz
0   Reacciones 00   P1      G       -1.0    0       162.0
1   Reacciones 00   P1      Q       -3.0    0       89.0
2   Reacciones 00   P1      Q       2.0     0       570.0
3   Reacciones 00   P2      G       5.0     0       19.0
4   Reacciones 00   P2      Q       4.0     0       5.0
5   Reacciones 00   P2      Q       2359.0  0       2894.0
6   Reacciones 00   P2      G       -4.0    0       15.0
7   Reacciones 00   P2      Q       -4.0    0       5.0
8   Reacciones 00   P2      Q       2740.0  0       -3464.0
0   Reacciones 01   P1      G       -0.0    0       16.0
1   Reacciones 01   P1      Q       0.0     0       89.0
2   Reacciones 01   P1      Q       -0.0    0       -22.0
3   Reacciones 01   P1      DL      1.0     3       -5.0
4   Reacciones 01   P1      DL      -2.0    4       4.0
5   Reacciones 01   P2      G       0.0     0       2.0
6   Reacciones 01   P2      Q       4.0     0       5.0
7   Reacciones 01   P2      Q       -92.0   0       -113.0
8   Reacciones 01   P2      DL      -3.0    0       50.0
9   Reacciones 01   P2      DL      -4.0    0       40.0
10  Reacciones 01   P2      G       -0.0    0       2.0
11  Reacciones 01   P2      Q       -4.0    0       5.0
12  Reacciones 01   P2      Q       -107.0  0       136.0
13  Reacciones 01   P2      DL      -10.0   0       50.0
14  Reacciones 01   P2      DL      5.0     0       -40.0

I get the maximum and minimum values of the columns Fx , Fy and Fz grouped by Modelo , Apoyo and Grupo with the following code:

agrupacion = ['Modelo', 'Apoyo', 'Grupo']
fuerzas = ['Fx', 'Fy', 'Fz']

reacciones.groupby(agrupacion).agg({x: [max, min] for x in fuerzas})

                                Fz              Fy          Fx
                                max     min     max min     max     min
Modelo          Apoyo   Grupo                       
Reacciones 00   P1      G       162.0   162.0    0  0       -1.0    -1.0
                        Q       570.0   89.0     0  0       2.0     -3.0
                P2      G       19.0    15.0     0  0       5.0     -4.0
                        Q       2894.0  -3464.0  0  0       2740.0  -4.0
Reacciones 01   P1      DL      4.0     -5.0     4  3       1.0     -2.0
                        G       16.0    16.0     0  0       -0.0    -0.0
                        Q       89.0    -22.0    0  0       0.0     0.0
                P2      DL      50.0    -40.0    0  0       5.0     -10.0
                        G       2.0     2.0      0  0       0.0     0.0
                        Q       136.0   -113.0   0  0       4.0     -107.0

The sum of the positive values and the sum of the negative values per group is obtained in the following way, for example for the column Fx :

reacciones.groupby(['Modelo', 'Apoyo', 'Grupo'])['Fx'].\
    agg([('suma_pos', lambda x : x[x > 0].sum()),
         ('suma_neg', lambda x : x[x < 0].sum())])

                                 suma_pos   suma_neg
Modelo          Apoyo   Grupo       
Reacciones 00   P1      G        NaN        -1.0
                        Q        2.0        -3.0
                P2      G        5.0        -4.0
                        Q        5103.0     -4.0
Reacciones 01   P1      DL       1.0        -2.0
                        G        NaN        NaN
                        Q        NaN        NaN
                P2      DL       5.0        -17.0
                        G        NaN        NaN
                        Q        4.0        -203.0

Now, how do I get a value from these grouped tables?

For example:

Reacciones 00 -> P1 -> Q -> Fx max = 2.0

Reacciones 00 -> P2 -> Q -> Fx suma_pos = 5103.0